CVPR 2022 论文列表

We introduce Ego4D, a massive-scale egocentric video dataset and benchmark suite. It offers 3,670 hours of daily-life activity video spanning hundreds of scenarios (household, outdoor, workplace, leisure, etc.) captured by 931 unique camera wearers from 74 worldwide locations and 9 different countries. The approach to collection is designed to uphold rigorous privacy and ethics standards, with consenting participants and robust de-identification procedures where relevant. Ego4D dramatically expands the volume of diverse egocentric video footage publicly available to the research community. Portions of the video are accompanied by audio, 3D meshes of the environment, eye gaze, stereo, and/or synchronized videos from multiple egocentric cameras at the same event. Furthermore, we present a host of new benchmark challenges centered around understanding the first-person visual experience in the past (querying an episodic memory), present (analyzing hand-object manipulation, audio-visual conversation, and social interactions), and future (forecasting activities). By publicly sharing this massive annotated dataset and benchmark suite, we aim to push the frontier of first-person perception. Project page: https://ego4d-data.org/

Unsupervised generation of high-quality multi-view-consistent images and 3D shapes using only collections of single-view 2D photographs has been a long-standing challenge. Existing 3D GANs are either compute-intensive or make approximations that are not 3D-consistent; the former limits quality and resolution of the generated images and the latter adversely affects multi-view consistency and shape quality. In this work, we improve the computational efficiency and image quality of 3D GANs without overly relying on these approximations. We introduce an expressive hybrid explicit-implicit network architecture that, together with other design choices, synthesizes not only high-resolution multi-view-consistent images in real time but also produces high-quality 3D geometry. By decoupling feature generation and neural rendering, our framework is able to leverage state-of-the-art 2D CNN generators, such as StyleGAN2, and inherit their efficiency and expressiveness. We demonstrate state-of-the-art 3D-aware synthesis with FFHQ and AFHQ Cats, among other experiments.

Locating 3D objects from a single RGB image via Perspective-n-Points (PnP) is a long-standing problem in computer vision. Driven by end-to-end deep learning, recent studies suggest interpreting PnP as a differentiable layer, so that 2D-3D point correspondences can be partly learned by backpropagating the gradient w.r.t. object pose. Yet, learning the entire set of unrestricted 2D-3D points from scratch fails to converge with existing approaches, since the deterministic pose is inherently non-differentiable. In this paper, we propose the EPro-PnP, a probabilistic PnP layer for general end-to-end pose estimation, which outputs a distribution of pose on the SE(3) manifold, essentially bringing categorical Softmax to the continuous domain. The 2D-3D coordinates and corresponding weights are treated as intermediate variables learned by minimizing the KL divergence between the predicted and target pose distribution. The underlying principle unifies the existing approaches and resembles the attention mechanism. EPro-PnP significantly outperforms competitive baselines, closing the gap between PnP-based method and the task-specific leaders on the LineMOD 6DoF pose estimation and nuScenes 3D object detection benchmarks.

Recently, learned image compression techniques have achieved remarkable performance, even surpassing the best manually designed lossy image coders. They are promising to be large-scale adopted. For the sake of practicality, a thorough investigation of the architecture design of learned image compression, regarding both compression performance and running speed, is essential. In this paper, we first propose uneven channel-conditional adaptive coding, motivated by the observation of energy compaction in learned image compression. Combining the proposed uneven grouping model with existing context models, we obtain a spatial-channel contextual adaptive model to improve the coding performance without damage to running speed. Then we study the structure of the main transform and propose an efficient model, ELIC, to achieve state-of-the-art speed and compression ability. With superior performance, the proposed model also supports extremely fast preview decoding and progressive decoding, which makes the coming application of learning-based image compression more promising.

Fine-grained image classification is a challenging computer vision task where various species share similar visual appearances, resulting in misclassification if merely based on visual clues. Therefore, it is helpful to leverage additional information, e.g., the locations and dates for data shooting, which can be easily accessible but rarely exploited. In this paper, we first demonstrate that existing multimodal methods fuse multiple features only on a single dimension, which essentially has insufficient help in feature discrimination. To fully explore the potential of multimodal information, we propose a dynamic MLP on top of the image representation, which interacts with multimodal features at a higher and broader dimension. The dynamic MLP is an efficient structure parameterized by the learned embeddings of variable locations and dates. It can be regarded as an adaptive nonlinear projection for generating more discriminative image representations in visual tasks. To our best knowledge, it is the first attempt to explore the idea of dynamic networks to exploit multimodal information in fine-grained image classification tasks. Extensive experiments demonstrate the effectiveness of our method. The t-SNE algorithm visually indicates that our technique improves the recognizability of image representations that are visually similar but with different categories. Furthermore, among published works across multiple fine-grained datasets, dynamic MLP consistently achieves SOTA results and takes third place in the iNaturalist challenge at FGVC8. Code is available at https://github.com/megvii-research/DynamicMLPForFinegrained.

