ICCV 2023 论文列表

3D plane recovery from a single image can usually be divided into several subtasks of plane detection, segmentation, parameter estimation and possibly depth estimation. Previous works tend to solve it by either extending the RCNN-based segmentation network or the dense pixel embedding-based clustering framework. However, none of them tried to integrate above related subtasks into a unified framework but treated them separately and sequentially, which we suspect is potentially a main source of performance limitation for existing approaches. Motivated by this finding and the success of query-based learning in enriching reasoning among semantic entities, in this paper, we propose PlaneRecTR, a Transformer-based architecture, which for the first time unifies all subtasks related to single-view plane recovery with a single compact model. Extensive quantitative and qualitative experiments demonstrate that our proposed unified learning achieves mutual benefits across subtasks, obtaining a new state-of-the-art performance on public ScanNet and NYUv2-Plane datasets.

Planar object tracking is a critical computer vision problem and has drawn increasing interest owing to its key roles in robotics, augmented reality, etc. Despite rapid progress, its further development, especially in the deep learning era, is largely hindered due to the lack of large-scale challenging benchmarks. Addressing this, we introduce PlanarTrack, a large-scale challenging planar tracking benchmark. Specifically, PlanarTrack consists of 1,000 videos with more than 490K images. All these videos are collected in complex unconstrained scenarios from the wild, which makes PlanarTrack, compared with existing benchmarks, more challenging but realistic for real-world applications. To ensure the high-quality annotation, each frame in PlanarTrack is manually labeled using four corners with multiple-round careful inspection and refinement. To our best knowledge, PlanarTrack, to date, is the largest and most challenging dataset dedicated to planar object tracking. In order to analyze the proposed PlanarTrack, we evaluate 10 planar trackers and conduct comprehensive comparisons and in-depth analysis. Our results, not surprisingly, demonstrate that current top-performing planar trackers degenerate significantly on the challenging PlanarTrack and more efforts are needed to improve planar tracking in the future. In addition, we further derive a variant named PlanarTrack_BB for generic object tracking from PlanarTrack. Our evaluation of 10 excellent generic trackers on PlanarTrack_BB manifests that, surprisingly, PlanarTrack_BB is even more challenging than several popular generic tracking benchmarks and more attention should be paid to handle such planar objects, though they are rigid. All benchmarks and evaluations will be released.

We present a simple but effective pixel-level self-supervised distillation framework friendly to dense prediction tasks. Our method, called Pixel-Wise Contrastive Distillation (PCD), distills knowledge by attracting the corresponding pixels from student's and teacher's output feature maps. PCD includes a novel design called SpatialAdaptor which "reshapes" a part of the teacher network while preserving the distribution of its output features. Our ablation experiments suggest that this reshaping behavior enables more informative pixel-to-pixel distillation. Moreover, we utilize a plug-in multi-head self-attention module that explicitly relates the pixels of student's feature maps to enhance the effective receptive field, leading to a more competitive student. PCD outperforms previous self-supervised distillation methods on various dense prediction tasks. A backbone of ResNet-18-FPN distilled by PCD achieves 37.4 AP-bbox and 34.0 AP-mask on COCO dataset using the detector of Mask R-CNN. We hope our study will inspire future research on how to pre-train a small model friendly to dense prediction tasks in a self-supervised fashion.

We present PARQ - a multi-view 3D object detector with transformer and pixel-aligned recurrent queries. Unlike previous works that use learnable features or only encode 3D point positions as queries in the decoder, PARQ leverages appearance-enhanced queries initialized from reference points in 3D space and updates their 3D location with recurrent cross-attention operations. Incorporating pixel-aligned features and cross attention enables the model to encode the necessary 3D-to-2D correspondences and capture global contextual information of the input images. PARQ outperforms prior best methods on the ScanNet and ARKitScenes datasets, learns and detects faster, is more robust to distribution shifts in reference points, can leverage additional input views without retraining, and can adapt inference compute by changing the number of recurrent iterations.

