Abstract Text‐to‐motion generative models span a wide range of 3D human actions but struggle with nuanced stylistic attributes such as a “Chicken” style. Due to the scarcity of style‐specific data, existing approaches pull the generative prior towards a reference style, which often results in out‐of‐distribution, low‐quality generations. In this work, we introduce LoRA‐MDM, a lightweight framework for motion stylization that generalizes to complex actions while maintaining editability. Our key insight is that adapting the generative prior to include the style, while preserving its overall distribution, is more effective than modifying each individual motion during generation. Building on this idea, LoRA‐MDM learns to adapt the prior to include the reference style using only a few samples. The style can then be used in the context of different textual prompts for generation. The low‐rank adaptation shifts the motion manifold in a semantically meaningful way, enabling realistic style infusion even for actions not present in the reference samples. Moreover, preserving the distribution structure enables advanced operations such as style blending and motion editing. We compare LoRA‐MDM to state‐of‐the‐art stylized motion generation methods and demonstrate a favorable balance between text fidelity and style consistency. Project page at https://haimsaw.github.io/LoRA‐MDM/

We propose a generative technique to edit 3D shapes, represented as meshes, NeRFs, or Gaussian Splats, in ≈ 3 seconds, without the need for running an SDS type of optimization. Our key insight is to cast 3D editing as a multiview image inpainting problem, as this representation is generic and can be mapped back to any 3D representation using the bank of available Large Reconstruction Models. We explore different fine-tuning strategies to obtain both multiview generation and inpainting capabilities within the same diffusion model. In particular, the design of the inpainting mask is an important factor of training an inpainting model, and we propose several masking strategies to mimic the types of edits a user would perform on a 3D shape. Our approach takes 3D generative editing from hours to seconds and produces higher-quality results compared to previous works.

Motion diffusion models and Reinforcement Learning (RL) based control for physics-based simulations have complementary strengths for human motion generation. The former is capable of generating a wide variety of motions, adhering to intuitive control such as text, while the latter offers physically plausible motion and direct interaction with the environment. In this work, we present a method that combines their respective strengths. CLoSD is a text-driven RL physics-based controller, guided by diffusion generation for various tasks. Our key insight is that motion diffusion can serve as an on-the-fly universal planner for a robust RL controller. To this end, CLoSD maintains a closed-loop interaction between two modules — a Diffusion Planner (DiP), and a tracking controller. DiP is a fast-responding autoregressive diffusion model, controlled by textual prompts and target locations, and the controller is a simple and robust motion imitator that continuously receives motion plans from DiP and provides feedback from the environment. CLoSD is capable of seamlessly performing a sequence of different tasks, including navigation to a goal location, striking an object with a hand or foot as specified in a text prompt, sitting down, and getting up.

We introduce a new interpretation of the attention matrix as a discrete-time Markov chain. Our interpretation sheds light on common operations involving attention scores such as selection, summation, and averaging in a unified framework. It further extends them by considering indirect attention, propagated through the Markov chain, as opposed to previous studies that only model immediate effects. Our key observation is that tokens linked to semantically similar regions form metastable states, i.e., regions where attention tends to concentrate, while noisy attention scores dissipate. Metastable states and their prevalence can be easily computed through simple matrix multiplication and eigenanalysis, respectively. Using these lightweight tools, we demonstrate state-of-the-art zero-shot segmentation. Lastly, we define TokenRank---the steady state vector of the Markov chain, which measures global token importance. We show that TokenRank enhances unconditional image generation, improving both quality (IS) and diversity (FID), and can also be incorporated into existing segmentation techniques to improve their performance over existing benchmarks. We believe our framework offers a fresh view of how tokens are being attended in modern visual transformers.

Light control in generated images is a difficult task, posing specific challenges, spanning over the entire image and frequency spectrum. Most approaches tackle this problem by training on extensive yet domain-specific datasets, limiting the inherent generalization and applicability of the foundational backbones used. Instead, PractiLight is a practical approach, effectively leveraging foundational understanding of recent generative models for the task. Our key insight is that lighting relationships in an image are similar in nature to token interaction in self-attention layers, and hence are best represented there. Based on this and other analyses regarding the importance of early diffusion iterations, PractiLight trains a lightweight LoRA regressor to produce the direct-irradiance map for a given image, using a small set of training images. We then employ this regressor to incorporate the desired lighting into the generation process of another image using Classifier Guidance. This careful design generalizes well to diverse conditions and image domains. We demonstrate state-of-the-art performance in terms of quality and control with proven parameter and data efficiency compared to leading works over a wide variety of scene types. We hope this work affirms that image lighting can feasibly be controlled by tapping into foundational knowledge, enabling practical and general relighting.

A sketch is one of the most intuitive and versatile tools humans use to convey their ideas visually. An animated sketch opens another dimension to the expression of ideas and is widely used by designers for a variety of purposes. Animating sketches is a laborious process, requiring extensive experience and professional design skills. In this work, we present a method that automatically adds motion to a single-subject sketch (hence, “breathing life into it”), merely by providing a text prompt indicating the desired motion. The output is a short animation provided in vector representation, which can be easily edited. Our method does not require extensive training, but instead leverages the motion prior of a large pretrained text-to- video diffusion model using a score-distillation loss to guide the placement of strokes. To promote natural and smooth motion and to better preserve the sketch's appearance, we model the learned motion through two components. The first governs small local deformations and the second controls global affine transformations. Surprisingly, wefind that even models that struggle to generate sketch videos on their own can still serve as a useful backbone for animating abstract representations.

