Recent mesh generation approaches typically tokenize triangle meshes into sequences of tokens and train autoregressive models to generate these tokens sequentially. Despite substantial progress, such token sequences inevitably reuse vertices multiple times to fully represent manifold meshes, as each vertex is shared by multiple faces. This redundancy leads to excessively long token sequences and inefficient generation processes. In this paper, we propose an efficient framework that generates artistic meshes by treating vertices and faces separately, significantly reducing redundancy. We employ an autoregressive model solely for vertex generation, decreasing the token count to approximately 23% of that required by the most compact existing tokenizer. Next, we leverage a bidirectional transformer to complete the mesh in a single step by capturing intervertex relationships and constructing the adjacency matrix that defines the mesh faces. To further improve the generation quality, we introduce a fidelity enhancer to refine vertex positioning into more natural arrangements and propose a post-processing framework to remove undesirable edge connections. Experimental results show that our method achieves more than$8 \times$faster speed on mesh generation compared to state-of-the-art approaches, while producing higher mesh quality.

The field of neural rendering has seen remarkable progress, driven by advancements in generative models and differentiable rendering techniques. While 2D diffusion has achieved notable success, the development of a unified 3D diffusion pipeline remains an open challenge. This paper presents a novel framework, LN3Diff++, designed to bridge this gap and facilitate fast, high-quality, and versatile conditional 3D generation. Our method leverages a 3D-aware architecture and a variational autoencoder (VAE) to encode input image(s) into a structured, compact 3D latent space. The latent representation is then decoded by a transformer-based decoder into a high-capacity 3D neural field. By training a diffusion model on this 3D-aware latent space, our method achieves superior performance for category-specific 3D generation on ShapeNet and FFHQ, as well as category-free image/text-conditioned 3D generation over Objaverse. Moreover, it surpasses existing 3D diffusion methods in inference speed, requiring no per-instance optimization.

Representing the cutting-edge technique of text-to-image models, the latest Multimodal Diffusion Transformer (MMDiT) largely mitigates many generation issues existing in previous models. However, we discover that it still suffers from subject neglect or mixing when the input text prompt contains multiple subjects of similar semantics or appearance. We identify three possible ambiguities within the MMDiT architecture that cause this problem: Inter-block Ambiguity, Text Encoder Ambiguity, and Semantic Ambiguity. To address these issues, we propose to repair the ambiguous latent on-the-fly by test-time optimization at early denoising steps. In detail, we design three loss functions: Block Alignment Loss, Text Encoder Alignment Loss, and Overlap Loss, each tailored to mitigate these ambiguities. Despite significant improvements, we observe that semantic ambiguity persists when generating multiple similar subjects, as the guidance provided by overlap loss is not explicit enough. Therefore, we further propose Overlap Online Detection and Back-to-Start Sampling Strategy to alleviate the problem. Experimental results on a newly constructed challenging dataset of similar subjects validate the effectiveness of our approach, showing superior generation quality and much higher success rates over existing methods. The consistent and substantial improvements observed across multiple MMDiT based text-to-image models such as SD3, SD3.5 and FLUX provide strong evidence of the general applicability of our method.

Advances in 3D-aware generative models have pushed the boundary of image synthesis with explicit camera control. To achieve high-resolution image synthesis, several attempts have been made to design efficient generators, such as hybrid architectures with both 3D and 2D components. However, such a design compromises multiview consistency, and the design of a pure 3D generator with high resolution is still an open problem. In this work, we present Generative Volumetric Primitives (GVP), the first pure 3D volumetric generative model that can sample and render 512-resolution images in real-time. GVP Jointly models a number of volumetric primitives and their spatial information, both of which can be efficiently generated via a 2D convolutional network. The mixture of these primitives naturally captures the sparsity in the 3D volume. The training of such a generator with a high degree of freedom is made possible through a combination of adversarial and knowledge distillation training. The learned model exhibits dense 3D correspondences between samples. We provide exhaustive qualitative and qualitative evaluations for dense correspondences. Experiments on several datasets demonstrate superior efficiency, 3D consistency, and the emergence of dense correspondences of GVP over the state-of-the-art.

Latent Diffusion Models (LDMs) are known to have an unstable generation process, where even small perturbations or shifts in the input noise can lead to significantly different outputs. This hinders their applicability in applications requiring consistent results. In this work, we redesign LDMs to enhance consistency by making them shift-equivariant. While introducing anti-aliasing operations can partially improve shift-equivariance, significant aliasing and inconsistency persist due to the unique challenges in LDMs, including 1) aliasing amplification during VAE training and multiple U-Net inferences, and 2) selfattention modules that inherently lack shift-equivariance. To address these issues, we redesign the attention modules to be shift-equivariant and propose an equivariance loss that effectively suppresses the frequency bandwidth of the features in the continuous domain. The resulting alias-free LDM (AF-LDM) achieves strong shift-equivariance and is also robust to irregular warping. Extensive experiments demonstrate that AF-LDM produces significantly more consistent results than vanilla LDM across various applications, including video editing and image-to-image translation. Code is available at: https://github.com/SingleZombie/AFLDM

Autoregressive models have demonstrated remarkable success across various fields, from large language models (LLMs) to large multimodal models (LMMs) and 2D content generation, moving closer to artificial general intelligence (AGI). Despite these advances, applying autoregres-sive approaches to 3D object generation and understanding remains largely unexplored. This paper introduces Scale AutoRegressive 3D (SAR3D), a novel framework that leverages a multi-scale 3D vector-quantized variational autoencoder (VQVAE) to tokenize 3D objects for efficient autoregressive generation and detailed understanding. By predicting the next scale in a multi-scale latent representation instead of the next single token, SAR3D reduces generation time significantly, achieving fast 3D object generation in just 0.82 seconds on an A6000 GPU. Additionally, given the tokens enriched with hierarchical 3D-aware information, we finetune a pretrained LLM on them, enabling multimodal comprehension of 3D content. Our experiments show that SAR3D surpasses current 3D generation methods in both speed and quality and allows LLMs to interpret and caption 3D models comprehensively.