Visual vibrometry is a highly useful tool for remote capture of audio, as well as the physical properties of materials, human heart rate, and more. While visually-observable vibrations can be captured directly with a high-speed camera, minute imperceptible object vibrations can be optically amplified by imaging the displacement of a speckle pattern, created by shining a laser beam on the vibrating surface. In this paper, we propose a novel method for sensing vibrations at high speeds (up to 63kHz), for multiple scene sources at once, using sensors rated for only 130Hz operation. Our method relies on simultaneously capturing the scene with two cameras equipped with rolling and global shutter sensors, respectively. The rolling shutter camera captures distorted speckle images that encode the highspeed object vibrations. The global shutter camera captures undistorted reference images of the speckle pattern, helping to decode the source vibrations. We demonstrate our method by capturing vibration caused by audio sources (e.g. speakers, human voice, and musical instruments) and analyzing the vibration modes of a tuning fork.

Learning spatial-temporal relation among multiple actors is crucial for group activity recognition. Different group activities often show the diversified interactions between actors in the video. Hence, it is often difficult to model complex group activities from a single view of spatial-temporal actor evolution. To tackle this problem, we propose a distinct Dual-path Actor Interaction (Dual-AI) framework, which flexibly arranges spatial and temporal transformers in two complementary orders, enhancing actor relations by integrating merits from different spatio-temporal paths. Moreover, we introduce a novel Multi-scale Actor Contrastive Loss (MAC-Loss) between two interactive paths of Dual-AI. Via self-supervised actor consistency in both frame and video levels, MAC-Loss can effectively distinguish individual actor representations to reduce action confusion among different actors. Consequently, our Dual-AI can boost group activity recognition by fusing such discriminative features of different actors. To evaluate the proposed approach, we conduct extensive experiments on the widely used benchmarks, including Volleyball, Collective Activity, and NBA datasets. The proposed Dual-AI achieves state-of-the-art performance on all these datasets. It is worth noting the proposed Dual-AI with 50% training data outperforms a number of recent approaches with 100% training data. This confirms the generalization power of Dual-AI for group activity recognition, even under the challenging scenarios of limited supervision.

Accurate change detection enables a wide range of tasks in visual surveillance, anomaly detection and mobile robotics. However, contemporary change detection approaches assume an ideal matching between the current and stored scenes, whereas only coarse matching is possible in real-world scenarios. Thus, contemporary approaches fail to show the reported performance in real-world settings. To overcome this limitation, we propose SimSaC. SimSaC concurrently conducts scene flow estimation and change detection and is able to detect changes with imperfect matches. To train SimSaC without additional manual labeling, we propose a training scheme with random geometric transformations and the cut-paste method. Moreover, we design an evaluation protocol which reflects performance in real-world settings. In designing the protocol, we collect a test benchmark dataset, which we claim as another contribution. Our comprehensive experiments verify that SimSaC displays robust performance even given imperfect matches and the performance margin compared to contemporary approaches is huge.

Due to the sophisticated imaging process, an identical scene captured by different cameras could exhibit distinct imaging patterns, introducing distinct proficiency among the super-resolution (SR) models trained on images from different devices. In this paper, we investigate a novel and practical task coded cross-device SR, which strives to adapt a real-world SR model trained on the paired images captured by one camera to low-resolution (LR) images captured by arbitrary target devices. The proposed task is highly challenging due to the absence of paired data from various imaging devices. To address this issue, we propose an unsupervised domain adaptation mechanism for real-world SR, named Dual ADversarial Adaptation (DADA), which only requires LR images in the target domain with available real paired data from a source camera. DADA employs the Domain-Invariant Attention (DIA) module to establish the basis of target model training even without HR supervision. Furthermore, the dual framework of DADA facilitates an Inter-domain Adversarial Adaptation (InterAA) in one branch for two LR input images from two domains, and an Intra-domain Adversarial Adaptation (IntraAA) in two branches for an LR input image. InterAA and IntraAA together improve the model transferability from the source domain to the target. We empirically conduct experiments under six Real to Real adaptation settings among three different cameras, and achieve superior performance compared with existing state-of-the-art approaches. We also evaluate the proposed DADA to address the adaptation to the video camera, which presents a promising research topic to promote the wide applications of real-world super-resolution. Our source code is publicly available at https://github.com/lonelyhope/DADA.git.