We introduce Multi-view Ancestral Sampling (MAS), a method for 3D motion generation, using 2D diffusion models that were trained on motions obtained from in-the-wild videos. As such, MAS opens opportunities to exciting and diverse fields of motion previously under-explored as 3D data is scarce and hard to collect. MAS works by simultaneously denoising multiple 2D motion sequences representing different views of the same 3D motion. It ensures consistency across all views at each diffusion step by combining the individual generations into a unified 3D sequence, and projecting it back to the original views. We demonstrate MAS on 2D pose data acquired from videos depicting professional basketball maneuvers, rhythmic gymnastic performances featuring a ball apparatus, and horse races. In each of these domains, 3D motion capture is arduous, and yet, MAS generates diverse and realistic 3D sequences. Unlike the Score Distillation approach, which optimizes each sample by repeatedly applying small fixes, our method uses a sampling process that was constructed for the diffusion framework. As we demonstrate, MAS avoids common issues such as out-of-domain sampling and mode-collapse. https://guytevet.github.io/mas-page/

Generating multi-instrument music from symbolic music representations is an important task in Music Information Retrieval (MIR). A central but still largely unsolved problem in this context is musically and acoustically informed control in the generation process. As the main contribution of this work, we propose enhancing control of multi-instrument synthesis by conditioning a generative model on a specific performance and recording environment, thus allowing for better guidance of timbre and style. Building on state-of-the-art diffusion-based music generative models, we introduce performance conditioning - a simple tool indicating the generative model to synthesize music with style and timbre of specific instruments taken from specific performances. Our prototype is evaluated using uncurated performances with diverse instrumentation and achieves state-of-the-art FAD realism scores while allowing novel timbre and style control. Our project page, including samples and demonstrations, is available at benadar293.github.io/midipm.

Synthesizing realistic animations of humans, animals, and even imaginary creatures, has long been a goal for artists and computer graphics professionals. Compared to the imaging domain, which is rich with large available datasets, the number of data instances for the motion domain is limited, particularly for the animation of animals and exotic creatures (e.g., dragons), which have unique skeletons and motion patterns. In this work, we introduce SinMDM, a Single Motion Diffusion Model. It is designed to learn the internal motifs of a single motion sequence with arbitrary topology and synthesize a variety of motions of arbitrary length that remain faithful to the learned motifs. We harness the power of diffusion models and present a denoising network explicitly designed for the task of learning from a single input motion. SinMDM is crafted as a lightweight architecture, which avoids overfitting by using a shallow network with local attention layers that narrow the receptive field and encourage motion diversity. Our work applies to multiple contexts, including spatial and temporal in-betweening, motion expansion, style transfer, and crowd animation. Our results show that SinMDM outperforms existing methods both qualitatively and quantitatively. Moreover, while prior network-based approaches require additional training for different applications, SinMDM supports these applications during inference. Our project page, which includes links to the code and trained models, is accessible at https://sinmdm.github.io/SinMDM-page.

Recent work has demonstrated the significant potential of denoising diffusion models for generating human motion, including text-to-motion capabilities. However, these methods are restricted by the paucity of annotated motion data, a focus on single-person motions, and a lack of detailed control. In this paper, we introduce three forms of composition based on diffusion priors: sequential, parallel, and model composition. Using sequential composition, we tackle the challenge of long sequence generation. We introduce DoubleTake, an inference-time method with which we generate long animations consisting of sequences of prompted intervals and their transitions, using a prior trained only for short clips. Using parallel composition, we show promising steps toward two-person generation. Beginning with two fixed priors as well as a few two-person training examples, we learn a slim communication block, ComMDM, to coordinate interaction between the two resulting motions. Lastly, using model composition, we first train individual priors to complete motions that realize a prescribed motion for a given joint. We then introduce DiffusionBlending, an interpolation mechanism to effectively blend several such models to enable flexible and efficient fine-grained joint and trajectory-level control and editing. We evaluate the composition methods using an off-the-shelf motion diffusion model, and further compare the results to dedicated models trained for these specific tasks.

Abstract The field of visual computing is rapidly advancing due to the emergence of generative artificial intelligence (AI), which unlocks unprecedented capabilities for the generation, editing, and reconstruction of images, videos, and 3D scenes. In these domains, diffusion models are the generative AI architecture of choice. Within the last year alone, the literature on diffusion‐based tools and applications has seen exponential growth and relevant papers are published across the computer graphics, computer vision, and AI communities with new works appearing daily on arXiv. This rapid growth of the field makes it difficult to keep up with all recent developments. The goal of this state‐of‐the‐art report (STAR) is to introduce the basic mathematical concepts of diffusion models, implementation details and design choices of the popular Stable Diffusion model, as well as overview important aspects of these generative AI tools, including personalization, conditioning, inversion, among others. Moreover, we give a comprehensive overview of the rapidly growing literature on diffusion‐based generation and editing, categorized by the type of generated medium, including 2D images, videos, 3D objects, locomotion, and 4D scenes. Finally, we discuss available datasets, metrics, open challenges, and social implications. This STAR provides an intuitive starting point to explore this exciting topic for researchers, artists, and practitioners alike.

