WACV 2024 论文列表

This paper focuses on building object-centric representations for long-term action anticipation in videos. Our key motivation is that objects provide important cues to recognize and predict human-object interactions, especially when the predictions are longer term, as an observed "background" object could be used by the human actor in the future. We observe that existing object-based video recognition frameworks either assume the existence of in-domain supervised object detectors or follow a fully weakly-supervised pipeline to infer object locations from action labels. We propose to build object-centric video representations by leveraging visual-language pretrained models. This is achieved by "object prompts", an approach to extract task-specific object-centric representations from general-purpose pretrained models without finetuning. To recognize and predict human-object interactions, we use a Transformer-based neural architecture which allows the "retrieval" of relevant objects for action anticipation at various time scales. We conduct extensive evaluations on the Ego4D, 50Salads, and EGTEA Gaze+ benchmarks. Both quantitative and qualitative results confirm the effectiveness of our proposed method.

Object re-identification (ReID) from images plays a critical role in application domains of image retrieval (surveillance, retail analytics, etc.) and multi-object tracking (autonomous driving, robotics, etc.). However, systems that additionally or exclusively perceive the world from depth sensors are becoming more commonplace without any corresponding methods for object ReID. In this work, we fill the gap by providing the first large-scale study of object ReID from point clouds and establishing its performance relative to image ReID. To enable such a study, we create two large-scale ReID datasets with paired image and LiDAR observations and propose a lightweight matching head that can be concatenated to any set or sequence processing backbone (e.g., PointNet or ViT), creating a family of comparable object ReID networks for both modalities. Run in Siamese style, our proposed point cloud ReID networks can make thousands of pairwise comparisons in real-time (10 hz). Our findings demonstrate that their performance increases with higher sensor resolution and approaches that of image ReID when observations are sufficiently dense. Our strongest network trained at the largest scale achieves ReID accuracy exceeding 90% for rigid objects and 85% for deformable objects (without any explicit skeleton normalization). To our knowledge, we are the first to study object re-identification from real point cloud observations.

Interactive Image Segmentation is a process of separating a user selected object from the background. This task requires building an effective class-agnostic segmentation model that performs well even on unseen categories. To achieve good accuracy with limited training dataset, it is important that the model has robust prior understanding of features of similar class objects. The model should also have good distinguishing capabilities of foreground objects with the background. In this paper, we propose Object Aware Click Embeddings (OACE) that represents user click aware foreground object features. OACE is obtained based on a prior network trained using the Contrastive Learning paradigm. The single-click object selection accuracy of our base interactive segmentation network is vastly improved with the OACE input. Additionally, we propose a Multi-Stage fusion approach to better utilize user click information. With the proposed method, we outperform existing state-of-the-art approaches by 21% on publicly available test-sets for click-based Interactive Image Segmentation.

Semantic segmentation is a fundamental task in visual scene understanding. We focus on the supervised setting, where ground-truth semantic annotations are available. Based on knowledge about the high regularity of real-world scenes, we propose a method for improving class predictions by learning to selectively exploit information from neighboring pixels. In particular, our method is based on the prior that for each pixel, there is a seed pixel in its close neighborhood sharing the same prediction with the former. Motivated by this prior, we design a novel two-head network, named Offset Vector Network (OVeNet), which generates both standard semantic predictions and a dense 2D offset vector field indicating the offset from each pixel to the respective seed pixel, which is used to compute an alternative, seed-based semantic prediction. The two predictions are adaptively fused at each pixel using a learnt dense confidence map for the predicted offset vector field. We supervise offset vectors indirectly via optimizing the seed-based prediction and via a novel loss on the confidence map. Compared to the baseline state-of-the-art architectures HRNet and HRNet+OCR on which OVeNet is built, the latter achieves significant performance gains on three prominent benchmarks for semantic segmentation, namely Cityscapes, ACDC and ADE20K. Code is available at https://github.com/stamatisalex/OVeNet.

