We present AvatarTex, a high-fidelity facial texture reconstruction framework capable of generating both stylized and photorealistic textures from a single image. Existing methods struggle with stylized avatars due to the lack of diverse multi-style datasets and challenges in maintaining geometric consistency in non-standard textures. To address these limitations, AvatarTex introduces a novel three-stage diffusion-to-GAN pipeline. Our key insight is that while diffusion models excel at generating diversified textures, they lack explicit UV constraints, whereas GANs provide a well-structured latent space that ensures style and topology consistency. By integrating these strengths, AvatarTex achieves high-quality topology-aligned texture synthesis with both artistic and geometric coherence. Specifically, our three-stage pipeline first completes missing texture regions via diffusion-based inpainting, refines style and structure consistency using GAN-based latent optimization, and enhances fine details through diffusion-based repainting. To address the need for a stylized texture dataset, we introduce TexHub, a high-resolution collection of 20,000 multi-style UV textures with precise UV-aligned layouts. By leveraging TexHub and our structured diffusion-to-GAN pipeline, AvatarTex establishes a new state-of-the-art in multi-style facial texture reconstruction. TexHub will be released upon publication to facilitate future research in this field.

Existing 3D human motion generation and understanding methods often exhibit limited interpretability, restricting effective mutual enhancement between these inherently related tasks. While current unified frameworks based on large language models (LLMs) leverage linguistic priors, they frequently encounter challenges in semantic alignment and task coherence. Moreover, the next-token prediction paradigm in LLMs is ill-suited for motion sequences, causing cumulative prediction errors. To address these limitations, we propose UniMo, a novel framework that integrates motion-language information and interpretable chain of thought (CoT) reasoning into the LLM via supervised fine-tuning (SFT). We further introduce reinforcement learning with Group Relative Policy Optimization (GRPO) as a post-training strategy that optimizes over groups of tokens to enforce structural correctness and semantic alignment, mitigating cumulative errors in motion token prediction. Extensive experiments demonstrate that UniMo significantly outperforms existing unified and task-specific models, achieving state-of-the-art performance in both motion generation and understanding.

Recovering sewing patterns from draped 3D garments is a challenging problem in human digitization research. Unlike draping from designed sewing pattern with mature physical simulation engines, the inverse process requires mapping complex garment surfaces back to parametric 2D patterns, which existing methods still struggle with. To this end, we propose a two-stage methodology that recovers sewing patterns from 3D garments via an intermediate representation, BoxMesh, which inherently encodes garment-level geometry and panel-level details in 3D space. In Stage I, a geometry-oriented auto-regressive model predicts BoxMesh from the input 3D garment. In Stage II, a semantic-aware auto-regressive model parses BoxMesh into parametric sewing patterns. This decomposition separately tackles geometric inversion and numerical reasoning, enabling more accurate recovery. Experiments have demonstrated that our method not only achieves state-of-the-art performance on the GarmentCodeData benchmark but also can be applied effectively to real scans and single-view images.

While recent advancements have demonstrated remarkable progress in general 3D shape generation, the challenge of automatically generating wearable 3D assets remains largely unexplored. To address this gap, we present BAG - a Body-aligned Asset Generation method that produces 3D wearable assets which can be automatically fitted onto given 3D human bodies. This is achieved by controlling the 3D generation process using human body shape and pose information. Specifically, we first construct a general single-image-to-consistent-multi-view diffusion model, and train it on the large-scale Objaverse dataset to ensure diversity and generalizability. We then train a body-conditioned multi-view ControlNet to guide the generator toward producing body-aligned multi-view images. The control signal leverages multi-view 2D projections of the target human body, where pixel values represent the XYZ coordinates of the body surface in a canonical space. The resulting body-conditioned multi-view diffusion outputs body-aligned images, which are subsequently fed into a native 3D diffusion model to reconstruct the 3D shape of the asset. Finally, we recover the similarity transformation using multi-view silhouette supervision and mitigate asset-body penetration using physics-based simulation, ensuring accurate asset fitting onto the target body. Experimental results demonstrate that our method significantly outperforms existing approaches in terms of prompt adherence, shape diversity, and shape quality.

Text-guided texture generation has been rapidly developed with the proliferation of generative artificial intelligence for creating three-dimensional textured objects. However, existing text-guided texture generation methods often suffer from artifacts such as inconsistent visual appearance across different views, the Janus problems and seams in texture maps. To address these issues, we propose a novel text-guided texture generation method, named WonderTex. It achieves the generation of high-quality, view-consistent, and seamless texture maps through a two-stage pipeline. Specifically, we fine-tune a Stable Diffusion model using a large dataset to obtain a multi-view image diffusion model capable of generating a 4-view grid. This model serves as the foundation for producing four consistent views and establishing the base texture in the first stage. Subsequently, an automatic view selection and inpainting strategy is employed to effectively fill and refine the texture maps in the second stage. Extensive experiments have shown that our method is effective and robust, capable of generating high-quality textures with various meshes and prompts, outperforming baseline methods in terms of texture details, view consistency, and other metrics.

We introduce SAMPro3D for zero-shot instance segmentation of 3D scenes. Given the 3D point cloud and multiple posed RGB-D frames of 3D scenes, our approach segments 3D instances by applying the pretrained Segment Anything Model (SAM) to 2D frames. Our key idea in-volves locating SAM prompts in 3D to align their projected pixel prompts across frames, ensuring the view consistency of SAM-predicted masks. Moreover, we suggest selecting prompts from the initial set guided by the information of SAM-predicted masks across all views, which enhances the overall performance. We further propose to consolidate different prompts if they are segmenting different surface parts of the same 3D instance, bringing a more comprehensive segmentation. Notably, our method does not require any additional training. Extensive experiments on diverse benchmarks show that our method achieves comparable or better performance compared to previous zero-shot or fully supervised approaches, and in many cases surpasses human annotations. Furthermore, since our fine-grained predictions often lack annotations in available datasets, we present ScanNet200-Fine50 test data which provides fine-grained annotations on 50 scenes from ScanNet200 dataset.

Using parts of existing models to rebuild new models, commonly termed as example-based modeling, is a classical methodology in the realm of computer graphics. Previous works mostly focus on shape composition, making them very hard to use for realistic composition of 3D objects captured from real-world scenes. This leads to combining multiple NeRFs into a single 3D scene to achieve seamless appearance blending. However, the current SeamlessNeRF method struggles to achieve interactive editing and harmonious stitching for real-world scenes due to its gradient-based strategy and grid-based representation. To this end, we present an example-based modeling method that combines multiple Gaussian fields in a point-based representation using sample-guided synthesis. Specifically, as for composition, we create a GUI to segment and transform multiple fields in real time, easily obtaining a semantically meaningful composition of models represented by 3D Gaussian Splatting (3DGS). For texture blending, due to the discrete and irregular nature of 3DGS, straightforwardly applying gradient propagation as SeamlssNeRF is not supported. Thus, a novel sampling-based cloning method is proposed to harmonize the blending while preserving the original rich texture and content. Our workflow consists of three steps: 1) real-time segmentation and transformation of 3DGS using a well-tailored GUI, 2) KNN analysis to identify boundary points in the intersecting area between the source and target models, and 3) two-phase optimization of the target model using sampling-based cloning and gradient constraints. Extensive experimental results validate that our approach significantly outperforms previous works in realistic synthesis, demonstrating its practicality.

Existing methods for capturing multi-person holistic human motions from a monocular video usually involve integrating the detector, the tracker, and the human pose & shape estimator into a cascaded system. Differently, we develop a one-stage multi-person holistic human motion capture system, which 1) employs only one network, enabling significant benefits from the end-to-end training on a large-scale dataset; 2) enables performance improving of the tracking module during training, avoiding being limited by a pre-trained tracker; 3) captures the motions of all individuals within a single shot, rather than tracking and estimating each person sequentially. In this system, each query within a temporal cross-attention module is responsible for the long motion of a specific individual, implicitly aggregating individual-specific information throughout the entire video. To further boost the proposed system from end-to-end training, we also construct a synthetic human video dataset, with multi-person and whole-body annotations. Extensive experiments across different datasets demonstrate both the efficacy and the efficiency of both the proposed method and the dataset. Codes are avaiable at https://github.com/KenkunLiu/MaQ.

Establishing character shape correspondence is a critical and fundamental task in computer vision and graphics, with diverse applications including re-topology, attribute transfer, and shape interpolation. Current dominant functional map methods, while effective in controlled scenarios, struggle in real situations with more complex challenges such as non-isometric shape discrepancies. In response, we revisit registration-for-correspondence methods and tap their potential for more stable shape correspondence estimation. To overcome their common issues including unstable deformations and the necessity for careful pre-alignment or high-quality initial 3D correspondences, we introduce Stable-SCore: A Stable Registration-based Framework for 3D Shape Correspondence. We first re-purpose a foundation model for 2D character correspondence that ensures reliable and stable 2D mappings. Crucially, we propose a novel Semantic Flow Guided Registration approach that leverages 2D correspondence to guide mesh deformations. Our framework significantly surpasses existing methods in challenging scenarios, and brings possibilities for a wide array of real applications, as demonstrated in our results.

This paper tackles the emerging challenge of multi-view virtual try-on, utilizing both front- and back-view clothing images as inputs. Extending frontal try-on methods to a multi-view context is not straightforward. Simply concatenating the two input views or encoding their features for a generative model, such as a diffusion model, often fails to produce satisfactory results. The main challenge lies in effectively extracting and fusing meaningful clothing features from these input views. Existing explicit warping-based methods, which establish direct correspondence between input and target views, tend to introduce artifacts, particularly when there is a significant disparity between the input and target views. Conversely, implicit encoding-based methods often lose spatial information about clothing, resulting in outputs that lack detail. To overcome these challenges, we propose Robust-MVTON, an end-to-end method for robust and high-quality multi-view try-ons. Our approach introduces a novel cross-pose feature alignment technique to guide the fusion of clothing features and incorporates a newly designed loss function for training. With the fused multi-scale clothing features, we employ a coarse-to-fine diffusion model to generate realistic and detailed results. Extensive experiments conducted on the Deepfashion and MPV datasets affirm the superiority of our method, achieving state-of-the-art performance.

We address key limitations in existing datasets and models for task-oriented hand-object interaction video generation, a critical approach of generating video demonstrations for robotic imitation learning. Current datasets, such as Ego4D [16], often suffer from inconsistent view perspectives and misaligned interactions, leading to reduced video quality and limiting their applicability for precise imitation learning tasks. Towards this end, we introduce TASTE-Rob — a pioneering large-scale dataset of 100,856 ego-centric hand-object interaction videos. Each video is meticulously aligned with language instructions and recorded from a consistent camera viewpoint to ensure interaction clarity. By fine-tuning a Video Diffusion Model (VDM) on TASTERob, we achieve realistic object interactions, though we observed occasional inconsistencies in hand grasping postures. To enhance realism, we introduce a three-stage pose-refinement pipeline that improves hand posture accuracy in generated videos. Our curated dataset, coupled with the specialized pose-refinement framework, provides notable performance gains in generating high-quality, task-oriented hand-object interaction videos, resulting in achieving superior generalizable robotic manipulation. To foster further advancements in the field, TASTE-Rob dataset and source code will be made publicly available on our website https://taste-rob.github.io.

3D data simulation aims to bridge the gap between simulated and real-captured 3D data, which is a fundamental problem for real-world 3D visual tasks. Most 3D data simulation methods inject predefined physical priors but struggle to capture the full complexity of real data. An optimal approach involves learning an implicit mapping from synthetic to realistic data in a data-driven manner, but progress in this solution has met stagnation in recent studies. This work explores a new solution path of data-driven 3D simulation, called Stable-Sim2Real, based on a novel two-stage depth diffusion model. The initial stage finetunes Stable-Diffusion to generate the residual between the real and synthetic paired depth, producing a stable but coarse depth, where some local regions may deviate from realistic patterns. To enhance this, both the synthetic and initial output depth are fed into a second-stage diffusion, where diffusion loss is adjusted to prioritize these distinct areas identified by a 3D discriminator. We provide a new benchmark scheme to evaluate 3D data simulation methods. Extensive experiments show that training the network with the 3D simulated data derived from our method significantly enhances performance in real-world 3D visual tasks. Moreover, the evaluation demonstrates the high similarity between our 3D simulated data and real-captured patterns. Project page: https://mutianxu.github.io/stable-sim2real/.

With the growing demand for high-fidelity 3D models from 2D images, existing methods still face significant challenges in accurately reproducing fine-grained geometric details due to limitations in domain gaps and inherent ambiguities in RGB images. To address these issues, we propose Hi3DGen, a novel framework for generating high-fidelity 3D geometry from images via normal bridging. Hi3DGen consists of three key components: (1) an image-to-normal estimator that decouples the low-high frequency image pattern with noise injection and dual-stream training to achieve generalizable, stable, and sharp estimation; (2) a normal-to-geometry learning approach that uses normal-regularized latent diffusion learning to enhance 3D geometry generation fidelity; and (3) a 3D data synthesis pipeline that constructs a high-quality dataset to support training. Extensive experiments demonstrate the effectiveness and superiority of our framework in generating rich geometric details, outperforming state-of-the-art methods in terms of fidelity. Our work provides a new direction for high-fidelity 3D geometry generation from images by leveraging normal maps as an intermediate representation.

Tri-plane-like representations have been widely adopted in 3D-aware GANs for head image synthesis and other 3D object/scene modeling tasks due to their efficiency. However, querying features via Cartesian coordinate projection often leads to feature entanglement, which results in mirroring artifacts. A recent work, SphereHead, attempted to address this issue by introducing spherical tri-planes based on a spherical coordinate system. While it successfully mitigates feature entanglement, SphereHead suffers from uneven mapping between the square feature maps and the spherical planes, leading to inefficient feature map utilization during rendering and difficulties in generating fine image details.Moreover, both tri-plane and spherical tri-plane representations share a subtle yet persistent issue: feature penetration across convolutional channels can cause interference between planes, particularly when one plane dominates the others (see Fig. 1). These challenges collectively prevent tri-plane-based methods from reaching their full potential. In this paper, we systematically analyze these problems for the first time and propose innovative solutions to address them. Specifically, we introduce a novel hybrid-plane (hy-plane for short) representation that combines the strengths of both planar and spherical planes while avoiding their respective drawbacks. We further enhance the spherical plane by replacing the conventional theta-phi warping with a novel near-equal-area warping strategy, which maximizes the effective utilization of the square feature map. In addition, our generator synthesizes a single-channel unified feature map instead of multiple feature maps in separate channels, thereby effectively eliminating feature penetration. With a series of technical improvements, our hy-plane representation enables our method, HyPlaneHead, to achieve state-of-the-art performance in full-head image synthesis.

Recent breakthroughs in video generation, powered by large-scale datasets and diffusion techniques, have shown that video diffusion models can function as implicit 4D novel view synthesizers. Nevertheless, current methods primarily concentrate on redirecting camera trajectory within the front view while struggling to generate 360-degree viewpoint changes. In this paper, we focus on human-centric subdomain and present MV-Performer, an innovative framework for creating synchronized novel view videos from monocular full-body captures. To achieve a 360-degree synthesis, we extensively leverage the MVHumanNet dataset and incorporate an informative condition signal. Specifically, we use the camera-dependent normal maps rendered from oriented partial point clouds, which effectively alleviate the ambiguity between seen and unseen observations. To maintain synchronization in the generated videos, we propose a multi-view human-centric video diffusion model that fuses information from the reference video, partial rendering, and different viewpoints. Additionally, we provide a robust inference procedure for in-the-wild video cases, which greatly mitigates the artifacts induced by imperfect monocular depth estimation. Extensive experiments on three datasets demonstrate our MV-Performer’s state-of-the-art effectiveness and robustness, setting a strong model for human-centric 4D novel view synthesis. Code is available at https://github.com/zyhbili/MV-Performer.