Reconstructing 3D shapes from planar cross-sections is a challenge inspired by downstream applications like medical imaging and geographic informatics. The input is an in/out indicator function fully defined on a sparse collection of planes in space, and the output is an interpolation of the indicator function to the entire volume. Previous works addressing this sparse and ill-posed problem either produce low quality results, or rely on additional priors such as target topology, appearance information, or input normal directions. In this paper, we present OReX, a method for 3D shape reconstruction from slices alone, featuring a Neural Field as the interpolation prior. A modest neural network is trained on the input planes to return an inside/outside estimate for a given 3D coordinate, yielding a powerful prior that induces smoothness and self-similarities. The main challenge for this approach is high-frequency details, as the neural prior is overly smoothing. To alleviate this, we offer an iterative estimation architecture and a hierarchical input sampling scheme that encourage coarse-to-fine training, allowing the training process to focus on high frequencies at later stages. In addition, we identify and analyze a ripple-like effect stemming from the mesh extraction step. We mitigate it by regularizing the spatial gradients of the indicator function around input in/out boundaries during network training, tackling the problem at the root. Through extensive qualitative and quantitative experimentation, we demonstrate our method is robust, accurate, and scales well with the size of the input. We report state-of-the-art results compared to previous approaches and recent potential solutions, and demonstrate the benefit of our individual contributions through analysis and ablation studies.11Code and data available at https://github.com/haimsaw/OReX

Text-to-image (T2I) personalization allows users to guide the creative image generation process by combining their own visual concepts in natural language prompts. Recently, encoder-based techniques have emerged as a new effective approach for T2I personalization, reducing the need for multiple images and long training times. However, most existing encoders are limited to a single-class domain, which hinders their ability to handle diverse concepts. In this work, we propose a domain-agnostic method that does not require any specialized dataset or prior information about the personalized concepts. We introduce a novel contrastive-based regularization technique to maintain high fidelity to the target concept characteristics while keeping the predicted embeddings close to editable regions of the latent space, by pushing the predicted tokens toward their nearest existing CLIP tokens. Our experimental results demonstrate the effectiveness of our approach and show how the learned tokens are more semantic than tokens predicted by unregularized models. This leads to a better representation that achieves state-of-the-art performance while being more flexible than previous methods.

Sheet Metal (SM) fabrication is perhaps one of the most common metalworking technique. Despite its prevalence, SM design is manual and costly, with rigorous practices that restrict the search space, yielding suboptimal results. In contrast, we present a framework for the first automatic design of SM parts. Focusing on load bearing applications, our novel system generates a high-performing manufacturable SM that adheres to the numerous constraints that SM design entails: The resulting part minimizes manufacturing costs while adhering to structural, spatial, and manufacturing constraints. In other words, the part should be strong enough, not disturb the environment, and adhere to the manufacturing process. These desiderata sum up to an elaborate, sparse, and expensive search space. Our generative approach is a carefully designed exploration process, comprising two steps. In Segment Discovery connections from the input load to attachable regions are accumulated, and during Segment Composition the most performing valid combination is searched for. For Discovery, we define a slim grammar, and sample it for parts using a Markov-Chain Monte Carlo (MCMC) approach, ran in intercommunicating instances (i.e, chains) for diversity. This, followed by a short continuous optimization, enables building a diverse and high-quality library of substructures. During Composition, a valid and minimal cost combination of the curated substructures is selected. To improve compliance significantly without additional manufacturing costs, we reinforce candidate parts onto themselves --- a unique SM capability called self-riveting. we provide our code and data in https://github.com/amir90/AutoSheetMetal. We show our generative approach produces viable parts for numerous scenarios. We compare our system against a human expert and observe improvements in both part quality and design time. We further analyze our pipeline's steps with respect to resulting quality, and have fabricated some results for validation. We hope our system will stretch the field of SM design, replacing costly expert hours with minutes of standard CPU, making this cheap and reliable manufacturing method accessible to anyone.

Text-to-image personalization aims to teach a pre-trained diffusion model to reason about novel, user provided concepts, embedding them into new scenes guided by natural language prompts. However, current personalization approaches struggle with lengthy training times, high storage requirements or loss of identity. To overcome these limitations, we propose an encoder-based domain-tuning approach. Our key insight is that by underfitting on a large set of concepts from a given domain, we can improve generalization and create a model that is more amenable to quickly adding novel concepts from the same domain. Specifically, we employ two components: First, an encoder that takes as an input a single image of a target concept from a given domain, e.g. a specific face, and learns to map it into a word-embedding representing the concept. Second, a set of regularized weight-offsets for the text-to-image model that learn how to effectively injest additional concepts. Together, these components are used to guide the learning of unseen concepts, allowing us to personalize a model using only a single image and as few as 5 training steps --- accelerating personalization from dozens of minutes to seconds , while preserving quality. Code and trained encoders will be available at our project page.