Hyperspectral Image (HSI) reconstruction has made gratifying progress with the deep unfolding framework by formulating the problem into a data module and a prior module. Nevertheless, existing methods still face the problem of insufficient matching with HSI data. The issues lie in three aspects: 1) fixed gradient descent step in the data module while the degradation of HSI is agnostic in the pixel-level. 2) inadequate prior module for 3D HSI cube. 3) stage interaction ignoring the differences in features at different stages. To address these issues, in this work, we propose a Pixel Adaptive Deep Unfolding Transformer (PADUT) for HSI reconstruction. In the data module, a pixel adaptive descent step is employed to focus on pixel-level agnostic degradation. In the prior module, we introduce the Non-local Spectral Transformer (NST) to emphasize the 3D characteristics of HSI for recovering. Moreover, inspired by the diverse expression of features in different stages and depths, the stage interaction is improved by the Fast Fourier Transform (FFT). Experimental results on both simulated and real scenes exhibit the superior performance of our method compared to state-of-the-art HSI reconstruction methods. The code is released at: https://github.com/MyuLi/PADUT

Image diffusion models, trained on massive image collections, have emerged as the most versatile image generator model in terms of quality and diversity. They support inverting real images and conditional (e.g., text) generation, making them attractive for high-quality image editing applications. We investigate how to use such pre-trained image models for text-guided video editing. The critical challenge is to achieve the target edits while still preserving the content of the source video. Our method works in two simple steps: first, we use a pre-trained structure-guided (e.g., depth) image diffusion model to perform text-guided edits on an anchor frame; then, in the key step, we progressively propagate the changes to the future frames via self-attention feature injection to adapt the core denoising step of the diffusion model. We then consolidate the changes by adjusting the latent code for the frame before continuing the process. Our approach is training-free and generalizes to a wide range of edits. We demonstrate the effectiveness of the approach by extensive experimentation and compare it against four different prior and parallel efforts (on ArXiv). We demonstrate that realistic text-guided video edits are possible, without any compute-intensive preprocessing or video-specific finetuning.

Vectorized high-definition map online construction has garnered considerable attention in the field of autonomous driving research. Most existing approaches model changeable map elements using a fixed number of points, or predict local maps in a two-stage autoregressive manner, which may miss essential details and lead to error accumulation. Towards precise map element learning, we propose a simple yet effective architecture named PivotNet, which adopts unified pivot-based map representations and is formulated as a direct set prediction paradigm. Concretely, we first propose a novel Point-to-Line Mask module to encode both the subordinate and geometrical point-line priors in the network. Then, a well-designed Pivot Dynamic Matching module is proposed to model the topology in dynamic point sequences by introducing the concept of sequence matching. Furthermore, to supervise the position and topology of the vectorized point predictions, we propose a Dynamic Vectorized Sequence loss. Extensive experiments and ablations show that PivotNet is remarkably superior to other SOTAs by 5.9 mAP at least. The code will be available soon.

Image distortion by atmospheric turbulence is a stochastic degradation, which is a critical problem in long-range optical imaging systems. A number of research has been conducted during the past decades, including model-based and emerging deep-learning solutions with the help of synthetic data. Although fast and physics-grounded simulation tools have been introduced to help the deep-learning models adapt to real-world turbulence conditions recently, the training of such models only relies on the synthetic data and ground truth pairs. This paper proposes the Physics-integrated Restoration Network (PiRN) to bring the physics-based simulator directly into the training process to help the network to disentangle the stochasticity from the degradation and the underlying image. Furthermore, to overcome the "average effect" introduced by deterministic models and the domain gap between the synthetic and real-world degradation, we further introduce PiRN with Stochastic Refinement (PiRN-SR) to boost its perceptual quality. Overall, our PiRN and PiRN-SR improve the generalization to real-world unknown turbulence conditions and provide a state-of-the-art restoration in both pixel-wise accuracy and perceptual quality.

In this paper, we introduce physics-augmented autoencoder (PAA), a framework for 3D skeleton-based human gait recognition. Specifically, we construct the autoencoder with a graph-convolution-based encoder and a physics-based decoder. The encoder takes the skeleton sequence as input and generates the generalized positions and forces of each joint, which are taken by the decoder to reconstruct the input skeleton based on the Lagrangian dynamics. In this way, the intermediate representations are physically plausible and discriminative. During the inference, the decoder is discared and a RNN-based classifier takes the output of the encoder for gait recognition. We evaluated our proposed method on three benchmark datasets including Gait3D, GREW, and KinectGait. Our method achieves state-of-the-art performance for 3D skeleton-based gait recognition. Furthermore, extensive ablation studies show that our method generalizes better and is more robust with small-scale training data by incorporating the physics knowledge. We also validated the physical plausibility of the intermediate representations by making force predictions on real data with physical annotations.