Recovering the geometry and materials of objects from a single image is challenging due to its under-constrained nature. In this paper, we present Neural LightRig, a novel framework that boosts intrinsic estimation by leveraging auxiliary multi-lighting conditions from 2D diffusion priors. Specifically, 1) we first leverage illumination priors from large-scale diffusion models to build our multi-light diffusion model on a synthetic relighting dataset with dedicated designs. This diffusion model generates multiple consistent images, each illuminated by point light sources in different directions. 2) By using these varied lighting images to reduce estimation uncertainty, we train a large G-buffer model with a U-Net backbone to accurately predict surface normals and materials. Extensive experiments validate that our approach significantly outperforms state-of-the-art methods, enabling accurate surface normal and PBR material estimation with vivid relighting effects. Code and dataset are available on our project page at https://projects.zxhezexin.com/neural-lightrig.

Despite advances in neural rendering, due to the scarcity of high-quality 3D datasets and the inherent limitations of multi-view diffusion models, view synthesis and 3D model generation are restricted to low resolutions with suboptimal multi-view consistency. In this study, we present a novel 3D enhancement pipeline, dubbed 3DENHANCER, which employs a multi-view latent diffusion model to enhance coarse 3D inputs while preserving multi-view consistency. Our method includes a pose-aware encoder and a diffusion-based denoiser to refine low-quality multi-view images, along with data augmentation and a multi-view attention module with epipolar aggregation to maintain consistent, high-quality 3D outputs across views. Unlike existing video-based approaches, our model supports seamless multi-view enhancement with improved coherence across diverse viewing angles. Extensive evaluations show that 3DENHANCER significantly outperforms existing methods, boosting both multi-view enhancement and per-instance 3D optimization tasks.

Generating consistent long videos is a complex challenge: while diffusion-based generative models generate visually impressive short clips, extending them to longer durations often leads to memory bottlenecks and long-term inconsistency. In this paper, we propose TokensGen, a novel two-stage framework that leverages condensed tokens to address these issues. Our method decomposes long video generation into three core tasks: (1) inner-clip semantic control, (2) long-term consistency control, and (3) inter-clip smooth transition. First, we train To2V (Token-to-Video), a short video diffusion model guided by text and video tokens, with a Video Tokenizer that condenses short clips into semantically rich tokens. Second, we introduce T2To (Text-to-Token), a video token diffusion transformer that generates all tokens at once, ensuring global consistency across clips. Finally, during inference, an adaptive FIFO-Diffusion strategy seamlessly connects adjacent clips, reducing boundary artifacts and enhancing smooth transitions. Experimental results demonstrate that our approach significantly enhances long-term temporal and content coherence without incurring prohibitive computational overhead. By leveraging condensed tokens and pre-trained short video models, our method provides a scalable, modular solution for long video generation, opening new possibilities for storytelling, cinematic production, and immersive simulations. Please see our project page at https://vicky0522.github.io/tokensgen-webpage/ .

The integration of Rotary Position Embedding (RoPE) in Multimodal Diffusion Transformer (MMDiT) has significantly enhanced text-to-image generation quality. However, the fundamental reliance of self-attention layers on positional embedding versus query-key similarity during generation remains an intriguing question. We present the first mechanistic analysis of RoPE-based MMDiT models (e.g., FLUX), introducing an automated probing strategy that disentangles positional information versus content dependencies by strategically manipulating RoPE during generation. Our analysis reveals distinct dependency patterns that do not straightforwardly correlate with depth, offering new insights into the layer-specific roles in RoPE-based MMDiT. Based on these findings, we propose a training-free, task-specific image editing framework that categorizes editing tasks into three types: position-dependent editing (e.g., object addition), content similarity-dependent editing (e.g., non-rigid editing), and region-preserved editing (e.g., background replacement). For each type, we design tailored key-value injection strategies based on the characteristics of the editing task. Extensive qualitative and quantitative evaluations demonstrate that our method outperforms state-of-the-art approaches, particularly in preserving original semantic content and achieving seamless modifications.

Textured 3D morphing creates smooth and plausible interpolation sequences between two 3D objects, focusing on transitions in both shape and texture. This is important for creative applications like visual effects in filmmaking. Previous methods rely on establishing point-to-point correspondences and determining smooth deformation trajectories, which inherently restrict them to shape-only morphing on untextured, topologically aligned datasets. This restriction leads to labor-intensive preprocessing and poor generalization. To overcome these challenges, we propose a method for 3D regenerative morphing using a 3D diffusion prior. Unlike previous methods that depend on explicit correspondences and deformations, our method eliminates the additional need for obtaining correspondence and uses the 3D diffusion prior to generate morphing. Specifically, we introduce a 3D diffusion model and interpolate the source and target information at three levels: initial noise, model parameters, and condition features. We then explore an Attention Fusion strategy to generate more smooth morphing sequences. To further improve the plausibility of semantic interpolation and the generated 3D surfaces, we propose two strategies: (a) Token Reordering, where we match approximate tokens based on semantic analysis to guide implicit correspondences in the denoising process of the diffusion model, and (b) Low-Frequency Enhancement, where we enhance low-frequency signals in the tokens to improve the quality of generated surfaces. Experimental results show that our method achieves superior smoothness and plausibility in 3D morphing across diverse cross-category object pairs, offering a novel regenerative method for 3D morphing with textured representations.