Recently, numerous facial editing techniques have been proposed that leverage the generative power of a pretrained StyleGAN. To successfully edit an image this way, one must first project (or invert) the image into the pretrained generator’s domain. As it turns out, StyleGAN’s latent space induces an inherent tradeoff between distortion and editability, i.e., between maintaining the original appearance and convincingly altering its attributes. Hence, it remains challenging to apply ID-preserving edits to real facial images. In this article, we present an approach to bridge this gap. The idea is Pivotal Tuning —a brief training process that preserves editing quality, while surgically changing the portrayed identity and appearance. In Pivotal Tuning Inversion, an initial inverted latent code serves as a pivot, around which the generator is fine-tuned. At the same time, a regularization term keeps nearby identities intact, to locally contain the effect. We further show that Pivotal Tuning also applies to accommodating for a multitude of faces, while introducing negligible distortion on the rest of the domain. We validate our technique through inversion and editing metrics and show preferable scores to state-of-the-art methods. Last, we present successful editing for harder cases, including elaborate make-up or headwear.

We introduce a high resolution spatially adaptive light source, or a projector, into a neural reflectance field that allows to both calibrate the projector and photo realistic light editing. The projected texture is fully differentiable with respect to all scene parameters, and can be optimized to yield a desired appearance suitable for applications in augmented reality and projection mapping. Our neural field consists of three neural networks, estimating geometry, material, and transmittance. Using an analytical BRDF model and carefully selected projection patterns, our acquisition process is simple and intuitive, featuring a fixed uncalibrated projected and a handheld camera with a co-located light source. As we demonstrate, the virtual projector incorporated into the pipeline improves scene understanding and enables various projection mapping applications, alleviating the need for time consuming calibration steps performed in a traditional setting per view or projector location. In addition to enabling novel viewpoint synthesis, we demonstrate state-of-the-art performance projector compensation for novel viewpoints, improvement over the baselines in material and scene reconstruction, and three simply implemented scenarios where projection image optimization is performed, including the use of a 2D generative model to consistently dictate scene appearance from multiple viewpoints. We believe that neural projection mapping opens up the door to novel and exciting downstream tasks, through the joint optimization of the scene and projection images.

The inversion of real images into StyleGAN's latent space is a well-studied problem. Nevertheless, applying existing approaches to real-world scenarios remains an open challenge, due to an inherent trade-off between reconstruction and editability: latent space regions which can accurately represent real images typically suffer from degraded semantic control. Recent work proposes to mitigate this trade-off by fine-tuning the generator to add the target image to well-behaved, editable regions of the latent space. While promising, this fine-tuning scheme is impractical for prevalent use as it requires a lengthy training phase for each new image. In this work, we introduce this approach into the realm of encoder-based inversion. We propose HyperStyle, a hypernetwork that learns to modulate StyleGAN's weights to faithfully express a given image in editable regions of the latent space. A naive modulation approach would require training a hypernetwork with over three billion parameters. Through careful network design, we reduce this to be in line with existing encoders. HyperStyle yields reconstructions comparable to those of optimization techniques with the near real-time inference capabilities of encoders. Lastly, we demonstrate HyperStyle's effectiveness on several applications beyond the inversion task, including the editing of out-of-domain images which were never seen during training. Code is available on our project page: https://yuval-alaluf.github.io/hyperstyle/.

Vision Transformers (ViT) serve as powerful vision models. Unlike convolutional neural networks, which dominated vision research in previous years, vision transformers enjoy the ability to capture long-range dependencies in the data. Nonetheless, an integral part of any transformer architecture, the self-attention mechanism, suffers from high latency and inefficient memory utilization, making it less suitable for high-resolution input images. To alleviate these shortcomings, hierarchical vision models locally employ self-attention on non-interleaving windows. This relaxation reduces the complexity to be linear in the input size; however, it limits the cross-window interaction, hurting the model performance. In this paper, we propose a new shift-invariant local attention layer, called query and attend (QnA), that aggregates the input locally in an overlapping manner, much like convolutions. The key idea behind QnA is to introduce learned queries, which allow fast and efficient implementation. We verify the effectiveness of our layer by incorporating it into a hierarchical vision transformer model. We show improvements in speed and memory complexity while achieving comparable accuracy with state-of-the-art models. Finally, our layer scales especially well with window size, requiring up to x10 less memory while being up to x5 faster than existing methods. The code is publicly available at https://github.com/moabarar/qna.

Multi-instrument Automatic Music Transcription (AMT), or the decoding of a musical recording into semantic musical content, is one of the holy grails of Music Information Retrieval. Current AMT approaches are restricted to piano and (some) guitar recordings, due to difficult data collection. In order to overcome data collection barriers, previous AMT approaches attempt to employ musical scores in the form of a digitized version of the same song or piece. The scores are typically aligned using audio features and strenuous human intervention to generate training labels. We introduce Note$_{EM}$, a method for simultaneously training a transcriber and aligning the scores to their corresponding performances, in a fully-automated process. Using this unaligned supervision scheme, complemented by pseudo-labels and pitch shift augmentation, our method enables training on in-the-wild recordings with unprecedented accuracy and instrumental variety. Using only synthetic data and unaligned supervision, we report SOTA note-level accuracy of the MAPS dataset, and large favorable margins on cross-dataset evaluations. We also demonstrate robustness and ease of use; we report comparable results when training on a small, easily obtainable, self-collected dataset, and we propose alternative labeling to the MusicNet dataset, which we show to be more accurate. Our project page is available at https://benadar293.github.io.