Geometry- and appearance-controlled full-body human image generation is an interesting but challenging task. Existing solutions are either unconditional or dependent on coarse conditions (e.g., pose, text), thus lacking explicit geometry and appearance control of body and garment. Sketching offers such editing ability and has been adopted in various sketch-based face generation and editing solutions. However, directly adapting sketch-based face generation to full-body generation often fails to produce high-fidelity and diverse results due to the high complexity and diversity in the pose, body shape, and garment shape and texture. Recent geometrically controllable diffusion-based methods mainly rely on prompts to generate appearance. It is hard to balance the realism and the faithfulness of their results to the sketch when the input is coarse. This work presents Sketch2Human, the first system for controllable full-body human image generation guided by a semantic sketch (for geometry control) and a reference image (for appearance control). Our solution is based on the latent space of StyleGAN-Human with inverted geometry and appearance latent codes as input. Specifically, we present a sketch encoder trained with a large synthetic dataset sampled from StyleGAN-Human's latent space and directly supervised by sketches rather than real images. Considering the entangled information of partial geometry and texture in StyleGAN-Human and the absence of disentangled datasets, we design a novel training scheme that creates geometry-preserved and appearance-transferred training data to tune a generator to achieve disentangled geometry and appearance control. Although our method is trained with synthetic data, it can also handle hand-drawn sketches. Qualitative and quantitative evaluations demonstrate the superior performance of our method to state-of-the-art methods.

In the field of 3D point cloud generation, numerous 3D generative models have demonstrated the ability to generate diverse and realistic 3D shapes. However, the majority of these approaches struggle to generate controllable 3D point cloud shapes that meet user-specific requirements, hindering the large-scale application of 3D point cloud generation. To address the challenge of lacking control in 3D point cloud generation, we are the first to propose controlling the generation of point clouds by shape structures that comprise part existences and part adjacency relationships. We manually annotate the adjacency relationships between the segmented parts of point cloud shapes, thereby constructing a StructureGraph representation. Based on this StructureGraph representation, we introduce StrucADT, a novel structure-controllable point cloud generation model, which consists of StructureGraphNet module to extract structure-aware latent features, cCNF Prior module to learn the distribution of the latent features controlled by the part adjacency, and Diffusion Transformer module conditioned on the latent features and part adjacency to generate structure-consistent point cloud shapes. Experimental results demonstrate that our structure-controllable 3D point cloud generation method produces high-quality and diverse point cloud shapes, enabling the generation of controllable point clouds based on user-specified shape structures and achieving state-of-the-art performance in controllable point cloud generation on the ShapeNet dataset.

The image-based Person-to-Person (P2P) virtual try-on, involving the direct transfer of garments from one person to another, is one of the most promising applications of human-centric image generation. However, existing approaches struggle to accurately learn the clothing deformation when directly warping the garment from the source pose onto the target pose. To address this, we propose Person-to-Person virtual try-on via Garment UNwrapping, a novel framework dubbed as ViTon-GUN. Specifically, we divide the P2P task into two subtasks: Person-to-Garment (P2G) and Garment-to-Person (G2P). The P2G aims to unwrap the target garment from a source pose to a canonical representation based on A-Pose. In the P2G stage, we enable the implementation of a flow-based P2G scheme by introducing an A-Pose estimator and establishing comprehensive training conditions. Building upon this step-wise strategy, we introduce a novel pipeline for P2P try-on. Once trained, the P2G strategy can serve as a "plug-and-play" module, which efficiently adapts existing diffusion-based pre-trained G2P models to P2P try-on without further training. Quantitative and qualitative experiments demonstrate that our ViTon-GUN performs remarkably well on P2P try-on, even for dresses with intricate design details.

In recent years, significant progress has been witnessed in the field of neural 3D avatar reconstruction. Among all related tasks, building an animatable avatar from monocular videos is one of the most challenging ones, yet it also has a wide range of applications. The "animatable" means that we need to transfer any arbitrary and unseen poses onto the avatar and generate new 3D videos. Thanks to the rise of the powerful representation of NeRF, generating a high-fidelity animatable avatar from videos has become easier and more accessible. Despite their impressive visual results, the substantial training and rendering overhead dramatically hamper their applications. 3D Gaussian Splatting, as a timely new representation, has demonstrated its high-quality and high-efficiency rendering. This has led to many concurrent works to introduce 3D-GS to animatable avatar building. Although they demonstrate very high-fidelity renderings for poses similar to the training video frames, poor results are produced when the poses are far from training. We argue that this is primarily because the Gaussian points lack structures. Thus, we suggest involving DMTet to represent the coarse geometry of the avatar. In our representation, the majority of Gaussian points are bound to the mesh vertices, while some free Gaussian is allowed to expand to better fit the given video. Furthermore, we develop a dual-space optimization framework to jointly optimize the DMTet, Gaussian points, and skinning weights under two spaces. In this sense, Gaussian points are transformed in a constrained way, which dramatically improves the generalization ability for unseen poses. This is well demonstrated via extensive experiments.

Digital subtraction angiography (DSA) is an essential diagnostic tool for analyzing and diagnosing cardiovascular diseases. However, patient movement during image acquisition can introduce motion artifacts in DSA images, and this degradation in image quality always hinders accurate vessel identification and surgical treatment. Recently, some deep learning-based studies have been presented to address the artifact problem in DSA images by leveraging a generative model to produce high-quality DSA images directly from contrast images. Motionless data (paired contrast and artifact-free DSA images) is always required for these methods to train a model in a supervised manner. However, we face a dilemma that motionless DSA data is hard to acquire in clinical practice, most of which contain varying degrees of artifacts. This raises issues of insufficient motionless data and imperfect motion data for training effective deep generative models. To address this problem, we propose a new Artifact-aware DSA image generation method (denoted as AaDSA), which aims to generate high-quality DSA images with decreased artifacts using only motion-induced data. Specifically, a Gradient Field Transformation-based (GFT-based) method is introduced to obtain an artifact mask that identifies the artifact regions in a DSA image with minimal manual labeling costs. We then train an AaDSA model using the artifact mask as guidance, avoiding the adverse effect of artifact regions for model training. In the inference phase, the proposed AaDSA model can automatically generate a DSA-like image with decreased artifacts from a single contrast image without any human intervention. Experimental results on a real head-and-neck DSA dataset demonstrate the superiority of our method compared to state-of-the-art methods and its potential for clinical use.

Instance shape reconstruction from a 3D scene involves recovering the full geometries of multiple objects at the se-mantic instance level. Many methods leverage data-driven learning due to the intricacies of scene complexity and sig-nificant indoor occlusions. Training these methods often requires a large-scale, high-quality dataset with aligned and paired shape annotations with real-world scans. Existing datasets are either synthetic or misaligned, restricting the performance of data-driven methods on real data. To this end, we introduce LASA, a Large-scale Aligned Shape Annotation Dataset comprising 10,412 high-quality CAD annotations aligned with 920 real-world scene scans from ArkitScenes, created manually by professional artists. On this top, we propose a novel Diffusion-based Cross-Modal Shape Reconstruction (DisCo) method. It is empowered by a hybrid feature aggregation design to fuse multi-modal in-puts and recover high-fidelity object geometries (see Fig. 1). Besides, we present an Occupancy-Guided 3D Object De-tection (OccGOD) method and demonstrate that our shape annotations provide scene occupancy clues that can further improve 3D object detection. Supported by LASA, extensive experiments show that our methods achieve state-of-the-art performance in both instance-level scene reconstruction and 3D object detection tasks.

In this paper, we propose a novel virtual try-on from unconstrained designs (ucVTON) task to enable photorealistic synthesis of personalized composite clothing on input human images. Unlike prior arts constrained by specific input types, our method allows flexible specification of style (text or image) and texture (full garment, cropped sections, or texture patches) conditions. To address the entanglement challenge when using full garment images as conditions, we develop a two-stage pipeline with explicit disentanglement of style and texture. In the first stage, we generate a human parsing map reflecting the desired style conditioned on the input. In the second stage, we composite textures onto the parsing map areas based on the texture input. To represent complex and non-stationary textures that have never been achieved in previous fashion editing works, we first propose extracting hierarchical and balanced CLIP features and applying position encoding in VTON. Experiments demonstrate superior synthesis quality and personalization enabled by our method. The flexible control over style and texture mixing brings virtual try-on to a new level of user experience for online shopping and fashion design.

Lifting 2D diffusion for 3D generation is a challenging problem due to the lack of geometric prior and the complex entanglement of materials and lighting in natural images. Existing methods have shown promise by first creating the geometry through score-distillation sampling (SDS) applied to rendered surface normals, followed by appearance modeling. However, relying on a 2D RGB diffusion model to optimize surface normals is suboptimal due to the distribution discrepancy between natural images and normals maps, leading to instability in optimization. In this paper, recognizing that the normal and depth information effectively describe scene geometry and be auto-matically estimated from images, we propose to learn a generalizable Normal-Depth diffusion model for 3D generation. We achieve this by training on the large-scale LAION dataset together with the generalizable image-to-depth and normal prior models. In an attempt to alleviate the mixed illumination effects in the generated materials, we introduce an albedo diffusion model to impose data-driven constraints on the albedo component. Our experiments show that when integrated into existing text-to-3D pipelines, our models significantly enhance the detail richness, achieving state-of-the-art results. Our project page is at https://aigc3d.github.io/richdreamer/.

Generalizable 3D object reconstructionfrom single-view RGB-D images remains a challenging task, particularly with real-world data. Current state-of-the-art methods develop Transformer-based implicit field learning, necessitating an intensive learning paradigm that requires dense query-supervision uniformly sampled throughout the entire space. We propose a novel approach, IPoD, which harmonizes implicit field learning with point diffusion. This approach treats the query points for implicit field learning as a noisy point cloud for iterative denoising, allowing for their dynamic adaptation to the target object shape. Such adaptive query points harness diffusion learning's capability for coarse shape recovery and also enhances the implicit representation's ability to delineate finer details. Besides, an additional self-conditioning mechanism is designed to use implicit predictions as the guidance of diffusion learning, leading to a cooperative system. Experiments conducted on the CO3D-v2 dataset affirm the superiority of IPoD, achieving 7.8% improvement in F-score and 28.6% in Chamfer distance over existing methods. The generalizability of IPoD is also demonstrated on the MVImgNet dataset. Our project page is at https://yushuang-wu.github.io/IPoD.

In this era, the success of large language models and text-to-image models can be attributed to the driving force of large-scale datasets. However, in the realm of 3D vision, while remarkable progress has been made with models trained on large-scale synthetic and real-captured object data like Objaverse and MVImgNet, a similar level of progress has not been observed in the domain of human-centric tasks partially due to the lack of a large-scale human dataset. Existing datasets of high-fidelity 3D human capture continue to be mid-sized due to the significant challenges in acquiring large-scale high-quality 3D human data. To bridge this gap, we present MVHuman-Net, a dataset that comprises multi-view human action sequences of 4,500 human identities. The primary focus of our work is on collecting human data that features a large number of diverse identities and everyday clothing using a multi- view human capture system, which facilitates easily scalable data collection. Our dataset contains 9,000 daily outfits, 60,000 motion sequences and 645 million frames with extensive annotations, including human masks, camera parameters, 2D and 3D keypoints, SMPUSMPLX parameters, and corresponding textual descriptions. To explore the potential of MVHumanNet in various 2D and 3D visual tasks, we conducted pilot studies on view-consistent action recognition, human NeRF reconstruction, text-driven view-unconstrained human image generation, as well as 2D view-unconstrained human image and 3D avatar generation. Extensive experiments demonstrate the performance improvements and effective applications enabled by the scale provided by MVHumanNet. As the current largest-scale 3D human dataset, we hope that the release of MVHu-manNet data with annotations will foster further innovations in the domain of 3D human-centric tasks at scale.

As for human avatar reconstruction, contemporary techniques commonly necessitate the acquisition of costly data and struggle to achieve satisfactory results from a small number of casual images. In this paper, we investigate this task from a few-shot unconstrained photo album. The reconstruction of human avatars from such data sources is challenging because of limited data amount and dynamic articulated poses. For handling dynamic data, we integrate a skinning mechanism with deep marching tetrahedra (DMTet) to form a drivable tetrahedral representation, which drives arbitrary mesh topologies generated by the DMTet for the adaptation of unconstrained images. To effectively mine instructive information from few-shot data, we devise a two-phase optimization method with few-shot reference and few-shot guidance. The former focuses on aligning avatar identity with reference images, while the latter aims to generate plausible appearances for unseen regions. Overall, our framework, called HaveFun, can undertake avatar reconstruction, rendering, and animation. Extensive experiments on our developed benchmarks demonstrate that HaveFun exhibits substantially superior performance in reconstructing the human body and hand.

Dealing with noisy labels (LNL) emerges as a critical challenge when applying deep learning (DL) in practical settings. Previous methodologies primarily concentrated on harnessing model predictions to mitigate the impact of noisy labels. Nevertheless, their efficacy is strongly contingent on the accuracy of model predictions, a factor that cannot be assured in the context of LNL. Our empirical analysis shows that in noisy datasets, the spatial information of latent feature representation combined with original noisy labels is more robust than the methods using model predictions. To mitigate the unreliability introduced by model predictions, we propose a novel Feature Representation method, which utilizes noisy labels of nearest neighbors for label Correction and sample Selection (FRCS). Extensive experiments on various benchmark datasets demonstrate the superiority of FRCS compared with SOTA methods. Our codes are available at https://github.com/tianfangjh/FRCS-Noisy-Labels.

To date, transformer-based frameworks have demonstrated impressive results in single-image super-resolution (SISR). However, under practical lightweight scenarios, the complex interaction of deep image feature extraction and similarity modeling limits the performance of these methods, since they require simultaneous layer-specific optimization of both two tasks. In this work, we introduce a novel Unified Projection Sharing algorithm(UPS) to decouple the feature extraction and similarity modeling, achieving notable performance. To do this, we establish a unified projection space defined by a learnable projection matrix, for similarity calculation across all self-attention layers. As a result, deep image feature extraction remains a per-layer optimization manner, while similarity modeling is carried out by projecting these image features onto the shared projection space. Extensive experiments demonstrate that our proposed UPS achieves state-of-the-art performance relative to leading lightweight SISR methods, as verified by various popular benchmarks. Moreover, our unified optimized projection space exhibits encouraging robustness performance for unseen data (degraded and depth images). Finally, UPS also demonstrates promising results across various image restoration tasks, including real-world and classic SISR, image denoising, and image deblocking.