A camera's auto-white-balance (AWB) module operates under the assumption that there is a single dominant illumination in a captured scene. AWB methods estimate an image's dominant illumination and use it as the target "white point" for correction. However, in natural scenes, there are often many light sources present. We performed a user study that revealed that non-dominant illuminations often produce visually pleasing white-balanced images and, in some cases, are even preferred over the dominant illumination. Motivated by this observation, we revisit AWB to predict a distribution of plausible illuminations for use in white balance. As part of this effort, we extend the Cube++ illumination estimation dataset to provide ground truth illumination distributions per image. Using this new ground truth data, we describe how to train a lightweight neural network method to predict the scene's illumination distribution. We describe how our idea can be used with existing image formats by embedding the estimated distribution in the RAW image to enable users to generate visually plausible white-balance images.

Denoising diffusion models hold great promise for generating diverse and realistic human motions. However, existing motion diffusion models largely disregard the laws of physics in the diffusion process and often generate physically-implausible motions with pronounced artifacts such as floating, foot sliding, and ground penetration. This seriously impacts the quality of generated motions and limits their real-world application. To address this issue, we present a novel physics-guided motion diffusion model (PhysDiff), which incorporates physical constraints into the diffusion process. Specifically, we propose a physics-based motion projection module that uses motion imitation in a physics simulator to project the denoised motion of a diffusion step to a physically-plausible motion. The projected motion is further used in the next diffusion step to guide the denoising diffusion process. Intuitively, the use of physics in our model iteratively pulls the motion toward a physically-plausible space, which cannot be achieved by simple post-processing. Experiments on large-scale human motion datasets show that our approach achieves state-of-the-art motion quality and improves physical plausibility drastically (>78% for all datasets).

Training a generative model with limited number of samples is a challenging task. Current methods primarily rely on few-shot model adaption to train the network. However, in scenarios where data is extremely limited (less than 10), the generative network tends to overfit and suffers from content degradation. To address these problems, we propose a novel phasic content fusing few-shot diffusion model with directional distribution consistency loss, which targets different learning objectives at distinct training stages of the diffusion model. Specifically, we design a phasic training strategy with phasic content fusion to help our model learn content and style information when t is large, and learn local details of target domain when t is small, leading to an improvement in the capture of content, style and local details. Furthermore, we introduce a novel directional distribution consistency loss that ensures the consistency between the generated and source distributions more efficiently and stably than the prior methods, preventing our model from overfitting. Finally, we propose a cross-domain structure guidance strategy that enhances structure consistency during domain adaptation. Theoretical analysis, qualitative and quantitative experiments demonstrate the superiority of our approach in few-shot generative model adaption tasks compared to state-of-the-art methods.

We present a novel motion prior, called PhaseMP, modeling a probability distribution on pose transitions conditioned by a frequency domain feature extracted from a periodic autoencoder. The phase feature further enforces the pose transitions to be unidirectional (i.e. no backward movement in time), from which more stable and natural motions can be generated. Specifically, our motion prior can be useful for accurately estimating 3D human motions in the presence of challenging input data, including long periods of spatial and temporal occlusion, as well as noisy sensor measurements. Through a comprehensive evaluation, we demonstrate the efficacy of our novel motion prior, showcasing its superiority over existing state-of-the-art methods by a significant margin across various applications, including video-to-motion and motion estimation from sparse sensor data, and etc.

Federated learning casts a light on the collaboration of distributed local clients with privacy protected to attain a more generic global model. However, significant distribution shift in input/label space across different clients makes it challenging to well generalize to all clients, which motivates personalized federated learning (PFL). Existing PFL methods typically customize the local model by fine-tuning with limited local supervision and the global model regularizer, which secures local specificity but risks ruining the global discriminative knowledge. In this paper, we propose a novel Personalized Semantics Excitation (PSE) mechanism to breakthrough this limitation by exciting and fusing personalized semantics from the global model during local model customization. Specifically, PSE explores channel-wise gradient differentiation across global and local models to identify important low-level semantics mostly from convolutional layers which are embedded into the client-specific training. In addition, PSE deploys the collaboration of global and local models to enrich high-level feature representations and facilitate the robustness of client classifier through a cross-model attention module. Extensive experiments and analysis on various image classification benchmarks demonstrate the effectiveness and advantage of our method over the state-of-the-art PFL methods.

Approximately, 350 million people, a proportion of 8%, suffer from color vision deficiency (CVD). While image generation algorithms have been highly successful in synthesizing high-quality images, CVD populations are unintentionally excluded from target users and have difficulties understanding the generated images as normal viewers do. Although a straightforward baseline can be formed by combining generation models and recolor compensation methods as the post-processing, the CVD friendliness of the result images is still limited since the input image content of recolor methods is not CVD-oriented and will keep fixed during the recolor compensation process. Besides, the CVD populations can't be fully served since the varying degrees of CVD are often neglected in recoloring methods. Instead, we propose a personalized CVD-friendly image generation algorithm with two key characteristics: (i) generating CVD-oriented images end-to-end; (ii) generating continuous personalized images for people with various CVD types and degrees through disentangling the color representation based on a triple-latent structure. Quantitative experiments and the user study indicate our proposed image generation model can generate practical and compelling results compared to the normal generation model and combination baselines on several datasets.