The ability of Generative Adversarial Networks to encode rich semantics within their latent space has been widely adopted for facial image editing. However, replicating their success with videos has proven challenging. Applying StyleGAN editing to real videos introduces two main challenges: (i) StyleGAN operates over aligned crops. When editing videos, these crops need to be pasted back into the frame, resulting in a spatial inconsistency. (ii) Videos introduce a fundamental barrier to overcome — temporal coherency. To address the first challenge, we propose a novel stitching-tuning procedure. The generator is carefully tuned to overcome the spatial artifacts at crop borders, resulting in smooth transitions even when difficult backgrounds are involved. Turning to temporal coherence, we propose that this challenge is largely artificial. The source video is already temporally coherent, and deviations arise in part due to careless treatment of individual components in the editing pipeline. We leverage the natural alignment of StyleGAN and the tendency of neural networks to learn low-frequency functions, and demonstrate that they provide a strongly consistent prior. These components are combined in an end-to-end framework for semantic editing of facial videos. We compare our pipeline to the current state-of-the-art and demonstrate significant improvements. Our method produces meaningful manipulations and maintains greater spatial and temporal consistency, even on challenging talking head videos which current methods struggle with. Our code and videos are available at https://stitch-time.github.io/.

Text entry for mobile devices nowadays is an equally crucial and time-consuming task, with no practical solution available for natural typing speeds without extra hardware. In this paper, we introduce a real-time method that is a significant step towards enabling touch typing on arbitrary flat surfaces (e.g., tables). The method employs only a simple video camera, placed in front of the user on the flat surface — at an angle practical for mobile usage. To achieve this, we adopt a classification framework, based on the observation that, in touch typing, similar hand configurations imply the same typed character across users. Importantly, this approach allows the convenience of un-calibrated typing, where the hand positions, with respect to the camera and each other, are not dictated.To improve accuracy, we propose a Language Processing scheme, which corrects the typed text and is specifically designed for real-time performance and integration with the vision-based signal. To enable feasible data collection and training, we propose a self-refinement approach that allows training on unlabeled flat-surface-typing footage; A network trained on (labeled) keyboard footage labels flat-surface videos using dynamic time warping, and is trained on them, in an Expectation Maximization (EM) manner.Using these techniques, we introduce the TypingHands26 Dataset, comprising videos of 26 different users typing on a keyboard, and 10 users typing on a flat surface, labeled at the frame level. We validate our approach and present a single camera-based system with character-level accuracy of 93.5% on average for known users, and 85.7% for unknown ones, outperforming pose-estimation-based methods by a large margin, despite performing at natural typing speeds of up to 80 Words Per Minute. Our method is the first to rely on a simple camera alone, and runs in interactive speeds, while still maintaining accuracy comparable to systems employing non-commodity equipment.

Abstract Generative Adversarial Networks (GANs) have established themselves as a prevalent approach to image synthesis. Of these, StyleGAN offers a fascinating case study, owing to its remarkable visual quality and an ability to support a large array of downstream tasks. This state‐of‐the‐art report covers the StyleGAN architecture, and the ways it has been employed since its conception, while also analyzing its severe limitations. It aims to be of use for both newcomers, who wish to get a grasp of the field, and for more experienced readers that might benefit from seeing current research trends and existing tools laid out. Among StyleGAN's most interesting aspects is its learned latent space. Despite being learned with no supervision, it is surprisingly well‐behaved and remarkably disentangled. Combined with StyleGAN's visual quality, these properties gave rise to unparalleled editing capabilities. However, the control offered by StyleGAN is inherently limited to the generator's learned distribution, and can only be applied to images generated by StyleGAN itself. Seeking to bring StyleGAN's latent control to real‐world scenarios, the study of GAN inversion and latent space embedding has quickly gained in popularity. Meanwhile, this same study has helped shed light on the inner workings and limitations of StyleGAN. We map out StyleGAN's impressive story through these investigations, and discuss the details that have made StyleGAN the go‐to generator. We further elaborate on the visual priors StyleGAN constructs, and discuss their use in downstream discriminative tasks. Looking forward, we point out StyleGAN's limitations and speculate on current trends and promising directions for future research, such as task and target specific fine‐tuning.