Diffusion models have demonstrated exceptional visual quality in video generation, making them promising for autonomous driving world modeling. However, existing video diffusion-based world models struggle with flexible-length, long-horizon predictions and integrating trajectory planning. This is because conventional video diffusion models rely on global joint distribution modeling of fixed-length frame sequences rather than sequentially constructing localized distributions at each timestep. In this work, we propose Epona, an autoregressive diffusion world model that enables localized spatiotemporal distribution modeling through two key innovations: 1) Decoupled spatiotemporal factorization that separates temporal dynamics modeling from fine-grained future world generation, and 2) Modular trajectory and video prediction that seamlessly integrate motion planning with visual modeling in an end-to-end framework. Our architecture enables high-resolution, long-duration generation while introducing a novel chain-of-forward training strategy to address error accumulation in autoregressive loops. Experimental results demonstrate state-of-the-art performance with 7.4\% FVD improvement and minutes longer prediction duration compared to prior works. The learned world model further serves as a real-time motion planner, outperforming strong end-to-end planners on NAVSIM benchmarks. Code will be publicly available at \href{https://github.com/Kevin-thu/Epona/}{https://github.com/Kevin-thu/Epona/}.

Recent advancements in diffusion models and large-scale datasets have revolutionized image and video generation, with increasing focus on 3D content generation. While existing methods show promise, they face challenges in input formats, latent space structures, and output representations. This paper introduces a novel 3D generation framework that addresses these issues, enabling scalable and high-quality 3D generation with an interactive Point Cloud-structured Latent space. Our approach utilizes a VAE with multi-view posed RGB-D-N renderings as input, features a unique latent space design that preserves 3D shape information, and incorporates a cascaded latent flow-based model for improved shape-texture disentanglement. The proposed method, GaussianAnything, supports multi-modal conditional 3D generation, allowing for point cloud, caption, and single-view image inputs. Experimental results demonstrate superior performance on various datasets, advancing the state-of-the-art in 3D content generation.

Recent advancements in video generation have been greatly driven by video diffusion models, with camera motion control emerging as a crucial challenge in creating view-customized visual content. This paper introduces trajectory attention, a novel approach that performs attention along available pixel trajectories for fine-grained camera motion control. Unlike existing methods that often yield imprecise outputs or neglect temporal correlations, our approach possesses a stronger inductive bias that seamlessly injects trajectory information into the video generation process. Importantly, our approach models trajectory attention as an auxiliary branch alongside traditional temporal attention. This design enables the original temporal attention and the trajectory attention to work in synergy, ensuring both precise motion control and new content generation capability, which is critical when the trajectory is only partially available. Experiments on camera motion control for images and videos demonstrate significant improvements in precision and long-range consistency while maintaining high-quality generation. Furthermore, we show that our approach can be extended to other video motion control tasks, such as first-frame-guided video editing, where it excels in maintaining content consistency over large spatial and temporal ranges.

World simulation has gained increasing popularity due to its ability to model virtual environments and predict the consequences of actions. However, the limited temporal context window often leads to failures in maintaining long-term consistency, particularly in preserving 3D spatial consistency. In this work, we present WorldMem, a framework that enhances scene generation with a memory bank consisting of memory units that store memory frames and states (e.g., poses and timestamps). By employing state-aware memory attention that effectively extracts relevant information from these memory frames based on their states, our method is capable of accurately reconstructing previously observed scenes, even under significant viewpoint or temporal gaps. Furthermore, by incorporating timestamps into the states, our framework not only models a static world but also captures its dynamic evolution over time, enabling both perception and interaction within the simulated world. Extensive experiments in both virtual and real scenarios validate the effectiveness of our approach.

We propose ArtiLatent, a generative framework that synthesizes human-made 3D objects with fine-grained geometry, accurate articulation, and realistic appearance. Our approach jointly models part geometry and articulation dynamics by embedding sparse voxel representations and associated articulation properties—including joint type, axis, origin, range, and part category—into a unified latent space via a variational autoencoder. A latent diffusion model is then trained over this space to enable diverse yet physically plausible sampling. To reconstruct photorealistic 3D shapes, we introduce an articulation-aware Gaussian decoder that accounts for articulation-dependent visibility changes (e.g., revealing the interior of a drawer when opened). By conditioning appearance decoding on articulation state, our method assigns plausible texture features to regions that are typically occluded in static poses, significantly improving visual realism across articulation configurations. Extensive experiments on furniture-like objects from PartNet-Mobility and ACD datasets demonstrate that ArtiLatent outperforms existing approaches in geometric consistency and appearance fidelity. Our framework provides a scalable solution for articulated 3D object synthesis and manipulation.

Recent advances in large-scale text-to-image models have revolutionized creative fields by generating visually captivating outputs from textual prompts; however, while traditional photography offers precise control over camera settings to shape visual aesthetics—such as depth-of-field via aperture—current diffusion models typically rely on prompt engineering to mimic such effects. This approach often results in crude approximations and inadvertently alters the scene content. In this work, we propose Bokeh Diffusion, a scene-consistent bokeh control framework that explicitly conditions a diffusion model on a physical defocus blur parameter. To overcome the scarcity of paired real-world images captured under different camera settings, we introduce a hybrid training pipeline that aligns in-the-wild images with synthetic blur augmentations, providing diverse scenes and subjects as well as supervision to learn the separation of image content from lens blur. Central to our framework is our grounded self-attention mechanism, trained on image pairs with different bokeh levels of the same scene, which enables blur strength to be adjusted in both directions while preserving the underlying scene. Extensive experiments demonstrate that our approach enables flexible, lens-like blur control, supports downstream applications such as real image editing via inversion, and generalizes effectively across both Stable Diffusion and FLUX architectures.