Image manipulation dates back long before the deep learning era. The classical prevailing approaches were based on maximizing patch similarity between the input and generated output. Recently, single-image GANs were introduced as a superior and more sophisticated solution to image manipulation tasks. Moreover, they offered the opportunity not only to manipulate a given image, but also to generate a large and diverse set of different outputs from a single natural image. This gave rise to new tasks, which are considered "DL-only". However, despite their impressiveness, single-image GANs require long training time (usually hours) for each image and each task and often suffer from visual artifacts. In this paper we revisit the classical patch-based methods, and show that - unlike previously believed -- classical methods can be adapted to tackle these novel "GAN-only" tasks. Moreover, they do so better and faster than single-image GAN-based methods. More specifically, we show that: (i) by introducing slight modifications, classical patch-based methods are able to unconditionally generate diverse images based on a single natural image; (ii) the generated output visual quality exceeds that of single-image GANs by a large margin (confirmed both quantitatively and qualitatively); (iii) they are orders of magnitude faster (runtime reduced from hours to seconds).

As an inherently ill-posed problem, depth estimation from single images is the most challenging part of monocular 3D object detection (M3OD). Many existing methods rely on preconceived assumptions to bridge the missing spatial information in monocular images, and predict a sole depth value for every object of interest. However, these assumptions do not always hold in practical applications. To tackle this problem, we propose a depth solving system that fully explores the visual clues from the subtasks in M3OD and generates multiple estimations for the depth of each target. Since the depth estimations rely on different assumptions in essence, they present diverse distributions. Even if some assumptions collapse, the estimations established on the remaining assumptions are still reliable. In addition, we develop a depth selection and combination strategy. This strategy is able to remove abnormal estimations caused by collapsed assumptions, and adaptively combine the remaining estimations into a single one. In this way, our depth solving system becomes more precise and robust. Exploiting the clues from multiple subtasks of M3OD and without introducing any extra information, our method surpasses the current best method by more than 20% relatively on the Moderate level of test split in the KITTI 3D object detection benchmark, while still maintaining real-time efficiency.

Digitizing physical objects into the virtual world has the potential to unlock new research and applications in embodied AI and mixed reality. This work focuses on recreating interactive digital twins of real-world articulated objects, which can be directly imported into virtual environments. We introduce Ditto to learn articulation model estimation and 3D geometry reconstruction of an articulated object through interactive perception. Given a pair of visual observations of an articulated object before and after interaction, Ditto reconstructs part-level geometry and estimates the articulation model of the object. We employ implicit neural representations for joint geometry and articulation modeling. Our experiments show that Ditto effectively builds digital twins of articulated objects in a category-agnostic way. We also apply Ditto to real-world objects and deploy the recreated digital twins in physical simulation. Code and additional results are available at https://ut-austin-rpl.github.io/Ditto/

Recent progress on neural architecture search (NAS) has demonstrated exciting results on automating deep network architecture designs. In order to overcome the unaffordable complexity of training each candidate architecture from scratch, the state-of-the-art one-shot NAS approaches adopt a weight-sharing strategy to improve training efficiency. Although the computational cost is greatly reduced, such one-shot process introduces a severe weight coupling problem that largely degrades the evaluation accuracy of each candidate. The existing approaches often address the problem by shrinking the search space, model distillation, or few-shot training. Instead, in this paper, we propose a novel distribution consistent one-shot neural architecture search algorithm. We first theoretically investigate how the weight coupling problem affects the network searching performance from a parameter distribution perspective, and then propose a novel supernet training strategy with a Distribution Consistent Constraint that can provide a good measurement for the extent to which two architectures can share weights. Our strategy optimizes the supernet through iteratively inferring network weights and corresponding local sharing states. Such joint optimization of supernet's weights and topologies can diminish the discrepancy between the weights inherited from the supernet and the ones that are trained with a stand-alone model. As a result, it enables a more accurate model evaluation phase and leads to a better searching performance. We conduct extensive experiments on benchmark datasets with multiple searching spaces. The resulting architecture achieves superior performance over the current state-of-the-art NAS algorithms with comparable search costs, which demonstrates the efficacy of our approach.