Abstract In unsupervised video retargeting, content is transferred from one video to another while preserving the original appearance and style, without any additional annotations. While this challenge has seen substantial advancements through the use of deep neural networks, current methods struggle when the source and target videos are of shapes that are different in limb lengths or other body proportions. In this work, we consider this task for the case of objects of different shapes and appearances, that consist of similar skeleton connectivity and depict similar motion. We introduce JOKR—a JOint Keypoint Representation that captures the geometry common to both videos, while being disentangled from their unique styles. Our model first extracts unsupervised keypoints from the given videos. From this representation, two decoders reconstruct geometry and appearance, one for each of the input sequences. By employing an affine‐invariant domain confusion term over the keypoints bottleneck, we enforce the unsupervised keypoint representations of both videos to be indistinguishable. This encourages the aforementioned disentanglement between motion and appearance, mapping similar poses from both domains to the same representation. This allows yielding a sequence with the appearance and style of one video, but the content of the other. Our applicability is demonstrated through challenging video pairs compared to state‐of‐the‐art methods. Furthermore, we demonstrate that this geometry‐driven representation enables intuitive control, such as temporal coherence and manual pose editing. Videos can be viewed in the supplement HTML.

Can a generative model be trained to produce images from a specific domain, guided only by a text prompt, without seeing any image? In other words: can an image generator be trained "blindly"? Leveraging the semantic power of large scale Contrastive-Language-Image-Pre-training (CLIP) models, we present a text-driven method that allows shifting a generative model to new domains, without having to collect even a single image. We show that through natural language prompts and a few minutes of training, our method can adapt a generator across a multitude of domains characterized by diverse styles and shapes. Notably, many of these modifications would be difficult or infeasible to reach with existing methods. We conduct an extensive set of experiments across a wide range of domains. These demonstrate the effectiveness of our approach, and show that our models preserve the latent-space structure that makes generative models appealing for downstream tasks. Code and videos available at: stylegan-nada.github.io/

Abstraction is at the heart of sketching due to the simple and minimal nature of line drawings. Abstraction entails identifying the essential visual properties of an object or scene, which requires semantic understanding and prior knowledge of high-level concepts. Abstract depictions are therefore challenging for artists, and even more so for machines. We present CLIPasso, an object sketching method that can achieve different levels of abstraction, guided by geometric and semantic simplifications. While sketch generation methods often rely on explicit sketch datasets for training, we utilize the remarkable ability of CLIP (Contrastive-Language-Image-Pretraining) to distill semantic concepts from sketches and images alike. We define a sketch as a set of Bézier curves and use a differentiable rasterizer to optimize the parameters of the curves directly with respect to a CLIP-based perceptual loss. The abstraction degree is controlled by varying the number of strokes. The generated sketches demonstrate multiple levels of abstraction while maintaining recognizability, underlying structure, and essential visual components of the subject drawn.

Abstract The task of unsupervised image‐to‐image translation has seen substantial advancements in recent years through the use of deep neural networks. Typically, the proposed solutions learn the characterizing distribution of two large, unpaired collections of images, and are able to alter the appearance of a given image, while keeping its geometry intact. In this paper, we explore the capabilities of neural networks to understand image structure given only a single pair of images, and . We seek to generate images that are structurally aligned: that is, to generate an image that keeps the appearance and style of , but has a structural arrangement that corresponds to . The key idea is to map between image patches at different scales. This enables controlling the granularity at which analogies are produced, which determines the conceptual distinction between style and content. In addition to structural alignment, our method can be used to generate high quality imagery in other conditional generation tasks utilizing images and only: guided image synthesis, style and texture transfer, text translation as well as video translation. Our code and additional results are available in https://github.com/rmokady/structural-analogy/

In recent years, considerable progress has been made in the visual quality of Generative Adversarial Networks (GANs). Even so, these networks still suffer from degradation in quality for high-frequency content, stemming from a spectrally biased architecture, and similarly unfavorable loss functions. To address this issue, we present a novel general-purpose Style and WAvelet based GAN (SWAGAN) that implements progressive generation in the frequency domain. SWAGAN incorporates wavelets throughout its generator and discriminator architectures, enforcing a frequency-aware latent representation at every step of the way. This approach, designed to directly tackle the spectral bias of neural networks, yields an improvement in the ability to generate medium and high frequency content, including structures which other networks fail to learn. We demonstrate the advantage of our method by integrating it into the SyleGAN2 framework, and verifying that content generation in the wavelet domain leads to more realistic high-frequency content, even when trained for fewer iterations. Furthermore, we verify that our model's latent space retains the qualities that allow StyleGAN to serve as a basis for a multitude of editing tasks, and show that our frequency-aware approach also induces improved high-frequency performance in downstream tasks.

Many applications, such as autonomous driving, heavily rely on multi-modal data where spatial alignment between the modalities is required. Most multi-modal registration methods struggle computing the spatial correspondence between the images using prevalent cross-modality similarity measures. In this work, we bypass the difficulties of developing cross-modality similarity measures, by training an image-to-image translation network on the two input modalities. This learned translation allows training the registration network using simple and reliable mono-modality metrics. We perform multi-modal registration using two networks - a spatial transformation network and a translation network. We show that by encouraging our translation network to be geometry preserving, we manage to train an accurate spatial transformation network. Compared to state-of-the-art multi-modal methods our presented method is unsupervised, requiring no pairs of aligned modalities for training, and can be adapted to any pair of modalities. We evaluate our method quantitatively and qualitatively on commercial datasets, showing that it performs well on several modalities and achieves accurate alignment.

