Recent advancements in text-to-image generative models have demonstrated a remarkable ability to capture a deep semantic understanding of images. In this work, we leverage this semantic knowledge to transfer the visual appearance between objects that share similar semantics but may differ significantly in shape. To achieve this, we build upon the self-attention layers of these generative models and introduce a cross-image attention mechanism that implicitly establishes semantic correspondences across images. Specifically, given a pair of images — one depicting the target structure and the other specifying the desired appearance — our cross-image attention combines the queries corresponding to the structure image with the keys and values of the appearance image. This operation, when applied during the denoising process, leverages the established semantic correspondences to generate an image combining the desired structure and appearance. In addition, to improve the output image quality, we harness three mechanisms that either manipulate the noisy latent codes or the model’s internal representations throughout the denoising process. Importantly, our approach is zero-shot, requiring no optimization or training. Experiments show that our method is effective across a wide range of object categories and is robust to variations in shape, size, and viewpoint between the two input images.
论文检索
输入标题、作者或关键词，从 252 篇学术成果中精准定位
We explore the problem of computationally generating special images that produce multi-arrangement optical illusions when physically arranged and viewed in a certain way, which we call ‘prime’ images. First, we propose a formal definition for this problem. Next, we introduce Diffusion Illusions, the first comprehensive pipeline designed to automatically generate a wide range of these multi-arrangement illusions. Specifically, we both adapt the existing ‘score distillation loss’ and propose a new ‘dream target loss’ to optimize a group of differentially parametrized prime images, using a frozen text-to-image diffusion model. We study three types of illusions, each where the prime images are arranged in different ways and optimized using the aforementioned losses such that images derived from them align with user-chosen text prompts or images. We conduct comprehensive experiments on these illusions and verify the effectiveness of our proposed method qualitatively and quantitatively. Additionally, we showcase the successful physical fabrication of our illusions — as they are all designed to work in the real world. Code and examples are publicly available at our interactive project website: https://diffusionillusion.github.io/
Neural fields have been broadly investigated as scene representations for the reproduction and novel generation of diverse outdoor scenes, including those autonomous vehicles and robots must handle. While successful approaches for RGB and LiDAR data exist, neural reconstruction methods for radar as a sensing modality have been largely unexplored. Operating at millimeter wavelengths, radar sensors are robust to scattering in fog and rain, and, as such, offer a complementary modality to active and passive optical sensing techniques. Moreover, existing radar sensors are highly cost-effective and deployed broadly in robots and vehicles that operate outdoors. We introduce Radar Fields –- a neural scene reconstruction method designed for active radar imagers. Our approach unites an explicit, physics-informed sensor model with an implicit neural geometry and reflectance model to directly synthesize raw radar measurements and extract scene occupancy. The proposed method does not rely on volume rendering. Instead, we learn fields in Fourier frequency space, supervised with raw radar data. We validate our method’s effectiveness across diverse outdoor scenarios, including urban scenes with dense vehicles and infrastructure, and harsh weather scenarios, where mm-wavelength sensing is favorable.
Human senses and perception are our mechanisms to explore the external world. In this context, visual saccades –rapid and coordinated eye movements– serve as a primary tool for awareness of our surroundings. Typically, our perception is not limited to visual stimuli alone but is enriched by cross-modal interactions, such as the combination of sight and hearing. In this work, we investigate the temporal and spatial relationship of these interactions, focusing on how auditory cues that precede visual stimuli influence saccadic latency –the time that it takes for the eyes to react and start moving towards a visual target. Our research, conducted within a virtual reality environment, reveals that auditory cues preceding visual information can significantly accelerate saccadic responses, but this effect plateaus beyond certain temporal thresholds. Additionally, while the spatial positioning of visual stimuli influences the speed of these eye movements, as reported in previous research, we find that the location of auditory cues with respect to their corresponding visual stimulus does not have a comparable effect. To validate our findings, we implement two practical applications: first, a basketball training task set in a more realistic environment with complex audiovisual signals, and second, an interactive farm game that explores previously untested values of our key factors. Lastly, we discuss various potential applications where our model could be beneficial.
