CVPR 2022 论文列表

We present a novel high-fidelity generative adversarial network (GAN) inversion framework that enables attribute editing with image-specific details well-preserved (e.g., background, appearance, and illumination). We first analyze the challenges of high-fidelity GAN inversion from the perspective of lossy data compression. With a low bit-rate latent code, previous works have difficulties in preserving high-fidelity details in reconstructed and edited images. Increasing the size of a latent code can improve the accuracy of GAN inversion but at the cost of inferior editability. To improve image fidelity without compromising editability, we propose a distortion consultation approach that employs a distortion map as a reference for high-fidelity reconstruction. In the distortion consultation inversion (DCI), the distortion map is first projected to a high-rate latent map, which then complements the basic low-rate latent code with more details via consultation fusion. To achieve high-fidelity editing, we propose an adaptive distortion alignment (ADA) module with a self-supervised training scheme, which bridges the gap between the edited and inversion images. Extensive experiments in the face and car domains show a clear improvement in both inversion and editing quality.

To segment 4K or 6K ultra high-resolution images needs extra computation consideration in image segmentation. Common strategies, such as down-sampling, patch cropping, and cascade model, cannot address well the balance issue between accuracy and computation cost. Motivated by the fact that humans distinguish among objects continuously from coarse to precise levels, we propose the Continuous Refinement Model(CRM) for the ultra high-resolution segmentation refinement task. CRM continuously aligns the feature map with the refinement target and aggregates features to reconstruct these images' details. Besides, our CRM shows its significant generalization ability to fill the resolution gap between low-resolution training images and ultra high-resolution testing ones. We present quantitative performance evaluation and visualization to show that our proposed method is fast and effective on image segmentation refinement.

Most self-supervised video representation learning approaches focus on action recognition. In contrast, in this paper we focus on self-supervised video learning for movie understanding and propose a novel hierarchical self-supervised pretraining strategy that separately pretrains each level of our hierarchical movie understanding model. Specifically, we propose to pretrain the low-level video backbone using a contrastive learning objective, while pretrain the higher-level video contextualizer using an event mask prediction task, which enables the usage of different data sources for pretraining different levels of the hierarchy. We first show that our self-supervised pretraining strategies are effective and lead to improved performance on all tasks and metrics on VidSitu benchmark (e.g., improving on semantic role prediction from 47% to 61% CIDEr scores). We further demonstrate the effectiveness of our contextualized event features on LVU tasks, both when used alone and when combined with instance features, showing their complementarity.

Dimensionality reduction is crucial both for visualization and preprocessing high dimensional data for machine learning. We introduce a novel method based on a hierarchy built on 1-nearest neighbor graphs in the original space which is used to preserve the grouping properties of the data distribution on multiple levels. The core of the proposal is an optimization-free projection that is competitive with the latest versions of t-SNE and UMAP in performance and visualization quality while being an order of magnitude faster at run-time. Furthermore, its interpretable mechanics, the ability to project new data, and the natural separation of data clusters in visualizations make it a general purpose unsupervised dimension reduction technique. In the paper, we argue about the soundness of the proposed method and evaluate it on a diverse collection of datasets with sizes varying from 1K to 11M samples and dimensions from 28 to 16K. We perform comparisons with other state-of-the-art methods on multiple metrics and target dimensions highlighting its efficiency and performance. Code is available at https://github.com/koulakis/h-nne

Video captioning aims to generate natural language descriptions according to the content, where representation learning plays a crucial role. Existing methods are mainly developed within the supervised learning framework via word-by-word comparison of the generated caption against the ground-truth text without fully exploiting linguistic semantics. In this work, we propose a hierarchical modular network to bridge video representations and linguistic semantics from three levels before generating captions. In particular, the hierarchy is composed of: (I) Entity level, which highlights objects that are most likely to be mentioned in captions. (II) Predicate level, which learns the actions conditioned on highlighted objects and is supervised by the predicate in captions. (III) Sentence level, which learns the global semantic representation and is supervised by the whole caption. Each level is implemented by one module. Extensive experimental results show that the proposed method performs favorably against the state-of-the-art models on the two widely-used benchmarks: MSVD 104.0% and MSR-VTT 51.5% in CIDEr score. Code will be made available at https://github.com/MarcusNerva/HMN.