The CLIP network measures the similarity between natural text and images; in this work, we investigate the entanglement of the representation of word images and natural images in its image encoder. First, we find that the image encoder has an ability to match word images with natural images of scenes described by those words. This is consistent with previous research that suggests that the meaning and the spelling of a word might be entangled deep within the network. On the other hand, we also find that CLIP has a strong ability to match nonsense words, suggesting that processing of letters is separated from processing of their meaning. To explicitly determine whether the spelling capability of CLIP is separable, we devise a procedure for identifying representation subspaces that selectively isolate or eliminate spelling capabilities. We benchmark our methods against a range of retrieval tasks, and we also test them by measuring the appearance of text in CLIP-guided generated images. We find that our methods are able to cleanly separate spelling capabilities of CLIP from the visual processing of natural images.

We study the problem of compositional zero-shot learning for object-attribute recognition. Prior works use visual features extracted with a backbone network, pre-trained for object classification and thus do not capture the subtly distinct features associated with attributes. To overcome this challenge, these studies employ supervision from the linguistic space, and use pre-trained word embeddings to better separate and compose attribute-object pairs for recognition. Analogous to linguistic embedding space, which already has unique and agnostic embeddings for object and attribute, we shift the focus back to the visual space and propose a novel architecture that can disentangle attribute and object features in the visual space. We use visual decomposed features to hallucinate embeddings that are representative for the seen and novel compositions to better regularize the learning of our model. Extensive experiments show that our method outperforms existing work with significant margin on three datasets: MIT-States, UT-Zappos, and a new benchmark created based on VAW.

Guided depth super-resolution (GDSR) is an essential topic in multi-modal image processing, which reconstructs high-resolution (HR) depth maps from low-resolution ones collected with suboptimal conditions with the help of HR RGB images of the same scene. To solve the challenges in interpreting the working mechanism, extracting cross-modal features and RGB texture over-transferred, we propose a novel Discrete Cosine Transform Network (DCTNet) to alleviate the problems from three aspects. First, the Discrete Cosine Transform (DCT) module reconstructs the multi-channel HR depth features by using DCT to solve the channel-wise optimization problem derived from the image domain. Second, we introduce a semi-coupled feature extraction module that uses shared convolutional kernels to extract common information and private kernels to extract modality-specific information. Third, we employ an edge attention mechanism to highlight the contours informative for guided upsampling. Extensive quantitative and qualitative evaluations demonstrate the effectiveness of our DCTNet, which outperforms previous state-of-the-art methods with a relatively small number of parameters. Codes are available at https://github.com/Zhaozixiang1228/GDSR-DCTNet.

We present a super-fast convergence approach to reconstructing the per-scene radiance field from a set of images that capture the scene with known poses. This task, which is often applied to novel view synthesis, is recently revolutionized by Neural Radiance Field (NeRF) for its state-of-the-art quality and flexibility. However, NeRF and its variants require a lengthy training time ranging from hours to days for a single scene. In contrast, our approach achieves NeRF-comparable quality and converges rapidly from scratch in less than 15 minutes with a single GPU. We adopt a representation consisting of a density voxel grid for scene geometry and a feature voxel grid with a shallow network for complex view-dependent appearance. Modeling with explicit and discretized volume representations is not new, but we propose two simple yet non-trivial techniques that contribute to fast convergence speed and high-quality output. First, we introduce the post-activation interpolation on voxel density, which is capable of producing sharp surfaces in lower grid resolution. Second, direct voxel density optimization is prone to suboptimal geometry solutions, so we robustify the optimization process by imposing several priors. Finally, evaluation on five inward-facing benchmarks shows that our method matches, if not surpasses, NeRF's quality, yet it only takes about 15 minutes to train from scratch for a new scene. We will make our code publicly available.

Diffusion probabilistic models (DPMs) have achieved remarkable quality in image generation that rivals GANs'. But unlike GANs, DPMs use a set of latent variables that lack semantic meaning and cannot serve as a useful representation for other tasks. This paper explores the possibility of using DPMs for representation learning and seeks to extract a meaningful and decodable representation of an input image via autoencoding. Our key idea is to use a learnable encoder for discovering the high-level semantics, and a DPM as the decoder for modeling the remaining stochastic variations. Our method can encode any image into a two-part latent code, where the first part is semantically meaningful and linear, and the second part captures stochastic details, allowing near-exact reconstruction. This capability enables challenging applications that currently foil GAN-based methods, such as attribute manipulation on real images. We also show that this two-level encoding improves denoising efficiency and naturally facilitates various downstream tasks including few-shot conditional sampling.