In this paper, we introduce a realistic and challenging domain adaptation problem called Universal Semi-supervised Model Adaptation (USMA), which i) requires only a pre-trained source model, ii) allows the source and target domain to have different label sets, i.e., they share a common label set and hold their own private label set, and iii) requires only a few labeled samples in each class of the target domain. To address USMA, we propose a collaborative consistency training framework that regularizes the prediction consistency between two models, i.e., a pre-trained source model and its variant pre-trained with target data only, and combines their complementary strengths to learn a more powerful model. The rationale of our framework stems from the observation that the source model performs better on common categories than the target-only model, while on target-private categories, the target-only model performs better. We also propose a two-perspective, i.e., sample-wise and class-wise, consistency regularization to improve the training. Experimental results demonstrate the effectiveness of our method on several benchmark datasets.

Graph Neural Networks (GNNs) have gained momentum in graph representation learning and boosted the state of the art in a variety of areas, such as data mining (e.g., social network analysis and recommender systems), computer vision (e.g., object detection and point cloud learning), and natural language processing (e.g., relation extraction and sequence learning), to name a few. With the emergence of Transformers in natural language processing and computer vision, graph Transformers embed a graph structure into the Transformer architecture to overcome the limitations of local neighborhood aggregation while avoiding strict structural inductive biases. In this paper, we present a comprehensive review of GNNs and graph Transformers in computer vision from a task-oriented perspective. Specifically, we divide their applications in computer vision into five categories according to the modality of input data, i.e., 2D natural images, videos, 3D data, vision + language, and medical images. In each category, we further divide the applications according to a set of vision tasks. Such a task-oriented taxonomy allows us to examine how each task is tackled by different GNN-based approaches and how well these approaches perform. Based on the necessary preliminaries, we provide the definitions and challenges of the tasks, in-depth coverage of the representative approaches, as well as discussions regarding insights, limitations, and future directions.

Reconstructing a 3D shape based on a single sketch image is challenging due to the inherent sparsity and ambiguity present in sketches. Existing methods lose fine details when extracting features to predict 3D objects from sketches. Upon analyzing the 3D-to-2D projection process, we observe that the density map, characterizing the distribution of 2D point clouds, can serve as a proxy to facilitate the reconstruction process. In this work, we propose a novel sketch-based 3D reconstruction model named SketchSampler. It initiates the process by translating a sketch through an image translation network into a more informative 2D representation, which is then used to generate a density map. Subsequently, a two-stage probabilistic sampling process is employed to reconstruct a 3D point cloud: first, recovering the 2D points (i.e., the x and y coordinates) by sampling the density map; and second, predicting the depth (i.e., the z coordinate) by sampling the depth values along the ray determined by each 2D point. Additionally, we convert the reconstructed point cloud into a 3D mesh for wider applications. To reduce ambiguity, we incorporate hidden lines in sketches. Experimental results demonstrate that our proposed approach significantly outperforms other baseline methods.

Neural radiance fields (NeRF) have shown great success in novel view synthesis. However, recovering high-quality details from real-world scenes is still challenging for the existing NeRF-based approaches, due to the potential imperfect calibration information and scene representation inaccuracy. Even with high-quality training frames, the synthetic novel views produced by NeRF models still suffer from notable rendering artifacts, such as noise and blur. To address this, we propose NeRFLiX, a general NeRF-agnostic restorer paradigm that learns a degradation-driven inter-viewpoint mixer. Specially, we design a NeRF-style degradation modeling approach and construct large-scale training data, enabling the possibility of effectively removing NeRF-native rendering artifacts for deep neural networks. Moreover, beyond the degradation removal, we propose an inter-viewpoint aggregation framework that fuses highly related high-quality training images, pushing the performance of cutting-edge NeRF models to entirely new levels and producing highly photo-realistic synthetic views. Based on this paradigm, we further present NeRFLiX++ with a stronger two-stage NeRF degradation simulator and a faster inter-viewpoint mixer, achieving superior performance with significantly improved computational efficiency. Notably, NeRFLiX++ is capable of restoring photo-realistic ultra-high-resolution outputs from noisy low-resolution NeRF-rendered views. Extensive experiments demonstrate the excellent restoration ability of NeRFLiX++ on various novel view synthesis benchmarks.

In this paper, we address a complex but practical scenario in Active Learning (AL) known as open-set AL, where the unlabeled data consists of both in-distribution (ID) and out-of-distribution (OOD) samples. Standard AL methods will fail in this scenario as OOD samples are highly likely to be regarded as uncertain samples, leading to their selection and wasting of the budget. Existing methods focus on selecting the highly likely ID samples, which tend to be easy and less informative. To this end, we introduce two criteria, namely contrastive confidence and historical divergence, which measure the possibility of being ID and the hardness of a sample, respectively. By balancing the two proposed criteria, highly informative ID samples can be selected as much as possible. Furthermore, unlike previous methods that require additional neural networks to detect the OOD samples, we propose a contrastive clustering framework that endows the classifier with the ability to identify the OOD samples and further enhances the network's representation learning. The experimental results demonstrate that the proposed method achieves state-of-the-art performance on several benchmark datasets.

Neural implicit functions have brought impressive advances to the state-of-the-art of clothed human digitization from multiple or even single images. However, despite the progress, current arts still have difficulty generalizing to unseen images with complex cloth deformation and body poses. In this work, we present GarVerseLOD, a new dataset and framework that paves the way to achieving unprecedented robustness in high-fidelity 3D garment reconstruction from a single unconstrained image. Inspired by the recent success of large generative models, we believe that one key to addressing the generalization challenge lies in the quantity and quality of 3D garment data. Towards this end, GarVerseLOD collects 6,000 high-quality cloth models with fine-grained geometry details manually created by professional artists. In addition to the scale of training data, we observe that having disentangled granularities of geometry can play an important role in boosting the generalization capability and inference accuracy of the learned model. We hence craft GarVerseLOD as a hierarchical dataset with levels of details (LOD) , spanning from detail-free stylized shape to pose-blended garment with pixel-aligned details. This allows us to make this highly under-constrained problem tractable by factorizing the inference into easier tasks, each narrowed down with smaller searching space. To ensure GarVerseLOD can generalize well to in-the-wild images, we propose a novel labeling paradigm based on conditional diffusion models to generate extensive paired images for each garment model with high photorealism. We evaluate our method on a massive amount of in-the-wild images. Experimental results demonstrate that GarVerseLOD can generate standalone garment pieces with significantly better quality than prior approaches while being robust against a large variation of pose, illumination, occlusion, and deformation. Code and dataset are available at garverselod.github.io.

MVImgNet is a large-scale dataset that contains multi-view images of ~220k real-world objects in 238 classes. As a counterpart of ImageNet, it introduces 3D visual signals via multi-view shooting, making a soft bridge between 2D and 3D vision. This paper constructs the MVImgNet2.0 dataset that expands MVImgNet into a total of ~520k objects and 515 categories, which derives a 3D dataset with a larger scale that is more comparable to ones in the 2D domain. In addition to the expanded dataset scale and category range, MVImgNet2.0 is of a higher quality than MVImgNet owing to four new features: (i) most shoots capture 360° views of the objects, which can support the learning of object reconstruction with completeness; (ii) the segmentation manner is advanced to produce foreground object masks of higher accuracy; (iii) a more powerful structure-from-motion method is adopted to derive the camera pose for each frame of a lower estimation error; (iv) higher-quality dense point clouds are reconstructed via advanced methods for objects captured in 360 ° views, which can serve for downstream applications. Extensive experiments confirm the value of the proposed MVImgNet2.0 in boosting the performance of large 3D reconstruction models. MVImgNet2.0 will be public at luyues.github.io/mvimgnet2 , including multi-view images of all 520k objects, the reconstructed high-quality point clouds, and data annotation codes, hoping to inspire the broader vision community.

This work addresses the challenge of high-quality surface normal estimation from monocular colored inputs (i.e., images and videos), a field which has recently been revolutionized by repurposing diffusion priors. However, previous attempts still struggle with stochastic inference, conflicting with the deterministic nature of the Image2Normal task, and costly ensembling step, which slows down the estimation process. Our method, StableNormal, mitigates the stochasticity of the diffusion process by reducing inference variance, thus producing "Stable-and-Sharp" normal estimates without any additional ensembling process. StableNormal works robustly under challenging imaging conditions, such as extreme lighting, blurring, and low quality. It is also robust against transparent and reflective surfaces, as well as cluttered scenes with numerous objects. Specifically, StableNormal employs a coarse-to-fine strategy, which starts with a one-step normal estimator (YOSO) to derive an initial normal guess, that is relatively coarse but reliable, then followed by a semantic-guided refinement process (SG-DRN) that refines the normals to recover geometric details. The effectiveness of StableNormal is demonstrated through competitive performance in standard datasets such as DIODE-indoor, iBims, ScannetV2 and NYUv2, and also in various downstream tasks, such as surface reconstruction and normal enhancement. These results evidence that StableNormal retains both the "stability" and "sharpness" for accurate normal estimation. StableNormal represents a baby attempt to repurpose diffusion priors for deterministic estimation. To democratize this, code and models have been publicly available in hf.co/Stable-X.

Sampling, grouping, and aggregation are three important components in the multi-scale analysis of point clouds. In this paper, we present a novel data-driven sampler learning strategy for point-wise analysis tasks. Unlike the widely used sampling technique, Farthest Point Sampling (FPS), we propose to learn sampling and downstream applications jointly. Our key insight is that uniform sampling methods like FPS are not always optimal for different tasks: sampling more points around boundary areas can make the point-wise classification easier for segmentation. Towards this end, we propose a novel sampler learning strategy that learns sampling point displacement supervised by task-related ground truth information and can be trained jointly with the underlying tasks. We further demonstrate our methods in various point-wise analysis tasks, including semantic part segmentation, point cloud completion, and keypoint detection. Our experiments show that jointly learning of the sampler and task brings better performance than using FPS in various point-based networks.

Modeling 3D avatars benefits various application scenarios such as AR/VR, gaming, and filming. Character faces contribute significant diversity and vividity as a vital component of avatars. However, building 3D character face models usually requires a heavy workload with commercial tools, even for experienced artists. Various existing sketch-based tools fail to support amateurs in modeling diverse facial shapes and rich geometric details. In this article, we present SketchMetaFace - a sketching system targeting amateur users to model high-fidelity 3D faces in minutes. We carefully design both the user interface and the underlying algorithm. First, curvature-aware strokes are adopted to better support the controllability of carving facial details. Second, considering the key problem of mapping a 2D sketch map to a 3D model, we develop a novel learning-based method termed "Implicit and Depth Guided Mesh Modeling" (IDGMM). It fuses the advantages of mesh, implicit, and depth representations to achieve high-quality results with high efficiency. In addition, to further support usability, we present a coarse-to-fine 2D sketching interface design and a data-driven stroke suggestion tool. User studies demonstrate the superiority of our system over existing modeling tools in terms of the ease to use and visual quality of results. Experimental analyses also show that IDGMM reaches a better trade-off between accuracy and efficiency.

Large-scale datasets and deep generative models have enabled impressive progress in human face reenactment. Existing solutions for face reenactment have focused on processing real face images through facial landmarks by generative models. Different from real human faces, artistic human faces (e.g., those in paintings, cartoons, etc.) often involve exaggerated shapes and various textures. Therefore, directly applying existing solutions to artistic faces often fails to preserve the characteristics of the original artistic faces (e.g., face identity and decorative lines along face contours) due to the domain gap between real and artistic faces. To address these issues, we present ReenactArtFace, the first effective solution for transferring the poses and expressions from human videos to various artistic face images. We achieve artistic face reenactment in a coarse-to-fine manner. First, we perform 3D artistic face reconstruction, which reconstructs a textured 3D artistic face through a 3D morphable model (3DMM) and a 2D parsing map from an input artistic image. The 3DMM can not only rig the expressions better than facial landmarks but also render images under different poses/expressions as coarse reenactment results robustly. However, these coarse results suffer from self-occlusions and lack contour lines. Second, we thus perform artistic face refinement by using a personalized conditional adversarial generative model (cGAN) fine-tuned on the input artistic image and the coarse reenactment results. For high-quality refinement, we propose a contour loss to supervise the cGAN to faithfully synthesize contour lines. Quantitative and qualitative experiments demonstrate that our method achieves better results than the existing solutions.

The mainstream of the existing approaches for video prediction builds up their models based on a Single-In-Single-Out (SISO) architecture, which takes the current frame as input to predict the next frame in a recursive manner. This way often leads to severe performance degradation when they try to extrapolate a longer period of future, thus limiting the practical use of the prediction model. Alternatively, a Multi-In-Multi-Out (MIMO) architecture that outputs all the future frames at one shot naturally breaks the recursive manner and therefore prevents error accumulation. However, only a few MIMO models for video prediction are proposed and they only achieve inferior performance due to the date. The real strength of the MIMO model in this area is not well noticed and is largely under-explored. Motivated by that, we conduct a comprehensive investigation in this paper to thoroughly exploit how far a simple MIMO architecture can go. Surprisingly, our empirical studies reveal that a simple MIMO model can outperform the state-of-the-art work with a large margin much more than expected, especially in dealing with long-term error accumulation. After exploring a number of ways and designs, we propose a new MIMO architecture based on extending the pure Transformer with local spatio-temporal blocks and a new multi-output decoder, namely MIMO-VP, to establish a new standard in video prediction. We evaluate our model in four highly competitive benchmarks. Extensive experiments show that our model wins 1st place on all the benchmarks with remarkable performance gains and surpasses the best SISO model in all aspects including efficiency, quantity, and quality. A dramatic error reduction is achieved when predicting 10 frames on Moving MNIST and Weather datasets respectively. We believe our model can serve as a new baseline to facilitate the future research of video prediction tasks. The code will be released.

Assisting people in efficiently producing visually plausible 3D characters has always been a fundamental research topic in computer vision and computer graphics. Recent learning-based approaches have achieved unprecedented accuracy and efficiency in the area of 3D real human digitization. However, none of the prior works focus on modeling 3D biped cartoon characters, which are also in great demand in gaming and filming. In this paper, we introduce 3DBiCar, the first large-scale dataset of 3D biped cartoon characters, and RaBit, the corresponding parametric model. Our dataset contains 1,500 topologically consistent high-quality 3D textured models which are manually crafted by professional artists. Built upon the data, RaBit is thus designed with a SMPL-like linear blend shape model and a StyleGAN-based neural UV-texture generator, simultaneously expressing the shape, pose, and texture. To demonstrate the practicality of 3DBiCar and RaBit, various applications are conducted, including single-view reconstruction, sketch-based modeling, and 3D cartoon animation. For the single-view reconstruction setting, we find a straightforward global mapping from input images to the output UV-based texture maps tends to lose detailed appearances of some local parts (e.g., nose, ears). Thus, a part-sensitive texture reasoner is adopted to make all important local areas perceived. Experiments further demonstrate the effectiveness of our method both qualitatively and quantitatively. 3DBiCar and RaBit are available at gaplab.cuhk.edu.cn/projects/RaBit.