Accurately predicting the future motions of surrounding traffic agents is critical for the safety of autonomous vehicles. Recently, vectorized approaches have dominated the motion prediction community due to their capability of capturing complex interactions in traffic scenes. However, existing methods neglect the symmetries of the problem and suffer from the expensive computational cost, facing the challenge of making real-time multi-agent motion prediction without sacrificing the prediction performance. To tackle this challenge, we propose Hierarchical Vector Transformer (HiVT) for fast and accurate multi-agent motion prediction. By decomposing the problem into local context extraction and global interaction modeling, our method can effectively and efficiently model a large number of agents in the scene. Meanwhile, we propose a translation-invariant scene representation and rotation-invariant spatial learning modules, which extract features robust to the geometric transformations of the scene and enable the model to make accurate predictions for multiple agents in a single forward pass. Experiments show that HiVT achieves the state-of-the-art performance on the Argoverse motion forecasting benchmark with a small model size and can make fast multi-agent motion prediction.

Hyperspectral imaging is an essential imaging modality for a wide range of applications, especially in remote sensing, agriculture, and medicine. Inspired by existing hyperspectral cameras that are either slow, expensive, or bulky, reconstructing hyperspectral images (HSIs) from a low-budget snapshot measurement has drawn wide attention. By mapping a truncated numerical optimization algorithm into a network with a fixed number of phases, recent deep unfolding networks (DUNs) for spectral snapshot compressive sensing (SCI) have achieved remarkable success. However, DUNs are far from reaching the scope of industrial applications limited by the lack of cross-phase feature interaction and adaptive parameter adjustment. In this paper, we propose a novel Hyperspectral Explicable Reconstruction and Optimal Sampling deep Network for SCI, dubbed HerosNet, which includes several phases under the ISTA-unfolding framework. Each phase can flexibly simulate the sensing matrix and contextually adjust the step size in the gradient descent step, and hierarchically fuse and interact the hidden states of previous phases to effectively recover current HSI frames in the proximal mapping step. Simultaneously, a hardware-friendly optimal binary mask is learned end-to-end to further improve the reconstruction performance. Finally, our HerosNet is validated to outperform the state-of-the-art methods on both simulation and real datasets by large margins. The source code is available at https://github.com/jianzhangcs/HerosNet.

In this paper, we propose HeadNeRF, a novel NeRF-based parametric head model that integrates the neural radiance field to the parametric representation of the human head. It can render high fidelity head images in real-time on modern GPUs, and supports directly controlling the generated images' rendering pose and various semantic attributes. Different from existing related parametric models, we use the neural radiance fields as a novel 3D proxy instead of the traditional 3D textured mesh, which makes that HeadNeRF is able to generate high fidelity images. However, the computationally expensive rendering process of the original NeRF hinders the construction of the parametric NeRF model. To address this issue, we adopt the strategy of integrating 2D neural rendering to the rendering process of NeRF and design novel loss terms. As a result, the rendering speed of HeadNeRF can be significantly accelerated, and the rendering time of one frame is reduced from 5s to 25ms. The well designed loss terms also improve the rendering accuracy, and the fine-level details of the human head, such as the gaps between teeth, wrinkles, and beards, can be represented and synthesized by HeadNeRF. Extensive experimental results and several applications demonstrate its effectiveness. The trained parametric model is available at https://github.com/CrisHY1995/headnerf.

In many real-life image analysis applications, particularly in biomedical research domains, the objects of interest undergo multiple transformations that alters their visual properties while keeping the semantic content unchanged. Disentangling images into semantic content factors and transformations can provide significant benefits into many domain-specific image analysis tasks. To this end, we propose a generic unsupervised framework, Harmony, that simultaneously and explicitly disentangles semantic content from multiple parameterized transformations. Harmony leverages a simple cross-contrastive learning framework with multiple explicitly parameterized latent representations to disentangle content from transformations. To demonstrate the efficacy of Harmony, we apply it to disentangle image semantic content from several parameterized transformations (rotation, translation, scaling, and contrast). Harmony achieves significantly improved disentanglement over the baseline models on several image datasets of diverse domains. With such disentanglement, Harmony is demonstrated to incentivize bioimage analysis research by modeling structural heterogeneity of macromolecules from cryo-ET images and learning transformation-invariant representations of protein particles from single-particle cryo-EM images. Harmony also performs very well in disentangling content from 3D transformations and can perform coarse and fast alignment of 3D cryo-ET subtomograms. Therefore, Harmony is generalizable to many other imaging domains and can potentially be extended to domains beyond imaging as well.

Hands are often severely occluded by objects, which makes 3D hand mesh estimation challenging. Previous works often have disregarded information at occluded regions. However, we argue that occluded regions have strong correlations with hands so that they can provide highly beneficial information for complete 3D hand mesh estimation. Thus, in this work, we propose a novel 3D hand mesh estimation network HandOccNet, that can fully exploits the information at occluded regions as a secondary means to enhance image features and make it much richer. To this end, we design two successive Transformer-based modules, called feature injecting transformer (FIT) and self-enhancing transformer (SET). FIT injects hand information into occluded region by considering their correlation. SET refines the output of FIT by using a self-attention mechanism. By injecting the hand information to the occluded region, our HandOccNet reaches the state-of-the-art performance on 3D hand mesh benchmarks that contain challenging hand-object occlusions. The codes are available in: https://github.com/namepllet/HandOccNet.