DiffSound: Differentiable Modal Sound Rendering and Inverse Rendering for Diverse Inference Tasks
PDF ↗Accurately estimating and simulating the physical properties of objects from real-world sound recordings is of great practical importance in the fields of vision, graphics, and robotics. However, the progress in these directions has been limited—prior differentiable rigid or soft body simulation techniques cannot be directly applied to modal sound synthesis due to the high sampling rate of audio, while previous audio synthesizers often do not fully model the accurate physical properties of the sounding objects. We propose DiffSound, a differentiable sound rendering framework for physics-based modal sound synthesis, which is based on an implicit shape representation, a new high-order finite element analysis module, and a differentiable audio synthesizer. Our framework can solve a wide range of inverse problems thanks to the differentiability of the entire pipeline, including physical parameter estimation, geometric shape reasoning, and impact position prediction. Experimental results demonstrate the effectiveness of our approach, highlighting its ability to accurately reproduce the target sound in a physics-based manner. DiffSound serves as a valuable tool for various sound synthesis and analysis applications.
Underwater perception and 3D surface reconstruction are challenging problems with broad applications in construction, security, marine archaeology, and environmental monitoring. Treacherous operating conditions, fragile surroundings, and limited navigation control often dictate that submersibles restrict their range of motion and, thus, the baseline over which they can capture measurements. In the context of 3D scene reconstruction, it is well-known that smaller baselines make reconstruction more challenging. Our work develops a physics-based multimodal acoustic-optical neural surface reconstruction framework (AONeuS) capable of effectively integrating high-resolution RGB measurements with low-resolution depth-resolved imaging sonar measurements. By fusing these complementary modalities, our framework can reconstruct accurate high-resolution 3D surfaces from measurements captured over heavily-restricted baselines. Through extensive simulations and in-lab experiments, we demonstrate that AONeuS dramatically outperforms recent RGB-only and sonar-only inverse-differentiable-rendering–based surface reconstruction methods.
In the wake of many new ML-inspired approaches for reconstructing and representing high-quality 3D content, recent hybrid and explicitly learned representations exhibit promising performance and quality characteristics. However, their scaling to higher dimensions is challenging, e.g. when accounting for dynamic content with respect to additional parameters such as material properties, illumination, or time. In this paper, we tackle these challenges for an explicit representations based on Gaussian mixture models. With our solutions, we arrive at efficient fitting of compact N-dimensional Gaussian mixtures and enable efficient evaluation at render time: For fast fitting and evaluation, we introduce a high-dimensional culling scheme that efficiently bounds N-D Gaussians, inspired by Locality Sensitive Hashing. For adaptive refinement yet compact representation, we introduce a loss-adaptive density control scheme that incrementally guides the use of additional capacity towards missing details. With these tools we can for the first time represent complex appearance that depends on many input dimensions beyond position or viewing angle within a compact, explicit representation optimized in minutes and rendered in milliseconds.
Despite significant advancements in Neural Radiance Fields (NeRFs), the renderings may still suffer from aliasing and blurring artifacts, since it remains a fundamental challenge to effectively and efficiently characterize anisotropic areas induced by the cone-casting procedure. This paper introduces a Ripmap-Encoded Platonic Solid representation to precisely and efficiently featurize 3D anisotropic areas, achieving high-fidelity anti-aliased renderings. Central to our approach are two key components: Platonic Solid Projection and Ripmap encoding. The Platonic Solid Projection factorizes the 3D space onto the unparalleled faces of a certain Platonic solid, such that the anisotropic 3D areas can be projected onto planes with distinguishable characterization. Meanwhile, each face of the Platonic solid is encoded by the Ripmap encoding, which is constructed by anisotropically pre-filtering a learnable feature grid, to enable featurzing the projected anisotropic areas both precisely and efficiently by the anisotropic area-sampling. Extensive experiments on both well-established synthetic datasets and a newly captured real-world dataset demonstrate that our Rip-NeRF attains state-of-the-art rendering quality, particularly excelling in the fine details of repetitive structures and textures, while maintaining relatively swift training times, as shown in Fig. 1. The source code and data for this paper are at https://github.com/JunchenLiu77/Rip-NeRF.