Holistic 3D scene understanding entails estimation of both layout configuration and object geometry in a 3D environment. Recent works have shown advances in 3D scene estimation from various input modalities (e.g., images, 3D scans), by leveraging 3D supervision (e.g., 3D bounding boxes or CAD models), for which collection at scale is ex-pensive and often intractable. To address this shortcoming, we propose a new method to learn 3D scene priors of layout and shape without requiring any 3D ground truth. Instead, we rely on 2D supervision from multi-view RGB images. Our method represents a 3D scene as a latent vector, from which we can progressively decode to a sequence of objects characterized by their class categories, 3D bounding boxes, and meshes. With our trained autoregressive decoder representing the scene prior, our method facilitates many downstream applications, including scene synthesis, interpolation, and single-view reconstruction. Experiments on 3D-FRONT and ScanNet show that our method outperforms state of the art in single-view reconstruction, and achieves state-of-the-art results in scene synthesis against baselines which require for 3D supervision. Project page: https://yinyunie.github.io/sceneprior-page/

Reconstructing dynamic 3D garment surfaces with open boundaries from monocular videos is an important problem as it provides a practical and low-cost solution for clothes digitization. Recent neural rendering methods achieve high-quality dynamic clothed human reconstruction results from monocular video, but these methods cannot separate the garment surface from the body. Moreover, despite existing garment reconstruction methods based on feature curve representation demonstrating impressive results for garment reconstruction from a single image, they struggle to generate temporally consistent surfaces for the video input. To address the above limitations, in this paper, we formulate this task as an optimization problem of 3D garment feature curves and surface reconstruction from monocular video. We introduce a novel approach, called REC-MV, to jointly optimize the explicit feature curves and the implicit signed distance field (SDF) of the garments. Then the open garment meshes can be extracted via garment template registration in the canonical space. Experiments on multiple casually captured datasets show that our approach outperforms existing methods and can produce high-quality dynamic garment surfaces. The source code is available at https://github.com/GAP-LAB-CUHK-SZ/REC-MV.

3D shape completion from point clouds is a challenging task, especially from scans of real-world objects. Considering the paucity of 3D shape ground truths for real scans, existing works mainly focus on benchmarking this task on synthetic data, e.g. 3D computer-aided design models. However, the domain gap between synthetic and real data limits the generalizability of these methods. Thus, we propose a new task, SCoDA, for the domain adaptation of real scan shape completion from synthetic data. A new dataset, ScanSalon, is contributed with a bunch of elaborate 3D models created by skillful artists according to scans. To address this new task, we propose a novel cross-domain feature fusion method for knowledge transfer and a novel volume-consistent self-training framework for robust learning from real data. Extensive experiments prove our method is effective to bring an improvement of 6%~7% mIoU.

Masked Modeling (MM) has demonstrated widespread success in various vision challenges, by reconstructing masked visual patches. Yet, applying MM for large-scale 3D scenes remains an open problem due to the data sparsity and scene complexity. The conventional random masking paradigm used in 2D images often causes a high risk of ambiguity when recovering the masked region of 3D scenes. To this end, we propose a novel informative-preserved reconstruction, which explores local statistics to discover and preserve the representative structured points, effectively enhancing the pretext masking task for 3D scene understanding. Integrated with a progressive reconstruction manner, our method can concentrate on modeling regional geometry and enjoy less ambiguity for masked reconstruction. Besides, such scenes with progressive masking ratios can also serve to self-distill their intrinsic spatial consistency, requiring to learn the consistent representations from unmasked areas. By elegantly combining informative-preserved reconstruction on masked areas and consistency self-distillation from unmasked areas, a unified framework called MM-3DScene is yielded. We conduct comprehensive experiments on a host of downstream tasks. The consistent improvement (e.g., +6.1% [email protected] on object detection and +2.2% mIoU on semantic segmentation) demonstrates the superiority of our approach.

Being data-driven is one of the most iconic properties of deep learning algorithms. The birth of ImageNet [24] drives a remarkable trend of ‘learning from large-scale data’ in computer vision. Pretraining on ImageNet to obtain rich universal representations has been manifested to benefit various 2D visual tasks, and becomes a standard in 2D vision. However, due to the laborious collection of real-world 3D data, there is yet no generic dataset serving as a counterpart of ImageNet in 3D vision, thus how such a dataset can impact the 3D community is unraveled. To remedy this defect, we introduce MVImgNet, a large-scale dataset of multi-view images, which is highly convenient to gain by shooting videos of real-world objects in human daily life. It contains 6.5 million frames from 219,188 videos crossing objects from 238 classes, with rich annotations of object masks, camera parameters, and point clouds. The multi-view attribute endows our dataset with 3D-aware signals, making it a soft bridge between 2D and 3D vision. We conduct pilot studies for probing the potential of MVImgNet on a variety of 3D and 2D visual tasks, including radiance field reconstruction, multi-view stereo, and view-consistent image understanding, where MVImgNet demonstrates promising performance, remaining lots of possibilities for future explorations. Besides, via dense reconstruction on MVImgNet, a 3D object point cloud dataset is derived, called MVPNet, covering 87,200 samples from 150 categories, with the class label on each point cloud. Experiments show that MVP-Net can benefit the real-world 3D object classification while posing new challenges to point cloud understanding. MVImgNet and MVPNet will be public, hoping to inspire the broader vision community.

In this work, we tackle the challenging problem of learning-based single-view 3D hair modeling. Due to the great difficulty of collecting paired real image and 3D hair data, using synthetic data to provide prior knowledge for real domain becomes a leading solution. This unfortunately introduces the challenge of domain gap. Due to the inherent difficulty of realistic hair rendering, existing methods typically use orientation maps instead of hair images as input to bridge the gap. We firmly think an intermediate representation is essential, but we argue that orientation map using the dominant filtering-based methods is sensitive to uncertain noise and far from a competent representation. Thus, we first raise this issue up and propose a novel intermediate representation, termed as HairStep, which consists of a strand map and a depth map. It is found that HairStep not only provides sufficient information for accurate 3D hair modeling, but also is feasible to be inferred from real images. Specifically, we collect a dataset of 1,250 portrait images with two types of annotations. A learning framework is further designed to transfer real images to the strand map and depth map. It is noted that, an extra bonus of our new dataset is the first quantitative metric for 3D hair modeling. Our experiments show that HairStep narrows the domain gap between synthetic and real and achieves state-of-the-art performance on single-view 3D hair reconstruction.

For video frame interpolation (VFI), existing deep-learning-based approaches strongly rely on the ground-truth (GT) intermediate frames, which sometimes ignore the non-unique nature of motion judging from the given adjacent frames. As a result, these methods tend to produce averaged solutions that are not clear enough. To alleviate this issue, we propose to relax the requirement of reconstructing an intermediate frame as close to the GT as possible. Towards this end, we develop a texture consistency loss (TCL) upon the assumption that the interpolated content should maintain similar structures with their counterparts in the given frames. Predictions satisfying this constraint are encouraged, though they may differ from the predefined GT. Without the bells and whistles, our plug-and-play TCL is capable of improving the performance of existing VFI frameworks consistently. On the other hand, previous methods usually adopt the cost volume or correlation map to achieve more accurate image or feature warping. However, the O (N2) (N refers to the pixel count) computational complexity makes it infeasible for high-resolution cases. In this work, we design a simple, efficient O (N) yet powerful guided cross-scale pyramid alignment (GCSPA) module, where multi-scale information is highly exploited. Extensive experiments justify the efficiency and effectiveness of the proposed strategy.

Neural radiance fields (NeRF) show great success in novel view synthesis. However, in real-world scenes, recovering high-quality details from the source images is still challenging for the existing NeRF-based approaches, due to the potential imperfect calibration information and scene representation inaccuracy. Even with high-quality training frames, the synthetic novel views produced by NeRF models still suffer from notable rendering artifacts, such as noise, blur, etc. Towards to improve the synthesis quality of NeRF-based approaches, we propose NeRFLiX, a general NeRF-agnostic restorer paradigm by learning a degradation-driven inter-viewpoint mixer. Specially, we design a NeRF-style degradation modeling approach and construct large-scale training data, enabling the possibility of effectively removing NeRF-native rendering artifacts for existing deep neural networks. Moreover, beyond the degradation removal, we propose an inter-viewpoint aggregation framework that is able to fuse highly related high-quality training images, pushing the performance of cutting-edge NeRF models to entirely new levels and producing highly photo-realistic synthetic views.

Extracting parametric edge curves from point clouds is a fundamental problem in 3D vision and geometry processing. Existing approaches mainly rely on keypoint detection, a challenging procedure that tends to generate noisy out-put, making the subsequent edge extraction error-prone. To address this issue, we propose to directly detect structured edges to circumvent the limitations of the previous point-wise methods. We achieve this goal by presenting NerVE, a novel neural volumetric edge representation that can be easily learned through a volumetric learning framework. NerVE can be seamlessly converted to a versatile piece-wise lin-ear (PWL) curve representation, enabling a unified strategy for learning all types offree-form curves. Furthermore, as NerVE encodes rich structural information, we show that edge extraction based on NerVE can be reduced to a simple graph search problem. After converting NerVE to the PWL representation, parametric curves can be obtained via off-the-shelf spline fitting algorithms. We evaluate our method on the challenging ABC dataset [19]. We show that a sim-ple network based on NerVE can already outperform the previous state-of-the-art methods by a great margin.

Albeit the notable performance on in-domain test points, it is non-trivial for deep neural networks to attain satisfactory accuracy when deploying in the open world, where novel domains and object classes often occur. In this paper, we study a practical problem of Domain Generalization under Category Shift (DGCS), which aims to simultaneously detect unknown-class samples and classify known-class samples in the target domains. Compared to prior DG works, we face two new challenges: 1) how to learn the concept of "unknown " during training with only source known- class samples, and 2) how to adapt the source-trained model to unseen environments for safe model deployment. To this end, we propose a novel Activate and Reject (ART) framework to reshape the model’s decision boundary to accommodate unknown classes and conduct post hoc modification to further discriminate known and unknown classes using unlabeled test data. Specifically, during training, we promote the response to the unknown by optimizing the unknown probability and then smoothing the overall output to mitigate the overconfidence issue. At test time, we introduce a step-wise online adaptation method that predicts the label by virtue of the cross-domain nearest neighbor and class prototype information without updating the network’s parameters or using threshold-based mechanisms. Experiments reveal that ART consistently improves the generalization capability of deep networks on different vision tasks. For image classification, ART improves the H-score by 6.1% on average compared to the previous best method. For object detection and semantic segmentation, we establish new benchmarks and achieve competitive performance.

Recent work on Neural Radiance Fields (NeRF) has demonstrated significant advances in high-quality view synthesis. A major limitation of NeRF is its low rendering efficiency due to the need for multiple network forwardings to render a single pixel. Existing methods to improve NeRF either reduce the number of required samples or optimize the implementation to accelerate the network forwarding. Despite these efforts, the problem of multiple sampling persists due to the intrinsic representation of radiance fields. In contrast, Neural Light Fields (NeLF) reduce the computation cost of NeRF by querying only one single network forwarding per pixel. To achieve a close visual quality to NeRF, existing NeLF methods require significantly larger network capacities which limits their rendering efficiency in practice. In this work, we propose a new representation called Neural Radiance Distribution Field (NeRDF) that targets efficient view synthesis in real-time. Specifically, we use a small network similar to NeRF while preserving the rendering speed with a single network forwarding per pixel as in NeLF. The key is to model the radiance distribution along each ray with frequency basis and predict frequency weights using the network. Pixel values are then computed via volume rendering on radiance distributions. Experiments show that our proposed method offers a better trade-off among speed, quality, and network size than existing methods: we achieve a ∼254× speed-up over NeRF with similar network size, with only a marginal performance decline. Our project page is at yushuang-wu.github.io/NeRDF.

Fast generation of high-quality 3D digital humans is important to a vast number of applications ranging from entertainment to professional concerns. Recent advances in differentiable rendering have enabled the training of 3D generative models without requiring 3D ground truths. However, the quality of the generated 3D humans still has much room to improve in terms of both fidelity and diversity. In this paper, we present Get3DHuman, a novel 3D human framework that can significantly boost the realism and diversity of the generated outcomes by only using a limited budget of 3D ground-truth data. Our key observation is that the 3D generator can profit from human-related priors learned through 2D human generators and 3D reconstructors. Specifically, we bridge the latent space of Get3DHuman with that of StyleGAN-Human [13] via a specially-designed prior network, where the input latent code is mapped to the shape and texture feature volumes spanned by the pixel-aligned 3D reconstructor [50]. The outcomes of the prior network are then leveraged as the supervisory signals for the main generator network. To ensure effective training, we further propose three tailored losses applied to the generated feature volumes and the intermediate feature maps. Extensive experiments demonstrate that Get3DHuman greatly outperforms the other state-of-the-art approaches and can support a wide range of applications including shape interpolation, shape re-texturing, and single-view reconstruction through latent inversion.

Radiance fields have gradually become a main representation of media. Although its appearance editing has been studied, how to achieve view-consistent recoloring in an efficient manner is still under explored. We present RecolorNeRF, a novel user-friendly color editing approach for the neural radiance fields. Our key idea is to decompose the scene into a set of pure-colored layers, forming a palette. By this means, color manipulation can be conducted by altering the color components of the palette directly. To support efficient palette-based editing, the color of each layer needs to be as representative as possible. In the end, the problem is formulated as an optimization problem, where the layers and their blending weights are jointly optimized with the NeRF itself. Extensive experiments show that our jointly-optimized layer decomposition can be used against multiple backbones and produce photo-realistic recolored novel-view renderings. We demonstrate that RecolorNeRF outperforms baseline methods both quantitatively and qualitatively for color editing even in complex real-world scenes.

The generalization capability of machine learning systems degenerates notably when the test distribution drifts from the training distribution. Recently, Domain Generalization (DG) has been gaining momentum in enabling machine learning models to generalize to unseen domains. However, most DG methods assume that training and test data share an identical label space, ignoring the potential unseen categories in many real-world applications. In this paper, we delve into a more general but difficult problem termed Open Test-Time DG (OTDG), where both domain shift and open class may occur on the unseen test data. We propose Compaction and Disambiguation (CODA), a novel two-stage framework for learning compact representations and adapting to open classes in the wild. To meaningfully regularize the model's decision boundary, CODA introduces virtual unknown classes and optimizes a new training objective to insert unknowns into the latent space by compacting the embedding space of source known classes. To adapt target samples to the source model, we then disambiguate the decision boundaries between known and unknown classes with a test-time training objective, mitigating the adaptivity gap and catastrophic forgetting challenges. Experiments reveal that CODA can significantly outperform the previous best method on standard DG datasets and harmonize the classification accuracy between known and unknown classes.

The past two years have witnessed a significant increase in interest concerning NeRF-based human body rendering. While this surge has propelled considerable advancements, it has also led to an influx of methods and datasets. This explosion complicates experimental settings and makes fair comparisons challenging. In this work, we design and execute thorough studies into unified evaluation settings and metrics to establish a fair and reasonable benchmark for human NeRF models. To reveal the effects of extant models, we benchmark them against diverse and hard scenes. Additionally, we construct a cross-subject benchmark pre-trained on large-scale datasets to assess generalizable methods. Finally, we analyze the essential components for animatability and generalizability, and make HumanNeRF from monocular videos generalizable, as the inaugural baseline. We hope these benchmarks and analyses could serve the community.

Neural Radiance Fields (NeRFs) have emerged as promising digital mediums of 3D objects and scenes, sparking a surge in research to extend the editing capabilities in this domain. The task of seamless editing and merging of multiple NeRFs, resembling the “Poisson blending” in 2D image editing, remains a critical operation that is under-explored by existing work. To fill this gap, we propose SeamlessNeRF, a novel approach for seamless appearance blending of multiple NeRFs. In specific, we aim to optimize the appearance of a target radiance field in order to harmonize its merge with a source field. We propose a well-tailored optimization procedure for blending, which is constrained by 1) pinning the radiance color in the intersecting boundary area between the source and target fields and 2) maintaining the original gradient of the target. Extensive experiments validate that our approach can effectively propagate the source appearance from the boundary area to the entire target field through the gradients. To the best of our knowledge, SeamlessNeRF is the first work that introduces gradient-guided appearance editing to radiance fields, offering solutions for seamless stitching of 3D objects represented in NeRFs. Our code and more results are available at https://sites.google.com/view/seamlessnerf.