Temporal action segmentation is typically achieved by discovering the dramatic variances in global visual descriptors. In this paper, we explore the merits of local features by proposing the unsupervised framework of Object-centric Temporal Action Segmentation (OTAS). Broadly speaking, OTAS consists of self-supervised global and local feature extraction modules as well as a boundary selection module that fuses the features and detects salient boundaries for action segmentation. As a second contribution, we discuss the pros and cons of existing frame-level and boundary-level evaluation metrics. Through extensive experiments, we find OTAS is superior to the previous state-of-the-art method by 41% on average in terms of our recommended F1 score. Surprisingly, OTAS even outperforms the ground-truth human annotations in the user study. Moreover, OTAS is efficient enough to allow real-time inference

Out-of-Distribution (OOD) detection is a crucial problem for the safe deployment of machine learning models identifying samples that fall outside of the training distribution, i.e. in-distribution data (ID). Most OOD works focus on the classification models trained with Cross Entropy (CE) and attempt to fix its inherent issues. In this work we leverage powerful representation learned with Supervised Contrastive (SupCon) training and propose a holistic approach to learn a classifier robust to OOD data. We extend SupCon loss with two additional contrast terms. The first term pushes auxiliary OOD representations away from ID representations without imposing any constraints on similarities among auxiliary data. The second term pushes OOD features far from the existing class prototypes, while pushing ID representations closer to their corresponding class prototype. When auxiliary OOD data is not available, we propose feature mixing techniques to efficiently generate pseudo-OOD features. Our solution is simple and efficient and acts as a natural extension of the closed-set supervised contrastive representation learning. We compare against different OOD detection methods on the common benchmarks and show state-of-the-art results.

Video anomaly detection in real-world scenarios is challenging due to the complex temporal blending of long and short-length anomalies with normal ones. Further, it is more difficult to detect those due to : (i) Distinctive features characterizing the short and long anomalies with sharp and progressive temporal cues respectively; (ii) Lack of precise temporal information (i.e. weak-supervision) limits the temporal dynamics modeling of anomalies from normal events. In this paper, we propose a novel 'temporal transformer' framework for weakly-supervised anomaly detection: OE-CTST. The proposed framework has two major components: (i) Outlier Embedder (OE) and (ii) Cross Temporal Scale Transformer (CTST). First, OE generates anomaly-aware temporal position encoding to allow the transformer to effectively model the temporal dynamics among the anomalies and normal events. Second, CTST encodes the cross-correlation between multi-temporal scale features to benefit short and long length anomalies by modeling the global temporal relations. The proposed OE-CTST is validated on three publicly available datasets i.e. UCF-Crime, XD-Violence, and IITB-Corridor, outperforming recently reported state-of-the-art approaches.

Human vision demonstrates higher robustness than current AI algorithms under out-of-distribution scenarios. It has been conjectured such robustness benefits from performing analysis-by-synthesis. Our paper formulates triple vision tasks in a consistent manner using approximate analysis-by-synthesis by render-and-compare algorithms on neural features. In this work, we introduce Neural Textured Deformable Meshes, which involve the object model with deformable geometry that allows optimization on both camera parameters and object geometries. The deformable mesh is parameterized as a neural field, and covered by whole-surface neural texture maps, which are trained to have spatial discriminability. During inference, we extract the feature map of the test image and subsequently optimize the 3D pose and shape parameters of our model using differentiable rendering to best reconstruct the target feature map. We show that our analysis-by-synthesis is much more robust than conventional neural networks when evaluated on real-world images and even in challenging out-of-distribution scenarios, such as occlusion and domain shift. Our algorithms are competitive with standard algorithms when tested on conventional performance measures.

Arbitrary neural style transfer is an advanced AI technique that can effectively synthesize pictures with an artistic style similar to a given source picture. However, if such an AI technique is leveraged by unauthorized individuals, it can significantly infringe upon the copyright of the source picture's owner. In this paper, we study how to protect the artistic style of source images against unauthorized style transfer by adding imperceptible perturbations to the original source pictures. In particular, our goal is to disable the neural style transfer models from producing high-quality pictures with a similar style to the source pictures with slight manipulating the source images. We introduce Neural Style Protection (NSP), which provides protection for source images against various neural style transfer models. Through extensive experiments, we demonstrate the effectiveness and generalizability of the proposed style protection algorithm across numerous style transfer models using varied metrics.

We study neural image compression based on the Sparse Visual Representation (SVR), where images are embedded into a discrete latent space spanned by learned visual codebooks. By sharing codebooks with the decoder, the encoder transfers integer codeword indices that are efficient and cross-platform robust, and the decoder retrieves the embedded latent feature using the indices for reconstruction. Previous SVR-based compression lacks effective mechanism for rate-distortion tradeoffs, where one can only pursue either high reconstruction quality or low transmission bitrate. We propose a Masked Adaptive Codebook learning (M-AdaCode) method that applies masks to the latent feature subspace to balance bitrate and reconstruction quality. A set of semantic-class-dependent basis codebooks are learned, which are weighted combined to generate a rich latent feature for high-quality reconstruction. The combining weights are adaptively derived from each input image, providing fidelity information with additional transmission costs. By masking out unimportant weights in the encoder and recovering them in the decoder, we can trade off reconstruction quality for transmission bits, and the masking rate controls the balance between bitrate and distortion. Experiments over the standard JPEG-AI dataset demonstrate the effectiveness of our M-AdaCode approach.