Novel view synthesis from a sparse set of input images is a challenging problem of great practical interest, especially when camera poses are absent or inaccurate. Direct optimization of camera poses and usage of estimated depths in neural radiance field algorithms usually do not produce good results because of the coupling between poses and depths, and inaccuracies in monocular depth estimation. In this paper, we leverage the recent 3D Gaussian splatting method to develop a novel construct-and-optimize method for sparse view synthesis without camera poses. Specifically, we construct a solution progressively by using monocular depth and projecting pixels back into the 3D world. During construction, we optimize the solution by detecting 2D correspondences between training views and the corresponding rendered images. We develop a unified differentiable pipeline for camera registration and adjustment of both camera poses and depths, followed by back-projection. We also introduce a novel notion of an expected surface in Gaussian splatting, which is critical to our optimization. These steps enable a coarse solution, which can then be low-pass filtered and refined using standard optimization methods. We demonstrate results on the Tanks and Temples and Static Hikes datasets with as few as three widely-spaced views, showing significantly better quality than competing methods, including those with approximate camera pose information. Moreover, our results improve with more views and outperform previous InstantNGP and Gaussian Splatting algorithms even when using half the dataset.
We present the novel grammar language Recomp for the procedural modeling of architecture. In grammar-based approaches, the procedural refinement process is based on shape subdivisions. This process of decomposition results in disconnected subparts, which not only restricts the geometric expressiveness but also limits the control over an appropriate shape granularity needed to coordinate design decisions. Recomp overcomes these limitations by extending grammar languages with the recomposition ability. Fundamental is the concept of rule inlining, allowing for the topological recomposition of edited subparts by collapsing a shape subtree into one single shape on which derivation can continue. This is completed with a versatile geometry tagging system, allowing authors to compile and transport context information at any level of detail and gain full control over the geometry independent of the structure of the shape tree. Through various examples, we demonstrate the power of Recomp in procedural layout and mass modeling, as well as its capabilities in facilitating context-sensitive design.
Recent advancements in physics-based character animation leverage deep learning to generate agile and natural motion, enabling characters to execute movements such as backflips, boxing, and tennis. However, reproducing the selection and use of diverse motor skills in dynamic environments to solve complex tasks, as humans do, still remains a challenge. We present a strategy and skill learning approach for physics-based table tennis animation. Our method addresses the issue of mode collapse, where the characters do not fully utilize the motor skills they need to perform to execute complex tasks. More specifically, we demonstrate a hierarchical control system for diversified skill learning and a strategy learning framework for effective decision-making. We showcase the efficacy of our method through comparative analysis with state-of-the-art methods, demonstrating its capabilities in executing various skills for table tennis. Our strategy learning framework is validated through both agent-agent interaction and human-agent interaction in Virtual Reality, handling both competitive and cooperative tasks.
SuperPADL: Scaling Language-Directed Physics-Based Control with Progressive Supervised Distillation
PDF ↗Physically-simulated models for human motion can generate high-quality responsive character animations, often in real-time. Natural language serves as a flexible interface for controlling these models, allowing expert and non-expert users to quickly create and edit their animations. Many recent physics-based animation methods, including those that use text interfaces, train control policies using reinforcement learning (RL). However, scaling these methods beyond several hundred motions has remained challenging. Meanwhile, kinematic animation models are able to successfully learn from thousands of diverse motions by leveraging supervised learning methods. Inspired by these successes, in this work we introduce SuperPADL, a scalable framework for physics-based text-to-motion that leverages both RL and supervised learning to train controllers on thousands of diverse motion clips. SuperPADL is trained in stages using progressive distillation, starting with a large number of specialized experts using RL. These experts are then iteratively distilled into larger, more robust policies using a combination of reinforcement learning and supervised learning. Our final SuperPADL controller is trained on a dataset containing over 5000 skills and runs in real time on a consumer GPU. Moreover, our policy can naturally transition between skills, allowing for users to interactively craft multi-stage animations. We experimentally demonstrate that SuperPADL significantly outperforms RL-based baselines at this large data scale.