Semantic segmentation of point cloud usually relies on dense annotation that is exhausting and costly, so it attracts wide attention to investigate solutions for the weakly supervised scheme with only sparse points annotated. Existing works start from the given labels and propagate them to highly-related but unlabeled points, with the guidance of data, e.g. intra-point relation. However, it suffers from (i) the inefficient exploitation of data information, and (ii) the strong reliance on labels thus is easily suppressed when given much fewer annotations. Therefore, we propose a novel framework, PointMatch, that stands on both data and label, by applying consistency regularization to sufficiently probe information from data itself and leveraging weak labels as assistance at the same time. By doing so, meaningful information can be learned from both data and label for better representation learning , which also enables the model more robust to the extent of label sparsity . Simple yet effective, the proposed PointMatch achieves the state-of-the-art performance under various weakly-supervised schemes on both ScanNet-v2 and S3DIS datasets, especially on the settings with extremely sparse labels, e.g. surpassing SQN by 21.2% and 17.2% on the 0.01% and 0.1% setting of ScanNet-v2, respectively.

Abstract Sketch‐based shape modeling aims to bridge the gap between 2D drawing and 3D modeling by providing an intuitive and accessible approach to create 3D shapes from 2D sketches. However, existing methods still suffer from limitations in reconstruction quality and multi‐view interaction friendliness, hindering their practical application. This paper proposes a faithful and user‐friendly iterative solution to tackle these limitations by learning geometry‐aligned deep implicit functions from one or multiple sketches. Our method lifts 2D sketches to volume‐based feature tensors, which align strongly with the output 3D shape, enabling accurate reconstruction and faithful editing. Such a geometry‐aligned feature encoding technique is well‐suited to iterative modeling since features from different viewpoints can be easily memorized or aggregated. Based on these advantages, we design a unified interactive system for sketch‐based shape modeling. It enables users to generate the desired geometry iteratively by drawing sketches from any number of viewpoints. In addition, it allows users to edit the generated surface by making a few local modifications. We demonstrate the effectiveness and practicality of our method with extensive experiments and user studies, where we found that our method outperformed existing methods in terms of accuracy, efficiency, and user satisfaction. The source code of this project is available at https://github.com/LordLiang/GA‐Sketching .

Abstract In this paper, we tackle the challenging problem of 3D keypoint estimation of general objects using a novel implicit representation. Previous works have demonstrated promising results for keypoint prediction through direct coordinate regression or heatmap‐based inference. However, these methods are commonly studied for specific subjects, such as human bodies and faces, which possess fixed keypoint structures. They also suffer in several practical scenarios where explicit or complete geometry is not given, including images and partial point clouds. Inspired by the recent success of advanced implicit representation in reconstruction tasks, we explore the idea of using an implicit field to represent keypoints. Specifically, our key idea is employing spheres to represent 3D keypoints, thereby enabling the learnability of the corresponding signed distance field. Explicit key‐points can be extracted subsequently by our algorithm based on the Hough transform. Quantitative and qualitative evaluations also show the superiority of our representation in terms of prediction accuracy.

Domain Adaptive Object Detection (DAOD) focuses on improving the generalization ability of object detectors via knowledge transfer. Recent advances in DAOD strive to change the emphasis of the adaptation process from global to local in virtue of fine-grained feature alignment methods. However, both the global and local alignment approaches fail to capture the topological relations among different foreground objects as the explicit dependencies and interactions between and within domains are neglected. In this case, only seeking one-vs-one alignment does not necessarily ensure the precise knowledge transfer. Moreover, conventional alignment-based approaches may be vulnerable to catastrophic overfitting regarding those less transferable regions (e.g., backgrounds) due to the accumulation of inaccurate localization results in the target domain. To remedy these issues, we first formulate DAOD as an open-set domain adaptation problem, in which the foregrounds and backgrounds are seen as the "known classes" and "unknown class" respectively. Accordingly, we propose a new and general framework for DAOD, named Foreground-aware Graph-based Relational Reasoning (FGRR), which incorporates graph structures into the detection pipeline to explicitly model the intra- and inter-domain foreground object relations on both pixel and semantic spaces, thereby endowing the DAOD model with the capability of relational reasoning beyond the popular alignment-based paradigm. FGRR first identifies the foreground pixels and regions by searching reliable correspondence and cross-domain similarity regularization respectively. The inter-domain visual and semantic correlations are hierarchically modeled via bipartite graph structures, and the intra-domain relations are encoded via graph attention mechanisms. Through message-passing, each node aggregates semantic and contextual information from the same and opposite domain to substantially enhance its expressive power. Empirical results demonstrate that the proposed FGRR exceeds the state-of-the-art performance on four DAOD benchmarks.

We present a deep reinforcement learning method of progressive view inpainting for colored semantic point cloud scene completion under volume guidance, achieving high-quality scene reconstruction from only a single RGB-D image with severe occlusion. Our approach is end-to-end, consisting of three modules: 3D scene volume reconstruction, 2D RGB-D and segmentation image inpainting, and multi-view selection for completion. Given a single RGB-D image, our method first predicts its semantic segmentation map and goes through the 3D volume branch to obtain a volumetric scene reconstruction as a guide to the next view inpainting step, which attempts to make up the missing information; the third step involves projecting the volume under the same view of the input, concatenating them to complete the current view RGB-D and segmentation map, and integrating all RGB-D and segmentation maps into the point cloud. Since the occluded areas are unavailable, we resort to a A3C network to glance around and pick the next best view for large hole completion progressively until a scene is adequately reconstructed while guaranteeing validity. All steps are learned jointly to achieve robust and consistent results. We perform qualitative and quantitative evaluations with extensive experiments on the 3D-FUTURE data, obtaining better results than state-of-the-arts.

Transformer, the model of choice for natural language processing, has drawn scant attention from the medical imaging community. Given the ability to exploit long-term dependencies, transformers are promising to help atypical convolutional neural networks to learn more contextualized visual representations. However, most of recently proposed transformer-based segmentation approaches simply treated transformers as assisted modules to help encode global context into convolutional representations. To address this issue, we introduce nnFormer (i.e., not-another transFormer), a 3D transformer for volumetric medical image segmentation. nnFormer not only exploits the combination of interleaved convolution and self-attention operations, but also introduces local and global volume-based self-attention mechanism to learn volume representations. Moreover, nnFormer proposes to use skip attention to replace the traditional concatenation/summation operations in skip connections in U-Net like architecture. Experiments show that nnFormer significantly outperforms previous transformer-based counterparts by large margins on three public datasets. Compared to nnUNet, the most widely recognized convnet-based 3D medical segmentation model, nnFormer produces significantly lower HD95 and is much more computationally efficient. Furthermore, we show that nnFormer and nnUNet are highly complementary to each other in model ensembling. Codes and models of nnFormer are available at https://git.io/JSf3i.

Chronic Glaucoma is an eye disease with progressive optic nerve damage. It is the second leading cause of blindness after cataract and the first leading cause of irreversible blindness. Glaucoma forecast can predict future eye state of a patient by analyzing the historical fundus images, which is helpful for early detection and intervention of potential patients and avoiding the outcome of blindness. In this paper, we propose a GLaucoma forecast transformer based on Irregularly saMpled fundus images named GLIM-Net to predict the probability of developing glaucoma in the future. The main challenge is that the existing fundus images are often sampled at irregular times, making it difficult to accurately capture the subtle progression of glaucoma over time. We therefore introduce two novel modules, namely time positional encoding and time-sensitive MSA (multi-head self-attention) modules, to address this challenge. Unlike many existing works that focus on prediction for an unspecified future time, we also propose an extended model which is further capable of prediction conditioned on a specific future time. The experimental results on the benchmark dataset SIGF show that the accuracy of our method outperforms the state-of-the-art models. In addition, the ablation experiments also confirm the effectiveness of the two modules we propose, which can provide a good reference for the optimization of Transformer models.

Recently deep neural networks, which require a large amount of annotated samples, have been widely applied in nuclei instance segmentation of H&E stained pathology images. However, it is inefficient and unnecessary to label all pixels for a dataset of nuclei images which usually contain similar and redundant patterns. Although unsupervised and semi-supervised learning methods have been studied for nuclei segmentation, very few works have delved into the selective labeling of samples to reduce the workload of annotation. Thus, in this paper, we propose a novel full nuclei segmentation framework that chooses only a few image patches to be annotated, augments the training set from the selected samples, and achieves nuclei segmentation in a semi-supervised manner. In the proposed framework, we first develop a novel consistency-based patch selection method to determine which image patches are the most beneficial to the training. Then we introduce a conditional single-image GAN with a component-wise discriminator, to synthesize more training samples. Lastly, our proposed framework trains an existing segmentation model with the above augmented samples. The experimental results show that our proposed method could obtain the same-level performance as a fully-supervised baseline by annotating less than 5% pixels on some benchmarks.

Eyebrows play a critical role in facial expression and appearance. Although the 3D digitization of faces is well explored, less attention has been drawn to 3D eyebrow modeling. In this work, we propose EMS, the first learning-based framework for single-view 3D eyebrow reconstruction. Following the methods of scalp hair reconstruction, we also represent the eyebrow as a set of fiber curves and convert the reconstruction to fibers growing problem. Three modules are then carefully designed: RootFinder firstly localizes the fiber root positions which indicate where to grow; OriPredictor predicts an orientation field in the 3D space to guide the growing of fibers; FiberEnder is designed to determine when to stop the growth of each fiber. Our OriPredictor directly borrows the method used in hair reconstruction. Considering the differences between hair and eyebrows, both RootFinder and FiberEnder are newly proposed. Specifically, to cope with the challenge that the root location is severely occluded, we formulate root localization as a density map estimation task. Given the predicted density map, a density-based clustering method is further used for finding the roots. For each fiber, the growth starts from the root point and moves step by step until the ending, where each step is defined as an oriented line segment with a constant length according to the predicted orientation field. To determine when to end, a pixel-aligned RNN architecture is designed to form a binary classifier, which outputs stop or not for each growing step. To support the training of all proposed networks, we build the first 3D synthetic eyebrow dataset that contains 400 high-quality eyebrow models manually created by artists. Extensive experiments have demonstrated the effectiveness of the proposed EMS pipeline on a variety of different eyebrow styles and lengths, ranging from short and sparse to long bushy eyebrows.

Domain generalization (DG) aims to improve the generalization performance for an unseen target domain by using the knowledge of multiple seen source domains. Mainstream DG methods typically assume that the domain label of each source sample is known a priori, which is challenged to be satisfied in many real-world applications. In this paper, we study a practical problem of compound DG, which relaxes the discrete domain assumption to the mixed source domains setting. On the other hand, current DG algorithms prioritize the focus on semantic invariance across domains (one-vs-one), while paying less attention to the holistic semantic structure (many-vs-many). Such holistic semantic structure, referred to as meta-knowledge here, is crucial for learning generalizable representations. To this end, we present COmpound domain generalization via Meta-knowledge ENcoding (COMEN), a general approach to automatically discover and model latent domains in two steps. Firstly, we introduce Style-induced Domain-specific Normalization (SDNorm) to re-normalize the multi-modal underlying distributions, thereby dividing the mixture of source domains into latent clusters. Secondly, we harness the prototype representations, the centroids of classes, to perform relational modeling in the embedding space with two parallel and complementary modules, which explicitly encode the semantic structure for the out-of-distribution generalization. Experiments on four standard DG benchmarks reveal that COMEN exceeds the state-of-the-art performance without the need of domain supervision.

Generating expressive talking heads is essential for creating virtual humans. However, existing one- or few-shot methods focus on lip-sync and head motion, ignoring the emotional expressions that make talking faces realistic. In this paper, we propose the Granularly Controlled Audio-Visual Talking Heads (GC-AVT), which controls lip movements, head poses, and facial expressions of a talking head in a granular manner. Our insight is to decouple the audio-visual driving sources through prior-based pre-processing designs. Detailedly, we disassemble the driving image into three complementary parts including: 1) a cropped mouth that facilitates lip-sync; 2) a masked head that implicitly learns pose; and 3) the upper face which works corporately and complementarily with a time-shifted mouth to contribute the expression. Interestingly, the encoded features from the three sources are integrally balanced through reconstruction training. Extensive experiments show that our method generates expressive faces with not only synced mouth shapes, controllable poses, but precisely animated emotional expressions as well.

Waste inspection for packaged waste is an important step in the pipeline of waste disposal. Previous methods either rely on manual visual checking or RGB image-based inspection algorithm, requiring costly preparation procedures (e.g., open the bag and spread the waste items). Moreover, occluded items are very likely to be left out. Inspired by the fact that X-ray has a strong penetrating power to see through the bag and overlapping objects, we propose to perform waste inspection efficiently using X-ray images without the need to open the bag. We introduce a novel problem of instance-level waste segmentation in X-ray image for intelligent waste inspection, and contribute a real dataset consisting of 5,038 X-ray images (totally 30,881 waste items) with high-quality annotations (i.e., waste categories, object boxes, and instance-level masks) as a benchmark for this problem. As existing segmentation methods are mainly designed for natural images and cannot take advantage of the characteristics of X-ray waste images (e.g., heavy occlusions and penetration effect), we propose a new instance segmentation method to explicitly take these image characteristics into account. Specifically, our method adopts an easy-to-hard disassembling strategy to use high confidence predictions to guide the segmentation of highly overlapped objects, and a global structure guidance module to better capture the complex contour information caused by the penetration effect. Extensive experiments demonstrate the effectiveness of the proposed method. Our dataset is released at WIXRayNet.

Automatic tooth instance segmentation on 3D dental models is a fundamental task for computer-aided orthodontic treatments. Existing learning-based methods rely heavily on expensive point-wise annotations. To alleviate this problem, we are the first to explore a low-cost annotation way for 3D tooth instance segmentation, i.e., labeling all tooth centroids and only a few teeth for each dental model. Regarding the challenge when only weak annotation is provided, we present a dental arch prior-assisted 3D tooth segmentation method, namely DArch. Our DArch consists of two stages, including tooth centroid detection and tooth instance segmentation. Accurately detecting the tooth centroids can help locate the individual tooth, thus benefiting the segmentation. Thus, our DArch proposes to leverage the dental arch prior to assist the detection. Specifically, we firstly propose a coarse-to-fine method to estimate the dental arch, in which the dental arch is initially generated by Bezier curve regression, and then a graph-based convolutional network (GCN) is trained to refine it. With the estimated dental arch, we then propose a novel Arch-aware Point Sampling (APS) method to assist the tooth centroid proposal generation. Meantime, a segmentor is independently trained using a patch-based training strategy, aiming to segment a tooth instance from a 3D patch centered at the tooth centroid. Experimental results on 4, 773 dental models have shown our DArch can accurately segment each tooth of a dental model, and its performance is superior to the state-of-the-art methods.