In this study, we present evidence suggesting that depthwise convolutional kernels are effectively replicating the structural intricacies of the biological receptive fields observed in the mammalian retina. We provide analytics of trained kernels from various state-of-the-art models substantiating this evidence. Inspired by this intriguing discovery, we propose an initialization scheme that draws inspiration from the biological receptive fields. Experimental analysis of the ImageNet dataset with multiple CNN architectures featuring depthwise convolutions reveals a marked enhancement in the accuracy of the learned model when initialized with biologically derived weights. This underlies the potential for biologically inspired computational models to further our understanding of vision processing systems and to improve the efficacy of convolutional networks.

Diffusion models are the current state-of-the-art in image generation, synthesizing high-quality images by breaking down the generation process into many fine-grained denoising steps. Despite their good performance, diffusion models are computationally expensive, requiring many neural function evaluations (NFEs). In this work, we propose an anytime diffusion-based method that can generate viable images when stopped at arbitrary times before completion. Using existing pretrained diffusion models, we show that the generation scheme can be recomposed as two nested diffusion processes, enabling fast iterative refinement of a generated image. In experiments on ImageNet and Stable Diffusion-based text-to-image generation, we show, both qualitatively and quantitatively, that our method's intermediate generation quality greatly exceeds that of the original diffusion model, while the final generation result remains comparable. We illustrate the applicability of Nested Diffusion in several settings, including for solving inverse problems, and for rapid text-based content creation by allowing user intervention throughout the sampling process.

We present NeRFEditor, an efficient learning framework for 3D scene editing, which takes a video as input and outputs a high quality, identity-preserving stylized 3D scene. Our goal is to bridge the gap between 2D and 3D editing, catering to a wide array of creative modifications such as reference-guided alterations, text-based prompts, and user interactions. We achieve this by encouraging a pre-trained StyleGAN model and a NeRF model to learn mutually consistent renderings. Specifically, we use NeRF to generate numerous (image, camera pose)-pairs to train an adjustor module, which adapts the StyleGAN latent code for generating high fidelity stylized images from any given viewing angle. To extrapolate edits to novel views, i.e., those not seen by StyleGAN pre-training, while maintaining 360 degree consistency, we propose a second self-supervised module that maps these views into the hidden space of StyleGAN. Together these two modules produce sufficient guidance for NeRF to learn consistent stylization effects across the full range of views. Experiments show that NeRFEditor outperforms prior work on benchmark and real-world scenes with better editability, fidelity, and identity preservation.

Common illumination sources like sunlight or artificial light may introduce hidden vulnerabilities to AI systems. Our paper delves into these potential threats, offering a novel approach to simulate varying light conditions, including sunlight, headlights, and flashlight illuminations. Moreover, unlike typical physical adversarial attacks requiring conspicuous alterations, our method utilizes a model-agnostic black-box attack integrated with the Zeroth Order Optimization (ZOO) algorithm to identify deceptive patterns in a physically-applicable space. Consequently, attackers can recreate these simulated conditions, deceiving machine learning models with seemingly natural light. Empirical results demonstrate the efficacy of our method, misleading models trained on the GTSRB and LISA datasets under natural-like physical environments with an attack success rate exceeding 70% across all digital datasets, and remaining effective against all evaluated real-world traffic signs. Importantly, after adversarial training using samples generated from our approach, models showcase enhanced robustness, underscoring the dual value of our work in both identifying and mitigating potential threats.

Achieving robust and real-time 3D perception is fundamental for autonomous vehicles. While most existing 3D perception methods prioritize detection accuracy, they often overlook critical aspects such as computational efficiency, onboard chip deployment friendliness, resilience to sensor mounting deviations, and adaptability to various vehicle types. To address these challenges, we present NVAutoNet: a specialized Bird's-Eye-View (BEV) perception network tailored explicitly for automated vehicles. NVAutoNet takes synchronized camera images as input and predicts 3D signals like obstacles, freespaces, and parking spaces. The core of NVAutoNet's architecture (image and BEV backbones) relies on efficient convolutional networks, optimized for high performance using TensorRT. Our image-to-BEV transformation employs simple linear layers and BEV look-up tables, ensuring rapid inference speed. Trained on an extensive proprietary dataset, NVAutoNet consistently achieves elevated perception accuracy, operating remarkably at 53 frames per second on the NVIDIA DRIVE Orin SoC. Notably, NVAutoNet demonstrates resilience to sensor mounting deviations arising from diverse car models. Moreover, NVAutoNet excels in adapting to varied vehicle types, facilitated by inexpensive model fine-tuning procedures that expedite compatibility adjustments.