Misfocus is ubiquitous for almost all video producers, degrading video quality and often causing expensive delays and reshoots. Current autofocus (AF) systems are vulnerable to sudden disturbances such as subject movement or lighting changes commonly present in real-world and on-set conditions. Single image defocus deblurring methods are temporally unstable when applied to videos and cannot recover details obscured by temporally varying defocus blur. In this paper, we present an end-to-end solution that allows users to correct misfocus during post-processing. Our method generates and parameterizes defocused videos into sharp layered neural atlases and propagates consistent focus tracking back to the video frames. We introduce a novel differentiable disk blur layer for more accurate point spread function (PSF) simulation, coupled with a circle of confusion (COC) map estimation module with knowledge transferred from the current single image defocus deblurring (SIDD) networks. Our pipeline offers consistent, sharp video reconstruction and effective subject-focus correction and tracking directly on the generated atlases. Furthermore, by adopting our approach, we achieve comparable results to the state-of-the-art optical flow estimation approach from defocus videos.
Mobile cameras, despite their significant advancements, still have difficulty in low-light imaging due to compact sensors and lenses, leading to longer exposures and motion blur. Traditional blind deconvolution methods and learning-based deblurring methods can be potential solutions to remove blur. However, achieving practical performance still remains a challenge. To address this, we propose a learning-based deblurring framework for smartphones, utilizing wide and ultra-wide cameras as a hybrid camera system. We simultaneously capture a long-exposure wide image and short-exposure burst ultra-wide images, and utilize the burst images to deblur the wide image. To fully exploit burst ultra-wide images, we present HCDeblur, a practical deblurring framework that includes novel deblurring networks, HC-DNet and HC-FNet. HC-DNet utilizes motion information extracted from burst images to deblur a wide image, and HC-FNet leverages burst images as reference images to further enhance a deblurred output. For training and evaluating the proposed method, we introduce the HCBlur dataset, which consists of synthetic and real-world datasets. Our experiments demonstrate that HCDeblur achieves state-of-the-art deblurring quality. Codes and datasets are available at https://cg.postech.ac.kr/research/HCDeblur.
Existing methods for scale-invariant monocular depth estimation (SI MDE) often struggle due to the complexity of the task, and limited and non-diverse datasets, hindering generalizability in real-world scenarios. This is while shift-and-scale-invariant (SSI) depth estimation, simplifying the task and enabling training with abundant stereo datasets achieves high performance. We present a novel approach that leverages SSI inputs to enhance SI depth estimation, streamlining the network’s role and facilitating in-the-wild generalization for SI depth estimation while only using a synthetic dataset for training. Emphasizing the generation of high-resolution details, we introduce a novel sparse ordinal loss that substantially improves detail generation in SSI MDE, addressing critical limitations in existing approaches. Through in-the-wild qualitative examples and zero-shot evaluation we substantiate the practical utility of our approach in computational photography applications, showcasing its ability to generate highly detailed SI depth maps and achieve generalization in diverse scenarios.
Optics designers use simulation tools to assist them in designing lenses for various applications. Commercial tools rely on finite differencing and sampling methods to perform gradient-based optimization of lens design objectives. Recently, differentiable rendering techniques have enabled more efficient gradient calculation of these objectives. However, these techniques are unable to optimize for light throughput, often an important metric for many applications. We develop a method for calculating the gradients of optical systems with respect to both focus and light throughput. We formulate lens performance as an integral loss over a dynamic domain, which allows for the use of differentiable rendering techniques to calculate the required gradients. We also develop a ray tracer specifically designed for refractive lenses and derive formulas for calculating gradients that simultaneously optimize for focus and light throughput. Explicitly optimizing for light throughput produces lenses that outperform traditional optimized lenses that tend to prioritize for only focus. To evaluate our lens designs, we simulate various applications where our lenses:
In this paper, we present a low-budget and high-authenticity bidirectional telepresence system, Tele-Aloha, targeting peer-to-peer communication scenarios. Compared to previous systems, Tele-Aloha utilizes only four sparse RGB cameras, one consumer-grade GPU, and one autostereoscopic screen to achieve high-resolution (2048x2048), real-time (30 fps), low-latency (less than 150ms) and robust distant communication. As the core of Tele-Aloha, we propose an efficient novel view synthesis algorithm for upper-body. Firstly, we design a cascaded disparity estimator for obtaining a robust geometry cue. Additionally a neural rasterizer via Gaussian Splatting is introduced to project latent features onto target view and to decode them into a reduced resolution. Further, given the high-quality captured data, we leverage weighted blending mechanism to refine the decoded image into the final resolution of 2K. Exploiting world-leading autostereoscopic display and low-latency iris tracking, users are able to experience a strong three-dimensional sense even without any wearable head-mounted display device. Altogether, our telepresence system demonstrates the sense of co-presence in real-life experiments, inspiring the next generation of communication.