Neural Radiance Fields (NeRF) has recently gained popularity for its impressive novel view synthesis ability. This paper studies the problem of hallucinated NeRF: i.e., recovering a realistic NeRF at a different time of day from a group of tourism images. Existing solutions adopt NeRF with a controllable appearance embedding to render novel views under various conditions, but they cannot render view-consistent images with an unseen appearance. To solve this problem, we present an end-to-end framework for constructing a hallucinated NeRF, dubbed as Ha-NeRF. Specifically, we propose an appearance hallucination module to handle time-varying appearances and transfer them to novel views. Considering the complex occlusions of tourism images, we introduce an anti-occlusion module to decompose the static subjects for visibility accurately. Experimental results on synthetic data and real tourism photo collections demonstrate that our method can hallucinate the desired appearances and render occlusion-free images from different views. The project and supplementary materials are available at https://rover-xingyu.github.io/Ha-NeRF/.

Removing hair from portrait images is challenging due to the complex occlusions between hair and face, as well as the lack of paired portrait data with/without hair. To this end, we present a dataset and a baseline method for removing hair from portrait images using generative adversarial networks (GANs). Our core idea is to train a fully connected network HairMapper to find the direction of hair removal in the latent space of StyleGAN for the training stage. We develop a new separation boundary and diffuse method to generate paired training data for males, and a novel "female-male-bald" pipeline for paired data of females. Experiments show that our method can naturally deal with portrait images with variations on gender, age, etc. We validate the superior performance of our method by comparing it to state-of-the-art methods through extensive experiments and user studies. We also demonstrate its applications in hair design and 3D face reconstruction.

Hair editing is an interesting and challenging problem in computer vision and graphics. Many existing methods require well-drawn sketches or masks as conditional inputs for editing, however these interactions are neither straightforward nor efficient. In order to free users from the tedious interaction process, this paper proposes a new hair editing interaction mode, which enables manipulating hair attributes individually or jointly based on the texts or reference images provided by users. For this purpose, we encode the image and text conditions in a shared embedding space and propose a unified hair editing framework by leveraging the powerful image text representation capability of the Contrastive Language-Image Pre-Training (CLIP) model. With the carefully designed network structures and loss functions, our framework can perform high-quality hair editing in a disentangled manner. Extensive experiments demonstrate the superiority of our approach in terms of manipulation accuracy, visual realism of editing results, and irrelevant attribute preservation.

We present a large-scale study of imitating human demonstrations on tasks that require a virtual robot to search for objects in new environments - (1) ObjectGoal Navigation (e.g. 'find & go to a chair') and (2) Pick&Place (e.g. 'find mug, pick mug, find counter, place mug on counter'). First, we develop a virtual teleoperation data-collection infrastructure - connecting Habitat simulator running in a web browser to Amazon Mechanical Turk, allowing remote users to teleoperate virtual robots, safely and at scale. We collect 80k demonstrations for ObjectNav and 12k demonstrations for Pick&Place, which is an order of magnitude larger than existing human demonstration datasets in simulation or on real robots. Our virtual teleoperation data contains 29.3M actions, and is equivalent to 22.6k hours of real-world teleoperation time, and illustrates rich, diverse strategies for solving the tasks. Second, we use this data to answer the question - how does large-scale imitation learning (IL) (which has not been hitherto possible) compare to reinforcement learning (RL) (which is the status quo)? On ObjectNav, we find that IL (with no bells or whistles) using 70k human demonstrations outperforms RL using 240k agent-gathered trajectories. This effectively establishes an 'exchange rate' - a single human demonstration appears to be worth 4 agent-gathered ones. More importantly, we find the IL-trained agent learns efficient object-search behavior from humans - it peeks into rooms, checks corners for small objects, turns in place to get a panoramic view - none of these are exhibited as prominently by the RL agent, and to induce these behaviors via contemporary RL techniques would require tedious reward engineering. Finally, accuracy vs. training data size plots show promising scaling behavior, suggesting that simply collecting more demonstrations is likely to advance the state of art further. On Pick&Place, the comparison is starker - IL agents achieve 18% success on episodes with new object-receptacle locations when trained with 9.5k human demonstrations, while RL agents fail to get beyond 0%. Overall, our work provides compelling evidence for investing in large-scale imitation learning. Project page: https://ram81.github.io/projects/habitat-web.