Long-range temporal alignment is critical yet challenging for video restoration tasks. Recently, some works attempt to divide the long-range alignment into several sub-alignments and handle them progressively. Although this operation is helpful in modeling distant correspondences, error accumulation is inevitable due to the propagation mechanism. In this work, we present a novel, generic iterative alignment module which employs a gradual refinement scheme for sub-alignments, yielding more accurate motion compensation. To further enhance the alignment accuracy and temporal consistency, we develop a non-parametric re-weighting method, where the importance of each neighboring frame is adaptively evaluated in a spatial-wise way for aggregation. By virtue of the proposed strategies, our model achieves state-of-the-art performance on multiple benchmarks across a range of video restoration tasks including video super-resolution, denoising and deblurring.

Fueled by the power of deep learning techniques and implicit shape learning, recent advances in single-image human digitalization have reached unprecedented accuracy and could recover fine-grained surface details such as garment wrinkles. However, a common problem for the implicit-based methods is that they cannot produce separated and topology-consistent mesh for each garment piece, which is crucial for the current 3D content creation pipeline. To address this issue, we proposed a novel geometry inference framework ReEF that reconstructs topology-consistent layered garment mesh by registering the explicit garment template to the whole-body implicit fields predicted from single images. Experiments demonstrate that our method notably outperforms the counterparts on single-image layered garment reconstruction and could bring high-quality digital assets for further content creation.

Excellent performance has been achieved on instance segmentation but the quality on the boundary area remains unsatisfactory, which leads to a rising attention on boundary refinement. For practical use, an ideal post-processing refinement scheme are required to be accurate, generic and efficient. However, most of existing approaches propose pixel-wise refinement, which either introduce a massive computation cost or design specifically for different backbone models. Contour-based models are efficient and generic to be incorporated with any existing segmentation methods, but they often generate over-smoothed contour and tend to fail on corner areas. In this paper, we propose an efficient contour-based boundary refinement approach, named SharpContour, to tackle the segmentation of boundary area. We design a novel contour evolution process together with an Instance-aware Point Classifier. Our method deforms the contour iteratively by updating offsets in a discrete manner. Differing from existing contour evolution methods, SharpContour estimates each offset more independently so that it predicts much sharper and accurate contours. Notably, our method is generic to seamlessly work with diverse existing models with a small computational cost. Experiments show that SharpContour achieves competitive gains whilst preserving high efficiency.

Semi-supervised domain adaptation (SSDA) aims to apply knowledge learned from a fully labeled source domain to a scarcely labeled target domain. In this paper, we propose a Multi-level Consistency Learning (MCL) framework for SSDA. Specifically, our MCL regularizes the consistency of different views of target domain samples at three levels: (i) at inter-domain level, we robustly and accurately align the source and target domains using a prototype-based optimal transport method that utilizes the pros and cons of different views of target samples; (ii) at intra-domain level, we facilitate the learning of both discriminative and compact target feature representations by proposing a novel class-wise contrastive clustering loss; (iii) at sample level, we follow standard practice and improve the prediction accuracy by conducting a consistency-based self-training. Empirically, we verified the effectiveness of our MCL framework on three popular SSDA benchmarks, i.e., VisDA2017, DomainNet, and Office-Home datasets, and the experimental results demonstrate that our MCL framework achieves the state-of-the-art performance.

A key challenge of real-world image super-resolution (SR) is to recover the missing details in low-resolution (LR) images with complex unknown degradations (\eg, downsampling, noise and compression). Most previous works restore such missing details in the image space. To cope with the high diversity of natural images, they either rely on the unstable GANs that are difficult to train and prone to artifacts, or resort to explicit references from high-resolution (HR) images that are usually unavailable. In this work, we propose Feature Matching SR (FeMaSR), which restores realistic HR images in a much more compact feature space. Unlike image-space methods, our FeMaSR restores HR images by matching distorted LR image features to their distortion-free HR counterparts in our pretrained HR priors, and decoding the matched features to obtain realistic HR images. Specifically, our HR priors contain a discrete feature codebook and its associated decoder, which are pretrained on HR images with a Vector Quantized Generative Adversarial Network (VQGAN). Notably, we incorporate a novel semantic regularization in VQGAN to improve the quality of reconstructed images. For the feature matching, we first extract LR features with an LR encoder consisting of several Swin Transformer blocks and then follow a simple nearest neighbour strategy to match them with the pretrained codebook. In particular, we equip the LR encoder with residual shortcut connections to the decoder, which is critical to the optimization of feature matching loss and also helps to complement the possible feature matching errors.Experimental results show that our approach produces more realistic HR images than previous methods. Code will be made publicly available.

We present PVSeRF, a learning framework that reconstructs neural radiance fields from single-view RGB images, for novel view synthesis. Previous solutions, such as pixelNeRF, rely only on pixel-aligned features and suffer from feature ambiguity issues. As a result, they struggle with the disentanglement of geometry and appearance, leading to implausible geometries and blurry results. To address this challenge, we propose to incorporate explicit geometry reasoning and combine it with pixel-aligned features for radiance field prediction. Specifically, in addition to pixel-aligned features, we further constrain the radiance field learning to be conditioned on i) voxel-aligned features learned from a coarse volumetric grid and ii) fine surface-aligned features extracted from a regressed point cloud. We show that the introduction of such geometry-aware features helps to achieve a better disentanglement between appearance and geometry, i.e. recovering more accurate geometries and synthesizing higher quality images of novel views. Extensive experiments against state-of-the-art methods on ShapeNet benchmarks demonstrate the superiority of our approach for single-image novel view synthesis.

Abstract Point cloud generation aims to synthesize point clouds that do not exist in supervised dataset. Generating a point cloud with certain semantic labels remains an under‐explored problem. This paper proposes a formulation called DiffusionPointLabel, which completes point‐label pair generation based on a DDPM generative model (Denoising Diffusion Probabilistic Model). Specifically, we use a point cloud diffusion generative model and aggregate the intermediate features of the generator. On top of this, we propose Feature Interpreter that transforms intermediate features into semantic labels. Furthermore, we employ an uncertainty measure to filter unqualified point‐label pairs for a better quality of generated point cloud dataset. Coupling these two designs enables us to automatically generate annotated point clouds, especially when supervised point‐labels pairs are scarce. Our method extends the application of point cloud generation models and surpasses state‐of‐the‐art models.

There is a steadily growing range of applications that can benefit from facial reconstruction techniques, leading to an increasing demand for reconstruction of high-quality 3D face models. While it is an important expressive part of the human face, the nose has received less attention than other expressive regions in the face reconstruction literature. When applying existing reconstruction methods to facial images, the reconstructed nose models are often inconsistent with the desired shape and expression. In this paper, we propose a coarse-to-fine 3D nose reconstruction and correction pipeline to build a nose model from a single image, where 3D and 2D nose curve correspondences are adaptively updated and refined. We first correct the reconstruction result coarsely using constraints of 3D-2D sparse landmark correspondences, and then heuristically update a dense 3D-2D curve correspondence based on the coarsely corrected result. A final refinement step is performed to correct the shape based on the updated 3D-2D dense curve constraints. Experimental results show the advantages of our method for 3D nose reconstruction over existing methods.

This paper focuses on the challenging task of learning 3D object surface reconstructions from RGB images. Existing methods achieve varying degrees of success by using different surface representations. However, they all have their own drawbacks, and cannot properly reconstruct the surface shapes of complex topologies, arguably due to a lack of constraints on the topological structures in their learning frameworks. To this end, we propose to learn and use the topology-preserved, skeletal shape representation to assist the downstream task of object surface reconstruction from RGB images. Technically, we propose the novel SkeletonNet design that learns a volumetric representation of a skeleton via a bridged learning of a skeletal point set, where we use parallel decoders each responsible for the learning of points on 1D skeletal curves and 2D skeletal sheets, as well as an efficient module of globally guided subvolume synthesis for a refined, high-resolution skeletal volume; we present a differentiable Point2Voxel layer to make SkeletonNet end-to-end and trainable. With the learned skeletal volumes, we propose two models, the Skeleton-Based Graph Convolutional Neural Network (SkeGCNN) and the Skeleton-Regularized Deep Implicit Surface Network (SkeDISN), which respectively build upon and improve over the existing frameworks of explicit mesh deformation and implicit field learning for the downstream surface reconstruction task. We conduct thorough experiments that verify the efficacy of our proposed SkeletonNet. SkeGCNN and SkeDISN outperform existing methods as well, and they have their own merits when measured by different metrics. Additional results in generalized task settings further demonstrate the usefulness of our proposed methods. We have made our implementation code publicly available at https://github.com/tangjiapeng/SkeletonNet.

Estimating 3D human pose from a single image is a challenging task. This work attempts to address the uncertainty of lifting the detected 2D joints to the 3D space by introducing an intermediate state - Part-Centric Heatmap Triplets (HEMlets), which shortens the gap between the 2D observation and the 3D interpretation. The HEMlets utilize three joint-heatmaps to represent the relative depth information of the end-joints for each skeletal body part. In our approach, a Convolutional Network (ConvNet) is first trained to predict HEMlets from the input image, followed by a volumetric joint-heatmap regression. We leverage on the integral operation to extract the joint locations from the volumetric heatmaps, guaranteeing end-to-end learning. Despite the simplicity of the network design, the quantitative comparisons show a significant performance improvement over the best-of-grade methods (e.g., 20 percent on Human3.6M). The proposed method naturally supports training with "in-the-wild" images, where only weakly-annotated relative depth information of skeletal joints is available. This further improves the generalization ability of our model, as validated by qualitative comparisons on outdoor images. Leveraging the strength of the HEMlets pose estimation, we further design and append a shallow yet effective network module to regress the SMPL parameters of the body pose and shape. We term the entire HEMlets-based human pose and shape recovery pipeline HEMlets PoSh. Extensive quantitative and qualitative experiments on the existing human body recovery benchmarks justify the state-of-the-art results obtained with our HEMlets PoSh approach.

In the game and film industries, modeling 3D heads plays a very important role in designing characters. Although human head modeling has been researched for a long time, few works have focused on animal-like heads, which are of more diverse shapes and richer geometric details. In this article, we present SAniHead, an interactive system for creating animal-like heads with a mesh representation from dual-view sketches. Our core technical contribution is a view-surface collaborative mesh generative network. Initially, a graph convolutional neural network (GCNN) is trained to learn the deformation of a template mesh to fit the shape of sketches, giving rise to a coarse model. It is then projected into vertex maps where image-to-image translation networks are performed for detail inference. After back-projecting the inferred details onto the meshed surface, a new GCNN is trained for further detail refinement. The modules of view-based detail inference and surface-based detail refinement are conducted in an alternating cascaded fashion, collaboratively improving the model. A refinement sketching interface is also implemented to support direct mesh manipulation. Experimental results show the superiority of our approach and the usability of our interactive system. Our work also contributes a 3D animal head dataset with corresponding line drawings.

Grounding referring expressions in RGBD image has been an emerging field. We present a novel task of 3D visual grounding in single-view RGBD image where the referred objects are often only partially scanned due to occlusion. In contrast to previous works that directly generate object proposals for grounding in the 3D scenes, we propose a bottom-up approach to gradually aggregate content-aware information, effectively addressing the challenge posed by the partial geometry. Our approach first fuses the language and the visual features at the bottom level to generate a heatmap that coarsely localizes the relevant regions in the RGBD image. Then our approach conducts an adaptive feature learning based on the heatmap and performs the object-level matching with another visio-linguistic fusion to finally ground the referred object. We evaluate the proposed method by comparing to the state-of-the-art methods on both the RGBD images extracted from the ScanRefer dataset and our newly collected SUNRefer dataset. Experiments show that our method outperforms the previous methods by a large margin (by 11.2% and 15.6% [email protected]) on both datasets.

Semantic scene understanding from point clouds is particularly challenging as the points reflect only a sparse set of the underlying 3D geometry. Previous works often convert point cloud into regular grids (e.g. voxels or bird-eye view images), and resort to grid-based convolutions for scene understanding. In this work, we introduce RfD-Net that jointly detects and reconstructs dense object surfaces directly from raw point clouds. Instead of representing scenes with regular grids, our method leverages the sparsity of point cloud data and focuses on predicting shapes that are recognized with high objectness. With this design, we decouple the instance reconstruction into global object localization and local shape prediction. It not only eases the difficulty of learning 2-D manifold surfaces from sparse 3D space, the point clouds in each object proposal convey shape details that support implicit function learning to reconstruct any high-resolution surfaces. Our experiments indicate that instance detection and reconstruction present complementary effects, where the shape prediction head shows consistent effects on improving object detection with modern 3D proposal network backbones. The qualitative and quantitative evaluations further demonstrate that our approach consistently outperforms the state-of-the-arts and improves over 11 of mesh IoU in object reconstruction.

Caricature is an artistic representation that deliberately exaggerates the distinctive features of a human face to convey humor or sarcasm. However, reconstructing a 3D caricature from a 2D caricature image remains a challenging task, mostly due to the lack of data. We propose to fill this gap by introducing 3DCaricShop, the first large-scale 3D caricature dataset that contains 2000 high-quality diversified 3D caricatures manually crafted by professional artists. 3DCaricShop also provides rich annotations including a paired 2D caricature image, camera parameters and 3D facial landmarks. To demonstrate the advantage of 3DCaricShop, we present a novel baseline approach for single-view 3D caricature reconstruction. To ensure a faithful reconstruction with plausible face deformations, we propose to connect the good ends of the detail-rich implicit functions and the parametric mesh representations. In particular, we first register a template mesh to the output of the implicit generator and iteratively project the registration result onto a pre-trained PCA space to resolve artifacts and self-intersections. To deal with the large deformation during non-rigid registration, we propose a novel view-collaborative graph convolution network (VC-GCN) to extract key points from the implicit mesh for accurate alignment. Our method is able to generate high-fidelity 3D caricature in a pre-defined mesh topology that is animation-ready. Extensive experiments have been conducted on 3DCaricShop to verify the significance of the database and the effectiveness of the proposed method. We will release 3DCaricShop upon publication.

Point completion refers to completing the missing geometries of an object from incomplete observations. Mainstream methods predict the missing shapes by decoding a global feature learned from the input point cloud, which often leads to deficient results in preserving topology consistency and surface details. In this work, we present MEPCN, a point completion network that leverages emptiness in 3D shape space. Given a single depth scan, previous methods often encode the occupied partial shapes while ignoring the empty regions (e.g. holes) in depth maps. In contrast, we argue that these ‘emptiness’ clues indicate shape boundaries that can be used to improve topology representation and detail granularity on surfaces. Specifically, our ME-PCN encodes both the occupied point cloud and the neighboring ‘empty points’. It estimates coarse-grained but complete and reasonable surface points in the first stage, followed by a refinement stage to produce fine-grained surface details. Comprehensive experiments verify that our ME-PCN presents better qualitative and quantitative performance against the state-of-the-art. Besides, we further prove that our ‘emptiness’ design is lightweight and easy to embed in existing methods, which shows consistent effectiveness in improving the CD and EMD scores.

The key point of language-guided person search is to construct the cross-modal association between visual and textual input. Existing methods focus on designing multimodal attention mechanisms and novel cross-modal loss functions to learn such association implicitly. We propose a representation learning method for language-guided person search based on color reasoning (LapsCore). It can explicitly build a fine-grained cross-modal association bidirectionally. Specifically, a pair of dual sub-tasks, image colorization and text completion, is designed. In the former task, rich text information is learned to colorize gray images, and the latter one requests the model to understand the image and complete color word vacancies in the captions. The two sub-tasks enable models to learn correct alignments between text phrases and image regions, so that rich multimodal representations can be learned. Extensive experiments on multiple datasets demonstrate the effectiveness and superiority of the proposed method.