Despite the emergence of successful NeRF inpainting methods built upon explicit RGB and depth 2D inpainting supervisions, these methods are inherently constrained by the capabilities of their underlying 2D inpainters. This is due to two key reasons: (i) independently inpainting constituent images results in view-inconsistent imagery, and (ii) 2D inpainters struggle to ensure high-quality geometry completion and alignment with inpainted RGB images. To overcome these limitations, we propose a novel approach called MVIP-NeRF that harnesses the potential of diffusion priors for NeRF inpainting, addressing both appearance and geometry aspects. MVIP-NeRF performs joint inpainting across multiple views to reach a consistent solution, which is achieved via an iterative optimization process based on Score Distillation Sampling (SDS). Apart from recovering the rendered RGB images, we also extract normal maps as a geometric representation and define a normal SDS loss that motivates accurate geometry inpaint- ing and alignment with the appearance. Additionally, we formulate a multi-view SDS score function to distill generative priors simultaneously from different view images, ensuring consistent visual completion when dealing with large view variations. Our experimental results show better appearance and geometry recovery than previous NeRF in painting methods.

Diffusion models have achieved remarkable image generation quality surpassing previous generative models. However, a notable limitation of diffusion models, in comparison to GANs, is their difficulty in smoothly interpolating between two image samples, due to their highly unstructured latent space. Such a smooth interpolation is intriguing as it naturally serves as a solution for the image mor-phing task with many applications. In this work, we address this limitation via DiffMorpher, an approach that enables smooth and natural image interpolation by harnessing the prior knowledge of a pretrained diffusion model. Our key idea is to capture the semantics of the two images by fitting two LoRAs to them respectively, and interpolate between both the LoRA parameters and the latent noises to ensure a smooth semantic transition, where correspon-dence automatically emerges without the need for annotation. In addition, we propose an attention interpolation and injection technique, an adaptive normalization adjustment method, and a new sampling schedule to further enhance the smoothness between consecutive images. Extensive experiments demonstrate that DiffMorpher achieves starkly better image morphing effects than previous methods across a variety of object categories, bridging a critical functional gap that distinguished diffusion models from GANs.

Video generation primarily aims to model authentic and customized motion across frames, making understanding and controlling the motion a crucial topic. Most diffusion-based studies on video motion focus on motion customization with training-based paradigms, which, however, demands substantial training resources and necessitates retraining for diverse models. Crucially, these approaches do not explore how video diffusion models encode cross-frame motion information in their features, lacking interpretability and transparency in their effectiveness. To answer this question, this paper introduces a novel perspective to understand, localize, and manipulate motion-aware features in video diffusion models. Through analysis using Principal Component Analysis (PCA), our work discloses that robust motion-aware feature already exists in video diffusion models. We present a new MOtion FeaTure (MOFT) by eliminating content correlation information and filtering motion channels. MOFT provides a distinct set of benefits, including the ability to encode comprehensive motion information with clear interpretability, extraction without the need for training, and generalizability across diverse architectures. Leveraging MOFT, we propose a novel training-free video motion control framework. Our method demonstrates competitive performance in generating natural and faithful motion, providing architecture-agnostic insights and applicability in a variety of downstream tasks.

The remarkable generative capabilities of diffusion models have motivated extensive research in both image and video editing. Compared to video editing which faces additional challenges in the time dimension, image editing has witnessed the development of more diverse, high-quality approaches and more capable software like Photoshop. In light of this gap, we introduce a novel and generic solution that extends the applicability of image editing tools to videos by propagating edits from a single frame to the entire video using a pre-trained image-to-video model. Our method, dubbed I2VEdit, adaptively preserves the visual and motion integrity of the source video depending on the extent of the edits, effectively handling global edits, local edits, and moderate shape changes, which existing methods cannot fully achieve. At the core of our method are two main processes: Coarse Motion Extraction to align basic motion patterns with the original video, and Appearance Refinement for precise adjustments using fine-grained attention matching. We also incorporate a skip-interval strategy to mitigate quality degradation from auto-regressive generation across multiple video clips. Experimental results demonstrate our framework’s superior performance in fine-grained video editing, proving its capability to produce high-quality, temporally consistent outputs. Our website is at https://i2vedit.github.io/.

Event cameras report events whenever an individual pixel changes brightness. The discrete and asynchronous nature of events makes recovering pixel brightness signals a challenging task, even if conventional brightness frames are recorded along with events. Recent works have addressed this task with neural networks, which tend to be biased towards their training distribution. All methods need to deal with noise in the events to produce very high output framerates. We introduce a new approach to event-based reconstruction, not learning-based: Our model assigns each event an explicit confidence weight to account for the uncertainty arising from noise. We also introduce a novel loss term to balance confidences against each other and show that interpolation of brightness signals between events can benefit from Bézier curves. We demonstrate that allowing brightness changes between exposures can improve reconstruction quality. Our evaluation shows that our method improves the state of the art in the tasks of event-based de-blurring and event-based frame interpolation.

Facial editing is to manipulate the facial attributes of a given face image. Nowadays, with the development of generative models, users can easily generate 2D and 3D facial images with high fidelity and 3D-aware consistency. However, existing works are incapable of delivering a continuous and fine-grained editing mode (e.g., editing a slightly smiling face to a big laughing one) with natural interactions with users. In this work, we propose Talk-to-Edit, an interactive facial editing framework that performs fine-grained attribute manipulation through dialog between the user and the system. Our key insight is to model a continual "semantic field" in the GAN latent space. 1) Unlike previous works that regard the editing as traversing straight lines in the latent space, here the fine-grained editing is formulated as finding a curving trajectory that respects fine-grained attribute landscape on the semantic field. 2) The curvature at each step is location-specific and determined by the input image as well as the users' language requests. 3) To engage the users in a meaningful dialog, our system generates language feedback by considering both the user request and the current state of the semantic field. We demonstrate the effectiveness of our proposed framework on both 2D and 3D-aware generative models. We term the semantic field for the 3D-aware models as "tri-plane" flow, as it corresponds to the changes not only in the color space but also in the density space. We also contribute CelebA-Dialog, a visual-language facial editing dataset to facilitate large-scale study. Specifically, each image has manually annotated fine-grained attribute annotations as well as template-based textual descriptions in natural language. Extensive quantitative and qualitative experiments demonstrate the superiority of our framework in terms of 1) the smoothness of fine-grained editing, 2) the identity/attribute preservation, and 3) the visual photorealism and dialog fluency. Notably, the user study validates that our overall system is consistently favored by around 80% of the participants.