We propose X-Portrait, an innovative conditional diffusion model tailored for generating expressive and temporally coherent portrait animation. Specifically, given a single portrait as appearance reference, we aim to animate it with motion derived from a driving video, capturing both highly dynamic and subtle facial expressions along with wide-range head movements. As its core, we leverage the generative prior of a pre-trained diffusion model as the rendering backbone, while achieve fine-grained head pose and expression control with novel controlling signals within the framework of ControlNet. In contrast to conventional coarse explicit controls such as facial landmarks, our motion control module is learned to interpret the dynamics directly from the original driving RGB inputs. The motion accuracy is further enhanced with a patch-based local control module that effectively enhance the motion attention to small-scale nuances like eyeball positions. Notably, to mitigate the identity leakage from the driving signals, we train our motion control modules with scaling-augmented cross-identity images, ensuring maximized disentanglement from the appearance reference modules. Experimental results demonstrate the universal effectiveness of X-Portrait across a diverse range of facial portraits and expressive driving sequences, and showcase its proficiency in generating captivating portrait animations with consistently maintained identity characteristics.
Motions in a video primarily consist of camera motion, induced by camera movement, and object motion, resulting from object movement. Accurate control of both camera and object motion is essential for video generation. However, existing works either mainly focus on one type of motion or do not clearly distinguish between the two, limiting their control capabilities and diversity. Therefore, this paper presents MotionCtrl, a unified and flexible motion controller for video generation designed to effectively and independently control camera and object motion. The architecture and training strategy of MotionCtrl are carefully devised, taking into account the inherent properties of camera motion, object motion, and imperfect training data. Compared to previous methods, MotionCtrl offers three main advantages: 1) It effectively and independently controls camera motion and object motion, enabling more fine-grained motion control and facilitating flexible and diverse combinations of both types of motion. 2) Its motion conditions are determined by camera poses and trajectories, which are appearance-free and minimally impact the appearance or shape of objects in generated videos. 3) It is a relatively generalizable model that can adapt to a wide array of camera poses and trajectories once trained. Extensive qualitative and quantitative experiments have been conducted to demonstrate the superiority of MotionCtrl over existing methods. Project page: https://wzhouxiff.github.io/projects/MotionCtrl/ .
Direct-a-Video: Customized Video Generation with User-Directed Camera Movement and Object Motion
PDF ↗Recent text-to-video diffusion models have achieved impressive progress. In practice, users often desire the ability to control object motion and camera movement independently for customized video creation. However, current methods lack the focus on separately controlling object motion and camera movement in a decoupled manner, which limits the controllability and flexibility of text-to-video models. In this paper, we introduce Direct-a-Video, a system that allows users to independently specify motions for multiple objects as well as camera’s pan and zoom movements, as if directing a video. We propose a simple yet effective strategy for the decoupled control of object motion and camera movement. Object motion is controlled through spatial cross-attention modulation using the model’s inherent priors, requiring no additional optimization. For camera movement, we introduce new temporal cross-attention layers to interpret quantitative camera movement parameters. We further employ an augmentation-based approach to train these layers in a self-supervised manner on a small-scale dataset, eliminating the need for explicit motion annotation. Both components operate independently, allowing individual or combined control, and can generalize to open-domain scenarios. Extensive experiments demonstrate the superiority and effectiveness of our method. Project page and code are available at https://direct-a-video.github.io/.