Preserving maximal information is one of principles of designing self-supervised learning methodologies. To reach this goal, contrastive learning adopts an implicit way which is contrasting image pairs. However, we believe it is not fully optimal to simply use the contrastive estimation for preservation. Moreover, it is necessary and complemental to introduce an explicit solution to preserve more information. From this perspective, we introduce Preservational Learning to reconstruct diverse image contexts in order to preserve more information in learned representations. Together with the contrastive loss, we present Preservational Contrastive Representation Learning (PCRL) for learning self-supervised medical representations. PCRL provides very competitive results under the pretraining-finetuning protocol, outperforming both self-supervised and supervised counterparts in 5 classification/segmentation tasks substantially. Codes are available at https://github.com/Luchixiang/PCRL.

Recent advances in unsupervised domain adaptation have achieved remarkable performance on semantic segmentation tasks. Despite such progress, existing works mainly focus on bridging the inter-domain gaps between the source and target domain, while only few of them noticed the intra-domain gaps within the target data. In this work, we propose a pixel-level intra-domain adaptation approach to reduce the intra-domain gaps within the target data. Compared with image-level methods, ours treats each pixel as an instance, which adapts the segmentation model at a more fine-grained level. Specifically, we first conduct the inter-domain adaptation between the source and target domain; Then, we separate the pixels in target images into the easy and hard subdomains; Finally, we propose a pixel-level adversarial training strategy to adapt a segmentation network from the easy to the hard subdomain. Moreover, we show that the segmentation accuracy can be further improved by incorporating a continuous indexing technique in the adversarial training. Experimental results show the effectiveness of our method against existing state-of-the-art approaches.

Head shapes play an important role in 3D character design. In this work, we propose SimpModeling, a novel sketch-based system for helping users, especially amateur users, easily model 3D animalmorphic heads - a prevalent kind of heads in character design. Although sketching provides an easy way to depict desired shapes, it is challenging to infer dense geometric information from sparse line drawings. Recently, deepnet-based approaches have been taken to address this challenge and try to produce rich geometric details from very few strokes. However, while such methods reduce users’ workload, they would cause less controllability of target shapes. This is mainly due to the uncertainty of the neural prediction. Our system tackles this issue and provides good controllability from three aspects: 1) we separate coarse shape design and geometric detail specification into two stages and respectively provide different sketching means; 2) in coarse shape designing, sketches are used for both shape inference and geometric constraints to determine global geometry, and in geometric detail crafting, sketches are used for carving surface details; 3) in both stages, we use the advanced implicit-based shape inference methods, which have strong ability to handle the domain gap between freehand sketches and synthetic ones used for training. Experimental results confirm the effectiveness of our method and the usability of our interactive system. We also contribute to a dataset of high-quality 3D animal heads, which are manually created by artists.

In this paper, we provide a bidirectional shadow rendering method to render shadows between real and virtual objects in the 360° videos in real time. We construct a 3D scene approximation from the current output viewpoint to approximate the real scene geometry nearby in the video. Then, we propose a ray casting based algorithm to determine the shadow regions on the virtual objects cast by the real objects. After that, we introduce an object-aware shadow mapping method to cast shadows from virtual objects to real objects. Finally, we use a shadow intensity estimation algorithm to determine the shadow intensity of virtual objects and real objects to obtain shadows consistent with the input 360° video. The experiment results prove the effectiveness of our bidirectional shadow rendering method for mixed 360° videos. Our method can generate visually realistic shadows for virtual objects and real objects in 360° video in realtime, and make virtual objects more natural to integrate with real scenes in 360° videos of the mixed reality applications.

Skeleton-based action recognition has attracted research attentions in recent years. One common drawback in currently popular skeleton-based human action recognition methods is that the sparse skeleton information alone is not sufficient to fully characterize human motion. This limitation makes several existing methods incapable of correctly classifying action categories which exhibit only subtle motion differences. In this paper, we propose a novel framework for employing human pose skeleton and joint-centered light-weight information jointly in a two-stream graph convolutional network, namely, JOLO-GCN. Specifically, we use Joint-aligned optical Flow Patches (JFP) to capture the local subtle motion around each joint as the pivotal joint-centered visual information. Compared to the pure skeleton-based baseline, this hybrid scheme effectively boosts performance, while keeping the computational and memory overheads low. Experiments on the NTU RGB+D, NTU RGB+D 120, and the Kinetics-Skeleton dataset demonstrate clear accuracy improvements attained by the proposed method over the state-of-the-art skeleton-based methods.

Abstract Modeling 3D objects on existing software usually requires a heavy amount of interactions, especially for users who lack basic knowledge of 3D geometry. Sketch‐based modeling is a solution to ease the modelling procedure and thus has been researched for decades. However, modelling a man‐made shape with complex structures remains challenging. Existing methods adopt advanced deep learning techniques to map holistic sketches to 3D shapes. They are still bottlenecked to deal with complicated topologies. In this paper, we decouple the task of sketch2shape into a part generation module and a part assembling module, where deep learning methods are leveraged for the implementation of both modules. By changing the focus from holistic shapes to individual parts, it eases the learning process of the shape generator and guarantees high‐quality outputs. With the learned automated part assembler, users only need a little manual tuning to obtain a desired layout. Extensive experiments and user studies demonstrate the usefulness of our proposed system.

In most recent years, Siamese trackers have drawn great attention because of their well-balanced accuracy and efficiency. Although these approaches have achieved great success, the discriminative power of the conventional Siamese trackers is still limited by the insufficient template-candidate representation. Most of the existing approaches take non-aligned features to learn a similarity function for template-candidate matching, while the target object's geometrical transformation is seldom explored. To address this problem, we propose a novel Siamese tracking framework, which enables to dynamically transform the template-candidate features to a more discriminative viewpoint for similarity matching. Specifically, we reformulate the template-candidate matching problem of the conventional Siamese tracker from the perspective of Lucas-Kanade (LK) image alignment approach. A Lucas-Kanade network (LKNet) is proposed and incorporated to the Siamese architecture to learn aligned feature representations in data-driven trainable manner, which is able to enhance the model adaptability in challenging scenarios. Within this framework, we propose two Siamese trackers named LK-Siam and LK-SiamRPN to validate the effectiveness. Extensive experiments conducted on the prevalent datasets show that the proposed method is more competitive over a number of state-of-the-art methods.

Existing tracking-by-detection approaches using deep features have achieved promising results in recent years. However, these methods mainly exploit feature representations learned from individual static frames, thus paying little attention to the temporal smoothness between frames. This easily leads trackers to drift in the presence of large appearance variations and occlusions. To address this issue, we propose a two-stream network to learn discriminative spatio-temporal feature representations to represent the target objects. The proposed network consists of a Spatial ConvNet module and a Temporal ConvNet module. Specifically, the Spatial ConvNet adopts 2D convolutions to encode the target-specific appearance in static frames, while the Temporal ConvNet models the temporal appearance variations using 3D convolutions and learns consistent temporal patterns in a short video clip. Then we propose a proposal refinement module to adjust the predicted bounding box, which can make the target localizing outputs to be more consistent in video sequences. In addition, to improve the model adaptation during online update, we propose a contrastive online hard example mining (OHEM) strategy, which selects hard negative samples and enforces them to be embedded in a more discriminative feature space. Extensive experiments conducted on the OTB, Temple Color and VOT benchmarks demonstrate that the proposed algorithm performs favorably against the state-of-the-art methods.

Recent deep generative models allow real-time generation of hair images from sketch inputs. Existing solutions often require a user-provided binary mask to specify a target hair shape. This not only costs users extra labor but also fails to capture complicated hair boundaries. Those solutions usually encode hair structures via orientation maps, which, however, are not very effective to encode complex structures. We observe that colored hair sketches already implicitly define target hair shapes as well as hair appearance and are more flexible to depict hair structures than orientation maps. Based on these observations, we present SketchHairSalon , a two-stage framework for generating realistic hair images directly from freehand sketches depicting desired hair structure and appearance. At the first stage, we train a network to predict a hair matte from an input hair sketch, with an optional set of non-hair strokes. At the second stage, another network is trained to synthesize the structure and appearance of hair images from the input sketch and the generated matte. To make the networks in the two stages aware of long-term dependency of strokes, we apply self-attention modules to them. To train these networks, we present a new dataset containing thousands of annotated hair sketch-image pairs and corresponding hair mattes. Two efficient methods for sketch completion are proposed to automatically complete repetitive braided parts and hair strokes, respectively, thus reducing the workload of users. Based on the trained networks and the two sketch completion strategies, we build an intuitive interface to allow even novice users to design visually pleasing hair images exhibiting various hair structures and appearance via freehand sketches. The qualitative and quantitative evaluations show the advantages of the proposed system over the existing or alternative solutions.

When a user walks freely in an unknown virtual scene and searches for multiple dynamic targets, the lack of a comprehensive understanding of the environment may have a negative impact on the execution of virtual reality tasks. Previous studies can help users with auxiliary tools, such as top view maps or trails, and exploration guidance, for example, automatically generated paths according to the user location and important static spots in virtual scenes. However, in some virtual reality applications, when the scene has complex occlusions, and the user cannot obtain any real-time position information of the dynamic target, the above assistance cannot help the user complete the task more effectively. We design a virtual camera priority-based assistance to help the user search dynamic targets efficiently. Instead of forcing users to go to destinations, we provide an optimized instant path to guide them to places where they are more likely to find dynamic targets when they ask for help. We assume that a certain number of virtual cameras are fixed in virtual scenes to obtain extra depth maps, which capture the depth information of the scene and the locations of the dynamic targets. Our methodautomatically analyzes the priority of these virtual cameras, chooses the destination, and generates an instant path to assist the user in finding the dynamic targets. Our method is suitable for various virtual reality applications that do not require manual supervision or input. A user study is designed to evaluate the proposed method. The results indicate that compared with three conventional navigation methods, such as the top-view method, our method can help users find dynamic targets more efficiently. The advantages include reducing the task completion time, reducing the number of resets, increasing the average distance between resets, and reducing user task load. We presented a method for improving dynamic target searching efficiency in virtual scenes by virtual camera priority-based path guidance. Compared with three conventional navigation methods, such as the top-view method, this method can help users find dynamic targets more effectively.

With the advent of deep neural networks, learning-based single-view reconstruction has gained popularity. However, in 3D, there is no absolutely dominant representation that is both computationally efficient and accurate yet allows for reconstructing high-resolution geometry of arbitrary topology. After all, the accurate implicit methods are time-consuming due to dense sampling and inference, while volumetric approaches are fast but limited to heavy memory usage and low accuracy. In this paper, we propose VIPNet, an end-to-end hybrid representation learning for fast and accurate single-view reconstruction under sparse implicit point guidance. Given an image, it first generates a volumetric result. Meanwhile, a corresponding implicit shape representation is learned. To balance the efficiency and accuracy, we adopt PointGenNet to learn some representative points for guiding the voxel refinement with the corresponding sparse implicit inference. A strategy of patch-based synthesis with global-local features under implicit guidance is also applied for reducing memory consumption required to generate high-resolution output. Extensive experiments demonstrate the effectiveness of our method both qualitatively and quantitatively, which indicates that our proposed hybrid learning outperforms separate representation learning. Specifically, our network not only runs 60 times faster than implicit methods but also contributes to accuracy gains. We hope it will inspire a re-thinking of hybrid representation learning.

We introduce FPConv, a novel surface-style convolution operator designed for 3D point cloud analysis. Unlike previous methods, FPConv doesn't require transforming to intermediate representation like 3D grid or graph and directly works on surface geometry of point cloud. To be more specific, for each point, FPConv performs a local flattening by automatically learning a weight map to softly project surrounding points onto a 2D grid. Regular 2D convolution can thus be applied for efficient feature learning. FPConv can be easily integrated into various network architectures for tasks like 3D object classification and 3D scene segmentation, and achieve comparable performance with existing volumetric-type convolutions. More importantly, our experiments also show that FPConv can be a complementary of volumetric convolutions and jointly training them can further boost overall performance into state-of-the-art results.

Semantic reconstruction of indoor scenes refers to both scene understanding and object reconstruction. Existing works either address one part of this problem or focus on independent objects. In this paper, we bridge the gap between understanding and reconstruction, and propose an end-to-end solution to jointly reconstruct room layout, object bounding boxes and meshes from a single image. Instead of separately resolving scene understanding and object reconstruction, our method builds upon a holistic scene context and proposes a coarse-to-fine hierarchy with three components: 1. room layout with camera pose; 2. 3D object bounding boxes; 3. object meshes. We argue that understanding the context of each component can assist the task of parsing the others, which enables joint understanding and reconstruction. The experiments on the SUN RGB-D and Pix3D datasets demonstrate that our method consistently outperforms existing methods in indoor layout estimation, 3D object detection and mesh reconstruction.

In this paper, we present a deep learning approach for shading decomposition from single garment images with complex textured albedos. We propose a multi-stream generator to infer albedo, shading, normal and lighting in a unified network and train it in a novel local-global adversarial learning framework. To support the training, we carefully build a large dataset with synthetic garment photos under various lighting conditions and textured albedos. Both quantitative and qualitative evaluations, on synthetic and real images, demonstrate the superiority of our method against the state-of-the-art.

In this paper, we focus on protecting the facial privacy for people under the surveillance scenarios, by changing some visual appearances of the faces while keeping them recognizable by the current face recognition systems. This is a challenging problem because we need to retain the most important structures of the captured facial images, while modify the salient facial regions to protect personal privacy. To address this problem, we introduce a novel individual face protection model, which can camouflage the face appearance from the perspective of human visual perception and preserve the identity features of faces used for face authentication. To that end, we develop an encoder-decoder network architecture which can separately disentangle the facial feature representation into an appearance code and an identification code. Specifically, we first randomly divide the input face image into two groups, the source and target sets, where the identity and appearance codes can be correspondingly extracted. Then, we recombine the identity and appearance codes to synthesize a new face, which has the same identity as the source subject. Finally, the synthesized faces are employed to replace the original face to protect the individual privacy. Note that our model is end-to-end with a multi-task loss function, which can better preserve the identity and stabilize the training process. Experiments conducted on Cross-Age Celebrity dataset demonstrate the effectiveness of our model and validate our superiority in terms of visual quality and scalability.

Temporal grounding of natural language in untrimmed videos is a fundamental yet challenging multimedia task facilitating cross-media visual content retrieval. We focus on the weakly supervised setting of this task that merely accesses to coarse video-level language description annotation without temporal boundary, which is more consistent with reality as such weak labels are more readily available in practice. In this paper, we propose a Boundary Adaptive Refinement (BAR) framework that resorts to reinforcement learning (RL) to guide the process of progressively refining the temporal boundary. To the best of our knowledge, we offer the first attempt to extend RL to temporal localization task with weak supervision. As it is non-trivial to obtain a straightforward reward function in the absence of pairwise granular boundary-query annotations, a cross-modal alignment evaluator is crafted to measure the alignment degree of segment-query pair to provide tailor-designed rewards. This refinement scheme completely abandons traditional sliding window based solution pattern and contributes to acquiring more efficient, boundary-flexible and content-aware grounding results. Extensive experiments on two public benchmarks Charades-STA and ActivityNet demonstrate that BAR outperforms the state-of-the-art weakly-supervised method and even beats some competitive fully-supervised ones.