Visualizing high-dimensional data on a 2D canvas is generally challenging. It becomes significantly more difficult when multiple time-steps are to be presented, as the visual clutter quickly increases. Moreover, the challenge to perceive the significant temporal evolution is even greater. In this paper, we present a method to plot temporal high-dimensional data in a static scatterplot; it uses the established PCA technique to project data from multiple time-steps. The key idea is to extend each individual displacement prior to applying PCA, so as to skew the projection process, and to set a projection plane that balances the directions of temporal change and spatial variance. We present numerous examples and various visual cues to highlight the data trajectories, and demonstrate the effectiveness of the method for visualizing temporal data.

The paradigm of image-to-image translation is leveraged for the benefit of sketch stylization via transfer of geometric textural details. Lacking the necessary volumes of data for standard training of translation systems, we advocate for operation at the patch level, where a handful of stylized sketches provide ample mining potential for patches featuring basic geometric primitives. Operating at the patch level necessitates special consideration of full sketch translation, as individual translation of patches with no regard to neighbors is likely to produce visible seams and artifacts at patch borders. Aligned pairs of styled and plain primitives are combined to form input hybrids containing styled elements around the border and plain elements within, and given as input to a seamless translation (ST) generator, whose output patches are expected to reconstruct the fully styled patch. An adversarial addition promotes generalization and robustness to diverse geometries at inference time, forming a simple and effective system for arbitrary sketch stylization, as demonstrated upon a variety of styles and sketches.

Learning disentangled representations of data is a fundamental problem in artificial intelligence. Specifically, disentangled latent representations allow generative models to control and compose the disentangled factors in the synthesis process. Current methods, however, require extensive supervision and training, or instead, noticeably compromise quality. In this paper, we present a method that learns how to represent data in a disentangled way, with minimal supervision, manifested solely using available pre-trained networks. Our key insight is to decouple the processes of disentanglement and synthesis, by employing a leading pre-trained unconditional image generator, such as StyleGAN. By learning to map into its latent space, we leverage both its state-of-the-art quality, and its rich and expressive latent space, without the burden of training it. We demonstrate our approach on the complex and high dimensional domain of human heads. We evaluate our method qualitatively and quantitatively, and exhibit its success with de-identification operations and with temporal identity coherency in image sequences. Through extensive experimentation, we show that our method successfully disentangles identity from other facial attributes, surpassing existing methods, even though they require more training and supervision.

Abstract Designing fonts and typefaces is a difficult process for both beginner and expert typographers. Existing workflows require the designer to create every glyph, while adhering to many loosely defined design suggestions to achieve an aesthetically appealing and coherent character set. This process can be significantly simplified by exploiting the similar structure character glyphs present across different fonts and the shared stylistic elements within the same font. To capture these correlations, we propose learning a stroke‐based font representation from a collection of existing typefaces. To enable this, we develop a stroke‐based geometric model for glyphs, a fitting procedure to reparametrize arbitrary fonts to our representation. We demonstrate the effectiveness of our model through a manifold learning technique that estimates a low‐dimensional font space. Our representation captures a wide range of everyday fonts with topological variations and naturally handles discrete and continuous variations, such as presence and absence of stylistic elements as well as slants and weights. We show that our learned representation can be used for iteratively improving fit quality, as well as exploratory style applications such as completing a font from a subset of observed glyphs, interpolating or adding and removing stylistic elements in existing fonts.

We propose the first system for live dynamic augmentation of human faces. Using projector-based illumination, we alter the appearance of human performers during novel performances. The key challenge of live augmentation is latency — an image is generated according to a specific pose, but is displayed on a different facial configuration by the time it is projected. Therefore, our system aims at reducing latency during every step of the process, from capture, through processing, to projection. Using infrared illumination, an optically and computationally aligned high-speed camera detects facial orientation as well as expression. The estimated expression blendshapes are mapped onto a lower dimensional space, and the facial motion and non-rigid deformation are estimated, smoothed and predicted through adaptive Kalman filtering. Finally, the desired appearance is generated interpolating precomputed offset textures according to time, global position, and expression. We have evaluated our system through an optimized CPU and GPU prototype, and demonstrated successful low latency augmentation for different performers and performances with varying facial play and motion speed. In contrast to existing methods, the presented system is the first method which fully supports dynamic facial projection mapping without the requirement of any physical tracking markers and incorporates facial expressions.

Computational manufacturing technologies such as 3D printing hold the potential for creating objects with previously undreamed-of combinations of functionality and physical properties. Human designers, however, typically cannot exploit the full geometric (and often material) complexity of which these devices are capable. This STAR examines recent systems developed by the computer graphics community in which designers specify higher-level goals ranging from structural integrity and deformation to appearance and aesthetics, with the final detailed shape and manufacturing instructions emerging as the result of computation. It summarizes frameworks for interaction, simulation, and optimization, as well as documents the range of general objectives and domain-specific goals that have been considered. An important unifying thread in this analysis is that different underlying geometric and physical representations are necessary for different tasks: we document over a dozen classes of representations that have been used for fabrication-aware design in the literature. We analyze how these classes possess obvious advantages for some needs, but have also been used in creative manners to facilitate unexpected problem solutions.