Multi-view volumetric rendering techniques have recently shown great potential in modeling and synthesizing high-quality head avatars. A common approach to capture full head dynamic performances is to track the underlying geometry using a mesh-based template or 3D cube-based graphics primitives. While these model-based approaches achieve promising results, they often fail to learn complex geometric details such as the mouth interior, hair, and topological changes over time. This article presents a novel approach to building highly photorealistic digital head avatars. Our method learns a canonical space via an implicit function parameterized by a neural network. It leverages multiresolution hash encoding in the learned feature space, allowing for high quality, faster training, and high-resolution rendering. At test time, our method is driven by a monocular RGB video. Here, an image encoder extracts face-specific features that also condition the learnable canonical space. This encourages deformation-dependent texture variations during training. We also propose a novel optical flow-based loss that ensures correspondences in the learned canonical space, thus encouraging artifact-free and temporally consistent renderings. We show results on challenging facial expressions and show free-viewpoint renderings at interactive real-time rates for a resolution of 480 x 270. Our method outperforms related approaches both visually and numerically. We will release our multiple-identity dataset to encourage further research.

The real world exhibits rich structure and detail across many scales of observation. It is difficult, however, to capture and represent a broad spectrum of scales using ordinary images. We devise a novel paradigm for learning a representation that captures an orders-of-magnitude variety of scales from an unstructured collection of ordinary images. We treat this collection as a distribution of scale-space slices to be learned using adversarial training, and additionally enforce coherency across slices. Our approach relies on a multiscale generator with carefully injected procedural frequency content, which allows to interactively explore the emerging continuous scale space. Training across vastly different scales poses challenges regarding stability, which we tackle using a supervision scheme that involves careful sampling of scales. We show that our generator can be used as a multiscale generative model, and for reconstructions of scale spaces from unstructured patches. Significantly outperforming the state of the art, we demonstrate zoom-in factors of up to 256x at high quality and scale consistency.

Purely MLP-based neural radiance fields (NeRF-based methods) often suffer from underfitting with blurred renderings on large-scale scenes due to limited model capacity. Recent approaches propose to geographically divide the scene and adopt multiple sub-NeRFs to model each region individually, leading to linear scale-up in training costs and the number of sub-NeRFs as the scene expands. An alternative solution is to use a feature grid representation, which is computationally efficient and can naturally scale to a large scene with increased grid resolutions. However, the feature grid tends to be less constrained and often reaches suboptimal solutions, producing noisy artifacts in renderings, especially in regions with complex geometry and texture. In this work, we present a new framework that realizes high-fidelity rendering on large urban scenes while being computationally efficient. We propose to use a compact multi-resolution ground feature plane representation to coarsely capture the scene, and complement it with positional encoding inputs through another NeRF branch for rendering in a joint learning fashion. We show that such an integration can utilize the advantages of two alternative solutions: a light-weighted NeRF is sufficient, under the guidance of the feature grid representation, to render photorealistic novel views with fine details; and the jointly optimized ground feature planes, can meanwhile gain further refinements, forming a more accurate and compact feature space and output much more natural rendering results.

Most in-the-wild images are stored in Low Dynamic Range (LDR) form, serving as a partial observation of the High Dynamic Range (HDR) visual world. Despite limited dynamic range, these LDR images are often captured with different exposures, implicitly containing information about the underlying HDR image distribution. Inspired by this intuition, in this work we present, to the best of our knowledge, the first method for learning a generative model of HDR images from in-the-wild LDR image collections in a fully unsupervised manner. The key idea is to train a generative adversarial network (GAN) to generate HDR images which, when projected to LDR under various exposures, are indistinguishable from real LDR images. The projection from HDR to LDR is achieved via a camera model that captures the stochasticity in exposure and camera response function. Experiments show that our method GlowGAN can synthesize photorealistic HDR images in many challenging cases such as landscapes, lightning, or windows, where previous supervised generative models produce overexposed images. With the assistance of GlowGAN, we showcase the novel application of unsupervised inverse tone mapping (GlowGAN-ITM) that sets a new paradigm in this field. Unlike previous methods that gradually complete information from LDR input, GlowGAN-ITM searches the entire HDR image manifold modeled by GlowGAN for the HDR images which can be mapped back to the LDR input. GlowGAN-ITM achieves more realistic reconstruction of overexposed regions compared to state-of-the-art supervised learning models, despite not requiring HDR images or paired multi-exposure images for training.

Both indoor and outdoor environments are inherently structured and repetitive. Traditional modeling pipelines keep an asset library storing unique object templates, which is both versatile and memory efficient in practice. Inspired by this observation, we propose AssetField, a novel neural scene representation that learns a set of object-aware ground feature planes to represent the scene, where an asset library storing template feature patches can be constructed in an unsupervised manner. Unlike existing methods which require object masks to query spatial points for object editing, our ground feature plane representation offers a natural visualization of the scene in the bird-eye view, allowing a variety of operations (e.g. translation, duplication, deformation) on objects to configure a new scene. With the template feature patches, group editing is enabled for scenes with many recurring items to avoid repetitive work on object individuals. We show that AssetField not only achieves competitive performance for novel-view synthesis but also generates realistic renderings for new scene configurations.