Recent research has demonstrated that adding some imperceptible perturbations to original images can fool deep learning models. However, the current adversarial perturbations are usually shown in the form of noises, and thus have no practical meaning. Image watermark is a technique widely used for copyright protection. We can regard image watermark as a king of meaningful noises and adding it to the original image will not affect people's understanding of the image content, and will not arouse people's suspicion. Therefore, it will be interesting to generate adversarial examples using watermarks. In this paper, we propose a novel watermark perturbation for adversarial examples (Adv-watermark) which combines image watermarking techniques and adversarial example algorithms. Adding a meaningful watermark to the clean images can attack the DNN models. Specifically, we propose a novel optimization algorithm, which is called Basin Hopping Evolution (BHE), to generate adversarial watermarks in the black-box attack mode. Thanks to the BHE, Adv-watermark only requires a few queries from the threat models to finish the attacks. A series of experiments conducted on ImageNet and CASIA-WebFace datasets show that the proposed method can efficiently generate adversarial examples, and outperforms the state-of-the-art attack methods. Moreover, Adv-watermark is more robust against image transformation defense methods.

We present a novel framework for human video motion transfer. Deviating from recent studies that use only single source image, we propose to allow users to supply multiple source images by simply imitating some poses in the desired target video. To aggregate the appearance from multiple input images, we propose a JAFPro framework that incorporates two modules: an appearance fusion module that adaptively fuses the information in the supplied images and an appearance propagation module that propagates textures through flow-based warping to further improve the result. An attractive feature of JAFPro is that the quality of its results progressively improves as more imitating images are supplied. Furthermore, we build a new dataset containing a large variety of dancing videos in the wild. Extensive experiments conducted on this dataset demonstrate JAFPro outperforms state-of-the-art methods both qualitatively and quantitatively. We will release our code and dataset upon publication of this work.

Point completion refers to complete the missing geometries of objects from partial point clouds. Existing works usually estimate the missing shape by decoding a latent feature encoded from the input points. However, real-world objects are usually with diverse topologies and surface details, which a latent feature may fail to represent to recover a clean and complete surface. To this end, we propose a skeleton-bridged point completion network (SK-PCN) for shape completion. Given a partial scan, our method first predicts its 3D skeleton to obtain the global structure, and completes the surface by learning displacements from skeletal points. We decouple the shape completion into structure estimation and surface reconstruction, which eases the learning difficulty and benefits our method to obtain on-surface details. Besides, considering the missing features during encoding input points, SK-PCN adopts a local adjustment strategy that merges the input point cloud to our predictions for surface refinement. Comparing with previous methods, our skeleton-bridged manner better supports point normal estimation to obtain the full surface mesh beyond point clouds. The qualitative and quantitative experiments on both point cloud and mesh completion show that our approach outperforms the existing methods on various object categories.

As a domain-specific super-resolution problem, facial image hallucination has enjoyed a series of breakthroughs thanks to the advances of deep convolutional neural networks. However, the direct migration of existing methods to video is still difficult to achieve good performance due to its lack of alignment and consistency modelling in temporal domain. Taking advantage of high inter-frame dependency in videos, we propose a selfenhanced convolutional network for facial video hallucination. It is implemented by making full usage of preceding super-resolved frames and a temporal window of adjacent low-resolution frames. Specifically, the algorithm first obtains the initial high-resolution inference of each frame by taking into consideration a sequence of consecutive low-resolution inputs through temporal consistency modelling. It further recurrently exploits the reconstructed results and intermediate features of a sequence of preceding frames to improve the initial super-resolution of the current frame by modelling the coherence of structural facial features across frames. Quantitative and qualitative evaluations demonstrate the superiority of the proposed algorithm against state-of-theart methods. Moreover, our algorithm also achieves excellent performance in the task of general video super-resolution in a single-shot setting.

In this paper, we propose the first sketching system for interactively personalized and photorealistic face caricaturing. Input an image of a human face, the users can create caricature photos by manipulating its facial feature curves. Our system first performs exaggeration on the recovered 3D face model, which is conducted by assigning the laplacian of each vertex a scaling factor according to the edited sketches. The mapping between 2D sketches and the vertex-wise scaling field is constructed by a novel deep learning architecture. Our approach allows outputting different exaggerations when applying the same sketching on different input figures in term of their different geometric characteristics, which makes the generated results "personalized". With the obtained 3D caricature model, two images are generated, one obtained by applying 2D warping guided by the underlying 3D mesh deformation and the other obtained by re-rendering the deformed 3D textured model. These two images are then seamlessly integrated to produce our final output. Due to the severe stretching of meshes, the rendered texture is of blurry appearances. A deep learning approach is exploited to infer the missing details for enhancing these blurry regions. Moreover, a relighting operation is invented to further improve the photorealism of the result. These further make our results "photorealistic". The qualitative experiment results validated the efficiency of our sketching system.

We present a new data-driven video inpainting method for recovering missing regions of video frames. A novel deep learning architecture is proposed which contains two subnetworks: a temporal structure inference network and a spatial detail recovering network. The temporal structure inference network is built upon a 3D fully convolutional architecture: it only learns to complete a low-resolution video volume given the expensive computational cost of 3D convolution. The low resolution result provides temporal guidance to the spatial detail recovering network, which performs imagebased inpainting with a 2D fully convolutional network to produce recovered video frames in their original resolution. Such two-step network design ensures both the spatial quality of each frame and the temporal coherence across frames. Our method jointly trains both sub-networks in an end-to-end manner. We provide qualitative and quantitative evaluation on three datasets, demonstrating that our method outperforms previous learning-based video inpainting methods.

We present a deep reinforcement learning method of progressive view inpainting for 3D point scene completion under volume guidance, achieving high-quality scene reconstruction from only a single depth image with severe occlusion. Our approach is end-to-end, consisting of three modules: 3D scene volume reconstruction, 2D depth map inpainting, and multi-view selection for completion. Given a single depth image, our method first goes through the 3D volume branch to obtain a volumetric scene reconstruction as a guide to the next view inpainting step, which attempts to make up the missing information; the third step involves projecting the volume under the same view of the input, concatenating them to complete the current view depth, and integrating all depth into the point cloud. Since the occluded areas are unavailable, we resort to a deep Q-Network to glance around and pick the next best view for large hole completion progressively until a scene is adequately reconstructed while guaranteeing validity. All steps are learned jointly to achieve robust and consistent results. We perform qualitative and quantitative evaluations with extensive experiments on the SUNCG data, obtaining better results than the state of the art.

This paper focuses on the challenging task of learning 3D object surface reconstructions from single RGB images. Existing methods achieve varying degrees of success by using different geometric representations. However, they all have their own drawbacks, and cannot well reconstruct those surfaces of complex topologies. To this end, we propose in this paper a skeleton-bridged, stage-wise learning approach to address the challenge. Our use of skeleton is due to its nice property of topology preservation, while being of lower complexity to learn. To learn skeleton from an input image, we design a deep architecture whose decoder is based on a novel design of parallel streams respectively for synthesis of curve- and surface-like skeleton points. We use different shape representations of point cloud, volume, and mesh in our stage-wise learning, in order to take their respective advantages. We also propose multi-stage use of the input image to correct prediction errors that are possibly accumulated in each stage. We conduct intensive experiments to investigate the efficacy of our proposed approach. Qualitative and quantitative results on representative object categories of both simple and complex topologies demonstrate the superiority of our approach over existing ones. We will make our ShapeNet-Skeleton dataset publicly available.

Reconstructing the 3D mesh of a general object from a single image is now possible thanks to the latest advances of deep learning technologies. However, due to the nontrivial difficulty of generating a feasible mesh structure, the state-of-the-art approaches often simplify the problem by learning the displacements of a template mesh that deforms it to the target surface. Though reconstructing a 3D shape with complex topology can be achieved by deforming multiple mesh patches, it remains difficult to stitch the results to ensure a high meshing quality. In this paper, we present an end-to-end single-view mesh reconstruction framework that is able to generate high-quality meshes with complex topologies from a single genus-0 template mesh. The key to our approach is a novel progressive shaping framework that alternates between mesh deformation and topology modification. While a deformation network predicts the per-vertex translations that reduce the gap between the reconstructed mesh and the ground truth, a novel topology modification network is employed to prune the error-prone faces, enabling the evolution of topology. By iterating over the two procedures, one can progressively modify the mesh topology while achieving higher reconstruction accuracy. Moreover, a boundary refinement network is designed to refine the boundary conditions to further improve the visual quality of the reconstructed mesh. Extensive experiments demonstrate that our approach outperforms the current state-of-the-art methods both qualitatively and quantitatively, especially for the shapes with complex topologies.

Estimating 3D human pose from a single image is a challenging task. This work attempts to address the uncertainty of lifting the detected 2D joints to the 3D space by introducing an intermediate state - Part-Centric Heatmap Triplets (HEMlets), which shortens the gap between the 2D observation and the 3D interpretation. The HEMlets utilize three joint-heatmaps to represent the relative depth information of the end-joints for each skeletal body part. In our approach, a Convolutional Network(ConvNet) is first trained to predict HEMlests from the input image, followed by a volumetric joint-heatmap regression. We leverage on the integral operation to extract the joint locations from the volumetric heatmaps, guaranteeing end-to-end learning. Despite the simplicity of the network design, the quantitative comparisons show a significant performance improvement over the best-of-grade method (by 20% on Human3.6M). The proposed method naturally supports training with "in-the-wild'' images, where only weakly-annotated relative depth information of skeletal joints is available. This further improves the generalization ability of our model, as validated by qualitative comparisons on outdoor images.

Abstract Generating plausible hair image given limited guidance, such as sparse sketches or low‐resolution image, has been made possible with the rise of Generative Adversarial Networks (GANs). Traditional image‐to‐image translation networks can generate recognizable results, but finer textures are usually lost and blur artifacts commonly exist. In this paper, we propose a two‐phase generative model for high‐quality hair image synthesis. The two‐phase pipeline first generates a coarse image by an existing image translation model, then applies a re‐generating network with self‐enhancing capability to the coarse image. The self‐enhancing capability is achieved by a proposed differentiable layer, which extracts the structural texture and orientation maps from a hair image. Extensive experiments on two tasks, Sketch2Hair and Hair Super‐Resolution, demonstrate that our approach is able to synthesize plausible hair image with finer details, and reaches the state‐of‐the‐art.

This work presents a novel architecture of deep neural networks to generate meshes approximating the surface of a 3D object from a single image. Compared to existing learning-based 3D reconstruction models, our architecture is characterized by (1) deep mesh deformation stacks with residual network design, where a simple mesh is transformed to approximate the target surface and undergoes multiple deformation steps to progressively refine the result and reduce the residuals, and (2) parallel paths per deformation step, which can exponentially enrich the generated meshes using deeper structure and more model parameters. We also propose novel regularization scheme that encourages the meshes to be both globally complementary to cover the target surface and locally consistent with each other. Empirical evaluation on benchmark datasets show advantage of the proposed architecture over existing methods.

We propose a data-driven method for recovering missing parts of 3D shapes. Our method is based on a new deep learning architecture consisting of two sub-networks: a global structure inference network and a local geometry refinement network. The global structure inference network incorporates a long short-term memorized context fusion module (LSTM-CF) that infers the global structure of the shape based on multi-view depth information provided as part of the input. It also includes a 3D fully convolutional (3DFCN) module that further enriches the global structure representation according to volumetric information in the input. Under the guidance of the global structure network, the local geometry refinement network takes as input local 3D patches around missing regions, and progressively produces a high-resolution, complete surface through a volumetric encoder-decoder architecture. Our method jointly trains the global structure inference and local geometry refinement networks in an end-to-end manner. We perform qualitative and quantitative evaluations on six object categories, demonstrating that our method outperforms existing state-of-the-art work on shape completion.

Face modeling has been paid much attention in the field of visual computing. There exist many scenarios, including cartoon characters, avatars for social media, 3D face caricatures as well as face-related art and design, where low-cost interactive face modeling is a popular approach especially among amateur users. In this paper, we propose a deep learning based sketching system for 3D face and caricature modeling. This system has a labor-efficient sketching interface, that allows the user to draw freehand imprecise yet expressive 2D lines representing the contours of facial features. A novel CNN based deep regression network is designed for inferring 3D face models from 2D sketches. Our network fuses both CNN and shape based features of the input sketch, and has two independent branches of fully connected layers generating independent subsets of coefficients for a bilinear face representation. Our system also supports gesture based interactions for users to further manipulate initial face models. Both user studies and numerical results indicate that our sketching system can help users create face models quickly and effectively. A significantly expanded face database with diverse identities, expressions and levels of exaggeration is constructed to promote further research and evaluation of face modeling techniques.

Computing discrete geodesic distance over triangle meshes is one of the fundamental problems in computational geometry and computer graphics. In this problem, an effective window pruning strategy can significantly affect the actual running time. Due to its importance, we conduct an in-depth study of window pruning operations in this paper, and produce an exhaustive list of scenarios where one window can make another window partially or completely redundant. To identify a maximal number of redundant windows using such pairwise cross checking, we propose a set of procedures to synchronize local window propagation within the same triangle by simultaneously propagating a collection of windows from one triangle edge to its two opposite edges. On the basis of such synchronized window propagation, we design a new geodesic computation algorithm based on a triangle-oriented region growing scheme. Our geodesic algorithm can remove most of the redundant windows at the earliest possible stage, thus significantly reducing computational cost and memory usage at later stages. In addition, by adopting triangles instead of windows as the primitive in propagation management, our algorithm significantly cuts down the data management overhead. As a result, it runs 4--15 times faster than MMP and ICH algorithms, 2-4 times faster than FWP-MMP and FWP-CH algorithms, and also incurs the least memory usage.

Identifying sparse salient structures from dense pixels is a longstanding problem in visual computing. Solutions to this problem can benefit both image manipulation and understanding. In this paper, we introduce an image transform based on theL1norm for piecewise image flattening. This transform can effectively preserve and sharpen salient edges and contours while eliminating insignificant details, producing a nearly piecewise constant image with sparse structures. A variant of this image transform can perform edge-preserving smoothing more effectively than existing state-of-the-art algorithms. We further present a new method for complex scene-level intrinsic image decomposition. Our method relies on the above image transform to suppress surface shading variations, and perform probabilistic reflectance clustering on the flattened image instead of the original input image to achieve higher accuracy. Extensive testing on theIntrinsic-Images-in-the-Wilddatabase indicates our method can perform significantly better than existing techniques both visually and numerically. The obtained intrinsic images have been successfully used in two applications, surface retexturing and 3D object compositing in photographs.

We present an easy-to-use image retouching technique for realistic reshaping of human bodies in a single image. A model-based approach is taken by integrating a 3D whole-body morphable model into the reshaping process to achieve globally consistent editing effects. A novel body-aware image warping approach is introduced to reliably transfer the reshaping effects from the model to the image, even under moderate fitting errors. Thanks to the parametric nature of the model, our technique parameterizes the degree of reshaping by a small set of semantic attributes, such as weight and height. It allows easy creation of desired reshaping effects by changing the full-body attributes, while producing visually pleasing results even for loosely-dressed humans in casual photographs with a variety of poses and shapes.