We introduce a co-analysis technique designed for correspondence inference within large shape collections. Such collections are naturally rich in variation, adding ambiguity to the notoriously difficult problem of correspondence computation. We leverage the robustness of correspondences between similar shapes to address the difficulties associated with this problem. In our approach, pairs of similar shapes are extracted from the collection, analyzed and matched in an efficient and reliable manner, culminating in the construction of a network of correspondences that connects the entire collection. The correspondence between any pair of shapes then amounts to a simple propagation along the minimax path between the two shapes in the network. At the heart of our approach is the introduction of a robust, structure-oriented shape matching method. Leveraging the idea of projective analysis, we partition 2D projections of a shape to obtain a set of 1D ordered regions, which are both simple and efficient to match. We lift the matched projections back to the 3D domain to obtain a pairwise shape correspondence. The emphasis given to structural compatibility is a central tool in estimating the reliability and completeness of a computed correspondence, uncovering any non-negligible semantic discrepancies that may exist between shapes. These detected differences are a deciding factor in the establishment of a network aiming to capture local similarities. We demonstrate that the combination of the presented observations into a co-analysis method allows us to establish reliable correspondences among shapes within large collections.

In recent years we have seen numerous improvements on 3D scanning and tracking of human faces, greatly advancing the creation of digital doubles for film and video games. However, despite the high-resolution quality of the reconstruction approaches available, current methods are unable to capture one of the most important regions of the face - the eye region. In this work we present the first method for detailed spatio-temporal reconstruction of eyelids. Tracking and reconstructing eyelids is extremely challenging, as this region exhibits very complex and unique skin deformation where skin is folded under while opening the eye. Furthermore, eyelids are often only partially visible and obstructed due to self-occlusion and eyelashes. Our approach is to combine a geometric deformation model with image data, leveraging multi-view stereo, optical flow, contour tracking and wrinkle detection from local skin appearance. Our deformation model serves as a prior that enables reconstruction of eyelids even under strong self-occlusions caused by rolling and folding skin as the eye opens and closes. The output is a person-specific, time-varying eyelid reconstruction with anatomically plausible deformations. Our high-resolution detailed eyelids couple naturally with current facial performance capture approaches. As a result, our method can largely increase the fidelity of facial capture and the creation of digital doubles.

The facial performance of an individual is inherently rich in subtle deformation and timing details. Although these subtleties make the performance realistic and compelling, they often elude both motion capture and hand animation. We present a technique for adding fine-scale details and expressiveness to low-resolution art-directed facial performances, such as those created manually using a rig, via marker-based capture, by fitting a morphable model to a video, or through Kinect reconstruction using recent faceshift technology. We employ a high-resolution facial performance capture system to acquire a representative performance of an individual in which he or she explores the full range of facial expressiveness. From the captured data, our system extracts an expressiveness model that encodes subtle spatial and temporal deformation details specific to that particular individual. Once this model has been built, these details can be transferred to low-resolution art-directed performances. We demonstrate results on various forms of input; after our enhancement, the resulting animations exhibit the same nuances and fine spatial details as the captured performance, with optional temporal enhancement to match the dynamics of the actor. Finally, we show that our technique outperforms the current state-of-the-art in example-based facial animation.

Animated animatronic figures are a unique way to give physical presence to a character. However, their movement and expressions are often limited due to mechanical constraints. In this paper, we propose a complete process for augmenting physical avatars using projector-based illumination, significantly increasing their expressiveness. Given an input animation, the system decomposes the motion into low-frequency motion that can be physically reproduced by the animatronic head and high-frequency details that are added using projected shading. At the core is a spatio-temporal optimization process that compresses the motion in gradient space, ensuring faithful motion replay while respecting the physical limitations of the system. We also propose a complete multi-camera and projection system, including a novel defocused projection and subsurface scattering compensation scheme. The result of our system is a highly expressive physical avatar that features facial details and motion otherwise unattainable due to physical constraints.

Abstract ShadowPixare white surfaces that display several prescribed images formed by the self‐shadowing of the surface when lit from certain directions. The effect is surprising and not commonly seen in the real world. We present algorithms for constructingShadowPixthat allow up to four images to be embedded in a single surface. ShadowPixcan produce a variety of unusual effects depending on the embedded images: moving the light can animate or relight the object in the image, or three colored lights may be used to produce a single colored image. ShadowPixare easy to manufacture using a 3D printer and we present photographs, videos, and renderings demonstrating these effects.

We describe a simple algorithm to reconstruct the surface of smooth three-dimensional multilabeled objects from sampled planar cross-sections of arbitrary orientation. The algorithm has the unique ability to handle cross-sections in which regions are classified as being inside the object, outside the object, or unknown. This is achieved by constructing a scalar function on R 3 , whose zero set is the desired surface. The function is constructed independently inside every cell of the arrangement of the cross-section planes using transfinite interpolation techniques based on barycentric coordinates. These guarantee that the function is smooth, and its zero set interpolates the cross-sections. The algorithm is highly parallelizable and may be implemented as an incremental update as each new cross-section is introduced. This leads to an efficient online version, performed on a GPU, which is suitable for interactive medical applications.