3D content creation from a single image is a long-standing yet highly desirable task. Recent advances introduce 2D diffusion priors, yielding reasonable results. However, existing methods are not hyper-realistic enough for post-generation usage, as users cannot view, render and edit the resulting 3D content from a full range. To address these challenges, we introduce HyperDreamer with several key designs and appealing properties: 1) Full-range viewable: 360° mesh modeling with high-resolution textures enables the creation of visually compelling 3D models from a full range of observation points. 2) Full-range renderable: Fine-grained semantic segmentation and data-driven priors are incorporated as guidance to learn reasonable albedo, roughness, and specular properties of the materials, enabling semantic-aware arbitrary material estimation. 3) Full-range editable: For a generated model or their own data, users can interactively select any region via a few clicks and efficiently edit the texture with text-based guidance. Extensive experiments demonstrate the effectiveness of HyperDreamer in modeling region-aware materials with high-resolution textures and enabling user-friendly editing. We believe that HyperDreamer holds promise for advancing 3D content creation and finding applications in various domains.

Capturing and editing full-head performances enables the creation of virtual characters with various applications such as extended reality and media production. The past few years witnessed a steep rise in the photorealism of human head avatars. Such avatars can be controlled through different input data modalities, including RGB, audio, depth, IMUs, and others. While these data modalities provide effective means of control, they mostly focus on editing the head movements such as the facial expressions, head pose, and/or camera viewpoint. In this paper, we propose AvatarStudio, a text-based method for editing the appearance of a dynamic full head avatar. Our approach builds on existing work to capture dynamic performances of human heads using Neural Radiance Field (NeRF) and edits this representation with a text-to-image diffusion model. Specifically, we introduce an optimization strategy for incorporating multiple keyframes representing different camera viewpoints and time stamps of a video performance into a single diffusion model. Using this personalized diffusion model, we edit the dynamic NeRF by introducing view-and-time-aware Score Distillation Sampling (VT-SDS) following a model-based guidance approach. Our method edits the full head in a canonical space and then propagates these edits to the remaining time steps via a pre-trained deformation network. We evaluate our method visually and numerically via a user study, and results show that our method outperforms existing approaches. Our experiments validate the design choices of our method and highlight that our edits are genuine, personalized, as well as 3D- and time-consistent.

In everyday photography, physical limitations of camera sensors and lenses frequently lead to a variety of degradations in captured images such as saturation or defocus blur. A common approach to overcome these limitations is to resort to image stack fusion, which involves capturing multiple images with different focal distances or exposures. For instance, to obtain an all-in-focus image, a set of multi-focus images is captured. Similarly, capturing multiple exposures allows for the reconstruction of high dynamic range. In this paper, we present a novel approach that combines neural fields with an expressive camera model to achieve a unified reconstruction of an all-in-focus high-dynamic-range image from an image stack. Our approach is composed of a set of specialized implicit neural representations tailored to address specific sub-problems along our pipeline: We use neural implicits to predict flow to overcome misalignments arising from lens breathing, depth, and all-in-focus images to account for depth of field, as well as tonemapping to deal with sensor responses and saturation - all trained using a physically inspired supervision structure with a differentiable thin lens model at its core. An important benefit of our approach is its ability to handle these tasks simultaneously or independently, providing flexible post-editing capabilities such as refocusing and exposure adjustment. By sampling the three primary factors in photography within our framework (focal distance, aperture, and exposure time), we conduct a thorough exploration to gain valuable insights into their significance and impact on overall reconstruction quality. Through extensive validation, we demonstrate that our method outperforms existing approaches in both depth-from-defocus and all-in-focus image reconstruction tasks. Moreover, our approach exhibits promising results in each of these three dimensions, showcasing its potential to enhance captured image quality and provide greater control in post-processing.

2D images are observations of the 3D physical world depicted with the geometry, material, and illumination components. Recovering these underlying intrinsic components from 2D images, also known as inverse rendering, usually requires a supervised setting with paired images collected from multiple viewpoints and lighting conditions, which is resource-demanding. In this work, we present GAN2X, a new method for unsupervised inverse rendering that only uses unpaired images for training. Unlike previous Shape-from-GAN approaches that mainly focus on 3D shapes, we take the first attempt to also recover non-Lambertian material properties by exploiting the pseudo paired data generated by a GAN. To achieve precise inverse rendering, we devise a specularity-aware neural surface representation that continuously models the geometry and material properties. A shading-based refinement technique is adopted to further distill information in the target image and recover more fine details. Experiments demonstrate that GAN2X can accurately decompose 2D images to 3D shape, albedo, and specular properties for different object categories, and achieves state-of-the-art performance for unsupervised single-view 3D face reconstruction. We also show its applications in downstream tasks including real image editing and lifting 2D GANs to decomposed 3D GANs.

3D generative models of objects enable photorealistic image synthesis with 3D control. Existing methods model the scene as a global scene representation, ignoring the compositional aspect of the scene. Compositional reasoning can enable a wide variety of editing applications, in addition to enabling generalizable 3D reasoning. In this paper, we present a compositional generative model, where each semantic part of the object is represented as an independent 3D representation learnt from only in-the-wild 2D data. We start with a global generative model (GAN) and learn to decompose it into different semantic parts using supervision from 2D segmentation masks. We then learn to composite independently sampled parts in order to create coherent global scenes. Different parts can be independently sampled, while keeping rest of the object fixed. We evaluate our method on a wide variety of objects and parts, and demonstrate editing applications.

Learning 3D generative models from a dataset of monocular images enables self-supervised 3D reasoning and controllable synthesis. State-of-the-art 3D generative models are GANs that use neural 3D volumetric representations for synthesis. Images are synthesized by rendering the volumes from a given camera. These models can disentangle the 3D scene from the camera viewpoint in any generated image. However, most models do not disentangle other factors of image formation, such as geometry and appearance. In this paper, we design a 3D GAN which can learn a disentangled model of objects, just from monocular observations. Our model can disentangle the geometry and appearance variations in the scene, i.e., we can independently sample from the geometry and appearance spaces of the generative model. This is achieved using a novel non-rigid deformable scene formulation. A 3D volume that represents an object instance is computed as a non-rigidly deformed canonical 3D volume. Our method learns the canonical volume, as well as its deformations, jointly during training. This formulation also helps us improve the disentanglement between the 3D scene and the camera viewpoints using a novel pose regularization loss defined on the 3D deformation field. In addition, we model the inverse deformations, enabling the computation of dense correspondences between images generated by our model. Finally, we design an approach to embed real images into the latent space of our model, enabling editing of real images.