Capturing and rendering life-like hair is particularly challenging due to its fine geometric structure, complex physical interaction and the non-trivial visual appearance that must be captured. Yet, it is a critical component to create believable avatars. In this paper, we address the aforementioned problems: 1) we use a novel, volumetric hair representation that is composed of thousands of primitives. Each primitive can be rendered efficiently, yet realistically, by building on the latest advances in neural rendering. 2) To have a reliable control signal, we present a novel way of tracking hair on strand level. To keep the computational effort manageable, we use guide hairs and classic techniques to expand those into a dense head of hair. 3) To better enforce temporal consistency and generalization ability of our model, we further optimize the 3D scene flow of our representation with multiview optical flow, using volumetric raymarching. Our method can not only create realistic renders of recorded multi-view sequences, but also create renderings for new hair configurations by providing new control signals. We compare our method with existing work on viewpoint synthesis and drivable animation and achieve state-of-the-art results. https://ziyanw1.github.io/hvh/

We propose Human-centered 4D Scene Capture (HSC4D) to accurately and efficiently create a dynamic digital world, containing large-scale indoor-outdoor scenes, diverse human motions, and rich interactions between humans and environments. Using only body-mounted IMUs and LiDAR, HSC4D is space-free without any external devices' constraints and map-free without pre-built maps. Considering that IMUs can capture human poses but always drift for long-period use, while LiDAR is stable for global localization but rough for local positions and orientations, HSC4D makes both sensors complement each other by a joint optimization and achieves promising results for long-term capture. Relationships between humans and environments are also explored to make their interaction more realistic. To facilitate many down-stream tasks, like AR, VR, robots, autonomous driving, etc., we propose a dataset containing three large scenes (1k-5k m^2 ) with accurate dynamic human motions and locations. Diverse scenarios (climbing gym, multi-story building, slope, etc.) and challenging human activities (exercising, walking up/down stairs, climbing, etc.) demonstrate the effectiveness and the generalization ability of HSC4D. The dataset and code is available at lidarhumanmotion.net/hsc4d.

Capsule networks are designed to present the objects by a set of parts and their relationships, which provide an insight into the procedure of visual perception. Although recent works have shown the success of capsule networks on simple objects like digits, the human faces with homologous structures, which are suitable for capsules to describe, have not been explored. In this paper, we propose a Hierarchical Parsing Capsule Network (HP-Capsule) for unsupervised face subpart-part discovery. When browsing large-scale face images without labels, the network first encodes the frequently observed patterns with a set of explainable subpart capsules. Then, the subpart capsules are assembled into part-level capsules through a Transformer-based Parsing Module (TPM) to learn the compositional relations between them. During training, as the face hierarchy is progressively built and refined, the part capsules adaptively encode the face parts with semantic consistency. HP-Capsule extends the application of capsule networks from digits to human faces and takes a step forward to show how the neural networks understand homologous objects without human intervention. Besides, HP-Capsule gives unsupervised face segmentation results by the covered regions of part capsules, enabling qualitative and quantitative evaluation. Experiments on BP4D and Multi-PIE datasets show the effectiveness of our method.

Pre-training has been adopted in a few of recent works for Vision-and-Language Navigation (VLN). However, previous pre-training methods for VLN either lack the ability to predict future actions or ignore the trajectory contexts, which are essential for a greedy navigation process. In this work, to promote the learning of spatio-temporal visual-textual correspondence as well as the agent's capability of decision making, we propose a novel history-and-order aware pre-training paradigm (HOP) with VLN-specific objectives that exploit the past observations and support future action prediction. Specifically, in addition to the commonly used Masked Language Modeling (MLM) and Trajectory-Instruction Matching (TIM), we design two proxy tasks to model temporal order information: Trajectory Order Modeling (TOM) and Group Order Modeling (GOM). Moreover, our navigation action prediction is also enhanced by introducing the task of Action Prediction with History (APH), which takes into account the history visual perceptions. Extensive experimental results on four downstream VLN tasks (R2R, REVERIE, NDH, RxR) demonstrate the effectiveness of our proposed method compared against several state-of-the-art agents.

We present HOI4D, a large-scale 4D egocentric dataset with rich annotations, to catalyze the research of category-level human-object interaction. HOI4D consists of 2.4M RGB-D egocentric video frames over 4000 sequences collected by 9 participants interacting with 800 different object instances from 16 categories over 610 different indoor rooms. Frame-wise annotations for panoptic segmentation, motion segmentation, 3D hand pose, category-level object pose and hand action have also been provided, together with reconstructed object meshes and scene point clouds. With HOI4D, we establish three benchmarking tasks to promote category-level HOI from 4D visual signals including semantic segmentation of 4D dynamic point cloud sequences, category-level object pose tracking, and egocentric action segmentation with diverse interaction targets. In-depth analysis shows HOI4D poses great challenges to existing methods and produces huge research opportunities.