Wide-scale use of visual surveillance in public spaces puts individual privacy at stake while increasing resource consumption (energy, bandwidth, and computation). Neuromorphic vision sensors (event-cameras) have been recently considered a valid solution to the privacy issue because they do not capture detailed RGB visual information of the subjects in the scene. However, recent deep learning architectures have been able to reconstruct images from event cameras with high fidelity, reintroducing a potential threat to privacy for event-based vision applications. In this paper, we aim to anonymize event-streams to protect the identity of human subjects against such image reconstruction attacks. To achieve this, we propose an end-to-end network architecture jointly optimized for the twofold objective of preserving privacy and performing a downstream task such as person ReId. Our network learns to scramble events, enforcing the degradation of images recovered from the privacy attacker. In this work, we also bring to the community the first ever event-based person ReId dataset gathered to evaluate the performance of our approach. We validate our approach with extensive experiments and report results on the synthetic event data simulated from the publicly available SoftBio dataset and our proposed Event-ReId dataset.

Humans are highly adaptable, swiftly switching between different modes to progressively handle different tasks, situations and contexts. In Human-object interaction (HOI) activities, these modes can be attributed to two mechanisms: (1) the large-scale consistent plan for the whole activity and (2) the small-scale children interactive actions that start and end along the timeline. While neuroscience and cognitive science have confirmed this multi-mechanism nature of human behavior, machine modeling approaches for human motion are trailing behind. While attempted to use gradually morphing structures (e.g., graph attention networks) to model the dynamic HOI patterns, they miss the expeditious and discrete mode-switching nature of the human motion. To bridge that gap, this work proposes to model two concurrent mechanisms that jointly control human motion: the Persistent process that runs continually on the global scale, and the Transient sub-processes that operate intermittently on the local context of the human while interacting with objects. These two mechanisms form an interactive Persistent-Transient Duality that synergistically governs the activity sequences. We model this conceptual duality by a parent-child neural network of Persistent and Transient channels with a dedicated neural module for dynamic mechanism switching. The framework is trialed on HOI motion forecasting. On two rich datasets and a wide variety of settings, the model consistently delivers superior performances, proving its suitability for the challenge.

We present a physics-based humanoid controller that achieves high-fidelity motion imitation and fault-tolerant behavior in the presence of noisy input (e.g. pose estimates from video or generated from language) and unexpected falls. Our controller scales up to learning ten thousand motion clips without using any external stabilizing forces and learns to naturally recover from fail-state. Given reference motion, our controller can perpetually control simulated avatars without requiring resets. At its core, we propose the progressive multiplicative control policy (PMCP), which dynamically allocates new network capacity to learn harder and harder motion sequences. PMCP allows efficient scaling for learning from large-scale motion databases and adding new tasks, such as fail-state recovery, without catastrophic forgetting. We demonstrate the effectiveness of our controller by using it to imitate noisy poses from video-based pose estimators and language-based motion generators in a live and real-time multi-person avatar use case.

Source-free object detection (SFOD) aims to adapt the source detector to unlabeled target domain data in the absence of source domain data. Most SFOD methods follow the same self-training paradigm using mean-teacher (MT) framework where the student model is guided by only one single teacher model. However, such paradigm can easily fall into a training instability problem that when the teacher model collapses uncontrollably due to the domain shift, the student model also suffers drastic performance degradation. To address this issue, we propose the Periodically Exchange Teacher-Student (PETS) method, a simple yet novel approach that introduces a multiple-teacher framework consisting of a static teacher, a dynamic teacher, and a student model. During the training phase, we periodically exchange the weights between the static teacher and the student model. Then, we update the dynamic teacher using the moving average of the student model that has already been exchanged by the static teacher. In this way, the dynamic teacher can integrate knowledge from past periods, effectively reducing error accumulation and enabling a more stable training process within the MT-based framework. Further, we develop a consensus mechanism to merge the predictions of two teacher models to provide higher-quality pseudo labels for student model. Extensive experiments on multiple SFOD benchmarks show that the proposed method achieves state-of-the-art performance compared with other related methods, demonstrating the effectiveness and superiority of our method on SFOD task.