Learning a good image prior is a long-term goal for image restoration and manipulation. While existing methods like deep image prior (DIP) capture low-level image statistics, there are still gaps toward an image prior that captures rich image semantics including color, spatial coherence, textures, and high-level concepts. This work presents an effective way to exploit the image prior captured by a generative adversarial network (GAN) trained on large-scale natural images. As shown in Fig. 1, the deep generative prior (DGP) provides compelling results to restore missing semantics, e.g., color, patch, resolution, of various degraded images. It also enables diverse image manipulation including random jittering, image morphing, and category transfer. Such highly flexible restoration and manipulation are made possible through relaxing the assumption of existing GAN inversion methods, which tend to fix the generator. Notably, we allow the generator to be fine-tuned on-the-fly in a progressive manner regularized by feature distance obtained by the discriminator in GAN. We show that these easy-to-implement and practical changes help preserve the reconstruction to remain in the manifold of nature images, and thus lead to more precise and faithful reconstruction for real images. Code is available at https://github.com/XingangPan/deep-generative-prior.

Facial editing is an important task in vision and graphics with numerous applications. However, existing works are incapable to deliver a continuous and fine-grained editing mode (e.g., editing a slightly smiling face to a big laughing one) with natural interactions with users. In this work, we propose Talk-to-Edit, an interactive facial editing framework that performs fine-grained attribute manipulation through dialog between the user and the system. Our key insight is to model a continual "semantic field" in the GAN latent space. 1) Unlike previous works that regard the editing as traversing straight lines in the latent space, here the fine-grained editing is formulated as finding a curving trajectory that respects fine-grained attribute landscape on the semantic field. 2) The curvature at each step is location-specific and determined by the input image as well as the users’ language requests. 3) To engage the users in a meaningful dialog, our system generates language feedback by considering both the user request and the current state of the semantic field.We also contribute CelebA-Dialog, a visual-language facial editing dataset to facilitate large-scale study. Specifically, each image has manually annotated fine-grained attribute annotations as well as template-based textual descriptions in natural language. Extensive quantitative and qualitative experiments demonstrate the superiority of our framework in terms of 1) the smoothness of fine-grained editing, 2) the identity/attribute preservation, and 3) the visual photorealism and dialog fluency. Notably, user study validates that our overall system is consistently favored by around 80% of the participants. Our project page is https://www.mmlab-ntu.com/project/talkedit/.

The advancement of generative radiance fields has pushed the boundary of 3D-aware image synthesis. Motivated by the observation that a 3D object should look realistic from multiple viewpoints, these methods introduce a multi-view constraint as regularization to learn valid 3D radiance fields from 2D images. Despite the progress, they often fall short of capturing accurate 3D shapes due to the shape-color ambiguity, limiting their applicability in downstream tasks. In this work, we address this ambiguity by proposing a novel shading-guided generative implicit model that is able to learn a starkly improved shape representation. Our key insight is that an accurate 3D shape should also yield a realistic rendering under different lighting conditions. This multi-lighting constraint is realized by modeling illumination explicitly and performing shading with various lighting conditions. Gradients are derived by feeding the synthesized images to a discriminator. To compensate for the additional computational burden of calculating surface normals, we further devise an efficient volume rendering strategy via surface tracking, reducing the training and inference time by 24% and 48%, respectively. Our experiments on multiple datasets show that the proposed approach achieves photorealistic 3D-aware image synthesis while capturing accurate underlying 3D shapes. We demonstrate improved performance of our approach on 3D shape reconstruction against existing methods, and show its applicability on image relighting. Our code is available at https://github.com/XingangPan/ShadeGAN.

The advent of generative radiance fields has significantly promoted the development of 3D-aware image synthesis. The cumulative rendering process in radiance fields makes training these generative models much easier since gradients are distributed over the entire volume, but leads to diffused object surfaces. In the meantime, compared to radiance fields occupancy representations could inherently ensure deterministic surfaces. However, if we directly apply occupancy representations to generative models, during training they will only receive sparse gradients located on object surfaces and eventually suffer from the convergence problem. In this paper, we propose Generative Occupancy Fields (GOF), a novel model based on generative radiance fields that can learn compact object surfaces without impeding its training convergence. The key insight of GOF is a dedicated transition from the cumulative rendering in radiance fields to rendering with only the surface points as the learned surface gets more and more accurate. In this way, GOF combines the merits of two representations in a unified framework. In practice, the training-time transition of start from radiance fields and march to occupancy representations is achieved in GOF by gradually shrinking the sampling region in its rendering process from the entire volume to a minimal neighboring region around the surface. Through comprehensive experiments on multiple datasets, we demonstrate that GOF can synthesize high-quality images with 3D consistency and simultaneously learn compact and smooth object surfaces. Our code is available at https://github.com/SheldonTsui/GOF_NeurIPS2021.

A typical domain adaptation approach is to adapt models trained on the annotated data in a source domain (e.g., sunny weather) for achieving high performance on the test data in a target domain (e.g., rainy weather). Whether the target contains a single homogeneous domain or multiple heterogeneous domains, existing works always assume that there exist clear distinctions between the domains, which is often not true in practice (e.g., changes in weather). We study an open compound domain adaptation (OCDA) problem, in which the target is a compound of multiple homogeneous domains without domain labels, reflecting realistic data collection from mixed and novel situations. We propose a new approach based on two technical insights into OCDA: 1) a curriculum domain adaptation strategy to bootstrap generalization across domains in a data-driven self-organizing fashion and 2) a memory module to increase the model's agility towards novel domains. Our experiments on digit classification, facial expression recognition, semantic segmentation, and reinforcement learning demonstrate the effectiveness of our approach.