With the increasing reliance on small Unmanned Aerial Systems (sUAS) for Emergency Response Scenarios, such as Search and Rescue, the integration of computer vision capabilities has become a key factor in mission success. Nevertheless, computer vision performance for detecting humans severely degrades when shifting from ground to aerial views. Several aerial datasets have been created to mitigate this problem, however, none of them has specifically addressed the issue of occlusion, a critical component in Emergency Response Scenarios. Natural Occluded Multi-scale Aerial Dataset (NOMAD) presents a benchmark for human detection under occluded aerial views, with five different aerial distances and rich imagery variance. NOMAD is composed of 100 different Actors, all performing sequences of walking, laying and hiding. It includes 42,825 frames, extracted from 5.4k resolution videos, and manually annotated with a bounding box and a label describing 10 different visibility levels, categorized according to the percentage of the human body visible inside the bounding box. This allows computer vision models to be evaluated on their detection performance across different ranges of occlusion. NOMAD is designed to improve the effectiveness of aerial search and rescue and to enhance collaboration between sUAS and humans, by providing a new benchmark dataset for human detection under occluded aerial views.

Eye contact is a crucial non-verbal interaction modality and plays an important role in our everyday social life. While humans are very sensitive to eye contact, the capabilities of machines to capture a person's gaze are still mediocre. We tackle this challenge and present NITEC, a hand-annotated eye contact dataset for ego-vision interaction. NITEC exceeds existing datasets for ego-vision eye contact in size and variety of demographics, social contexts, and lighting conditions, making it a valuable resource for advancing ego-vision-based eye contact research. Our extensive evaluations on NITEC demonstrate strong cross-dataset performance, emphasizing its effectiveness and adaptability in various scenarios, that allows seamless utilization to the fields of computer vision, human-computer interaction, and social robotics. We make our NITEC dataset publicly available to foster reproducibility and further exploration in the field of ego-vision interaction.

We propose a novel and effective purification-based adversarial defense method against pre-processor blind white- and black-box attacks, without requiring any adversarial training or retraining of the classification model. Based on the observation of the adversarial noise, we propose a simple iterative Gaussian Smoothing (GS) that smoothes out adversarial noise and achieves substantially high robust accuracy. To further improve the method, we propose Neural Contextual Iterative Smoothing (NCIS), which trains a blind-spot network (BSN) in a self-supervised manner to reconstruct the discriminative features of the smoothed original image. From the extensive experiments on the large-scale ImageNet, we show that our method achieves both competitive standard accuracy and state-of-the-art robust accuracy against most strong purifier-blind white- and black-box attacks. Also, we propose a new evaluation benchmark based on commercial image classification APIs, including AWS, Azure, Clarifai, and Google, and demonstrate that users can use our method to increase the adversarial robustness of APIs.

Prompt Tuning, conditioning on task-specific learned prompt vectors, has emerged as a data-efficient and parameter-efficient method for adapting large pretrained vision-language models to multiple downstream tasks. However, existing approaches usually consider learning prompt vectors for each task independently from scratch, thereby failing to exploit the rich shareable knowledge across different vision-language tasks. In this paper, we propose multitask vision-language prompt tuning (MVLPT), which incorporates cross-task knowledge into prompt tuning for vision-language models. Specifically, (i) we demonstrate the effectiveness of learning a single transferable prompt from multiple source tasks to initialize the prompt for each target task; (ii) we show many target tasks can benefit each other from sharing prompt vectors and thus can be jointly learned via multitask prompt tuning. We benchmark the proposed MVLPT using three representative prompt tuning methods, namely text prompt tuning, visual prompt tuning, and the unified vision-language prompt tuning. Results in 20 vision tasks demonstrate that the proposed approach outperforms all single-task baseline prompt tuning methods, setting the new state-of-the-art on the few-shot ELEVATER benchmarks and cross-task generalization benchmarks. To understand where the cross-task knowledge is most effective, we also conduct a large-scale study on task transferability with 20 vision tasks in 400 combinations for each prompt tuning method. It shows that the most performant MVLPT for each prompt tuning method prefers different task combinations and many tasks can benefit each other, depending on their visual similarity and label similarity.