Natural scene understanding is a challenging task, particularly when encountering images of multiple objects that are partially occluded. This obstacle is given rise by varying object ordering and positioning. Existing scene understanding paradigms are able to parse only the visible parts, resulting in incomplete and unstructured scene interpretation. In this paper, we investigate the problem of scene de-occlusion, which aims to recover the underlying occlusion ordering and complete the invisible parts of occluded objects. We make the first attempt to address the problem through a novel and unified framework that recovers hidden scene structures without ordering and amodal annotations as supervisions. This is achieved via Partial Completion Network (PCNet)-mask (M) and -content (C), that learn to recover fractions of object masks and contents, respectively, in a self-supervised manner. Based on PCNet-M and PCNet-C, we devise a novel inference scheme to accomplish scene de-occlusion, via progressive ordering recovery, amodal completion and content completion. Extensive experiments on real-world scenes demonstrate the superior performance of our approach to other alternatives. Remarkably, our approach that is trained in a self-supervised manner achieves comparable results to fully-supervised methods. The proposed scene de-occlusion framework benefits many applications, including high-quality and controllable image manipulation and scene recomposition (see Fig. 1), as well as the conversion of existing modal mask annotations to amodal mask annotations.

Convolutional Neural Networks (CNNs) are typically constructed by stacking multiple building blocks, each of which contains a normalization layer such as batch normalization (BN) and a rectified linear function such as ReLU. However, this work shows that the combination of normalization and rectified linear function leads to inhibited channels, which have small magnitude and contribute little to the learned feature representation, impeding the generalization ability of CNNs. Unlike prior arts that simply removed the inhibited channels, we propose to “wake them up” during training by designing a novel neural building block, termed Channel Equilibrium (CE) block, which enables channels at the same layer to contribute equally to the learned representation. We show that CE is able to prevent inhibited channels both empirically and theoretically. CE has several appealing benefits. (1) It can be integrated into many advanced CNN architectures such as ResNet and MobileNet, outperforming their original networks. (2) CE has an interesting connection with the Nash Equilibrium, a well-known solution of a non-cooperative game. (3) Extensive experiments show that CE achieves state-of-the-art performance on various challenging benchmarks such as ImageNet and COCO.

Intelligent agent naturally learns from motion. Various self-supervised algorithms have leveraged the motion cues to learn effective visual representations. The hurdle here is that motion is both ambiguous and complex, rendering previous works either suffer from degraded learning efficacy, or resort to strong assumptions on object motions. In this work, we design a new learning-from-motion paradigm to bridge these gaps. Instead of explicitly modeling the motion probabilities, we design the pretext task as a conditional motion propagation problem. Given an input image and several sparse flow guidance on it, our framework seeks to recover the full-image motion. Compared to other alternatives, our framework has several appealing properties: (1) Using sparse flow guidance during training resolves the inherent motion ambiguity, and thus easing feature learning. (2) Solving the pretext task of conditional motion propagation encourages the emergence of kinematically-sound representations that poss greater expressive power. Extensive experiments demonstrate that our framework learns structural and coherent features; and achieves state-of-the-art self-supervision performance on several downstream tasks including semantic segmentation, instance segmentation and human parsing. Furthermore, our framework is successfully extended to several useful applications such as semi-automatic pixel-level annotation.

Normalization methods are essential components in convolutional neural networks (CNNs). They either standardize or whiten data using statistics estimated in predefined sets of pixels. Unlike existing works that design normalization techniques for specific tasks, we propose Switchable Whitening (SW), which provides a general form unifying different whitening methods as well as standardization methods. SW learns to switch among these operations in an end-to-end manner. It has several advantages. First, SW adaptively selects appropriate whitening or standardization statistics for different tasks (see Fig.1), making it well suited for a wide range of tasks without manual design. Second, by integrating benefits of different normalizers, SW shows consistent improvements over its counterparts in various challenging benchmarks. Third, SW serves as a useful tool for understanding the characteristics of whitening and standardization techniques. We show that SW outperforms other alternatives on image classification (CIFAR-10/100, ImageNet), semantic segmentation (ADE20K, Cityscapes), domain adaptation (GTA5, Cityscapes), and image style transfer (COCO). For example, without bells and whistles, we achieve state-of-the-art performance with 45.33% mIoU on the ADE20K dataset.

Convolutional neural networks (CNNs) are usually built by stacking convolutional operations layer-by-layer. Although CNN has shown strong capability to extract semantics from raw pixels, its capacity to capture spatial relationships of pixels across rows and columns of an image is not fully explored. These relationships are important to learn semantic objects with strong shape priors but weak appearance coherences, such as traffic lanes, which are often occluded or not even painted on the road surface as shown in Fig. 1 (a). In this paper, we propose Spatial CNN (SCNN), which generalizes traditional deep layer-by-layer convolutions to slice-by-slice convolutions within feature maps, thus enabling message passings between pixels across rows and columns in a layer. Such SCNN is particular suitable for long continuous shape structure or large objects, with strong spatial relationship but less appearance clues, such as traffic lanes, poles, and wall. We apply SCNN on a newly released very challenging traffic lane detection dataset and Cityscapse dataset. The results show that SCNN could learn the spatial relationship for structure output and significantly improves the performance. We show that SCNN outperforms the recurrent neural network (RNN) based ReNet and MRF+CNN (MRFNet) in the lane detection dataset by 8.7% and 4.6% respectively. Moreover, our SCNN won the 1st place on the TuSimple Benchmark Lane Detection Challenge, with an accuracy of 96.53%.