Vision-based occupancy prediction, also known as 3D Semantic Scene Completion (SSC), presents a significant challenge in computer vision. Previous methods, confined to onboard processing, struggle with simultaneous geometric and semantic estimation, continuity across varying viewpoints, and single-view occlusion. Our paper introduces OccFiner, a novel offboard framework designed to enhance the accuracy of vision-based occupancy predictions. OccFiner operates in two hybrid phases: 1) a multi-to-multi local propagation network that implicitly aligns and processes multiple local frames for correcting onboard model errors and consistently enhancing occupancy accuracy across all distances. 2) the region-centric global propagation, focuses on refining labels using explicit multi-view geometry and integrating sensor bias, especially to increase the accuracy of distant occupied voxels. Extensive experiments demonstrate that OccFiner improves both geometric and semantic accuracy across various types of coarse occupancy, setting a new state-of-the-art performance on the SemanticKITTI dataset. Notably, OccFiner elevates vision-based SSC models to a level even surpassing that of LiDAR-based onboard SSC models.
Relying on paired synthetic data, existing learning-based Computational Aberration Correction (CAC) methods are confronted with the intricate and multifaceted synthetic-to-real domain gap, which leads to suboptimal performance in real-world applications. In this paper, in contrast to improving the simulation pipeline, we deliver a novel insight into real-world CAC from the perspective of Unsupervised Domain Adaptation (UDA). By incorporating readily accessible unpaired real-world data into training, we formalize the Domain Adaptive CAC (DACAC) task, and then introduce a comprehensive Real-world aberrated images (Realab) dataset to benchmark it. The setup task presents a formidable challenge due to the intricacy of understanding the target aberration domain. To this intent, we propose a novel Quntized Domain-Mixing Representation (QDMR) framework as a potent solution to the issue. QDMR adapts the CAC model to the target domain from three key aspects: (1) reconstructing aberrated images of both domains by a VQGAN to learn a Domain-Mixing Codebook (DMC) which characterizes the degradation-aware priors; (2) modulating the deep features in CAC model with DMC to transfer the target domain knowledge; and (3) leveraging the trained VQGAN to generate pseudo target aberrated images from the source ones for convincing target domain supervision. Extensive experiments on both synthetic and real-world benchmarks reveal that the models with QDMR consistently surpass the competitive methods in mitigating the synthetic-to-real gap, which produces visually pleasant real-world CAC results with fewer artifacts. Codes and datasets will be made publicly available.
Understanding human actions from body poses is critical for assistive robots sharing space with humans in order to make informed and safe decisions about the next interaction. However, precise temporal localization and annotation of activity sequences is time-consuming and the resulting labels are often noisy. If not effectively addressed, label noise negatively affects the model's training, resulting in lower recognition quality. Despite its importance, addressing label noise for skeleton-based action recognition has been overlooked so far. In this study, we bridge this gap by implementing a framework that augments well-established skeleton-based human action recognition methods with label-denoising strategies from various research areas to serve as the initial benchmark. Observations reveal that these baselines yield only marginal performance when dealing with sparse skeleton data. Consequently, we introduce a novel methodology, NoiseEraSAR, which integrates global sample selection, co-teaching, and Cross-Modal Mixture-of-Experts (CM-MOE) strategies, aimed at mitigating the adverse impacts of label noise. Our proposed approach demonstrates better performance on the established benchmark, setting new state-of-the-art standards. The source code for this study will be made accessible at https://github.com/xuyizdby/NoiseEraSAR.
This paper introduces the task of Auditory Referring Multi-Object Tracking (AR-MOT), which dynamically tracks specific objects in a video sequence based on audio expressions and appears as a challenging problem in autonomous driving. Due to the lack of semantic modeling capacity in audio and video, existing works have mainly focused on text-based multi-object tracking, which often comes at the cost of tracking quality, interaction efficiency, and even the safety of assistance systems, limiting the application of such methods in autonomous driving. In this paper, we delve into the problem of AR-MOT from the perspective of audio-video fusion and audio-video tracking. We put forward EchoTrack, an end-to-end AR-MOT framework with dual-stream vision transformers. The dual streams are intertwined with our Bidirectional Frequency-domain Cross-attention Fusion Module (Bi-FCFM), which bidirectionally fuses audio and video features from both frequency- and spatiotemporal domains. Moreover, we propose the Audio-visual Contrastive Tracking Learning (ACTL) regime to extract homogeneous semantic features between expressions and visual objects by learning homogeneous features between different audio and video objects effectively. Aside from the architectural design, we establish the first set of large-scale AR-MOT benchmarks, including Echo-KITTI, Echo-KITTI+, and Echo-BDD. Extensive experiments on the established benchmarks demonstrate the effectiveness of the proposed EchoTrack model and its components. The source code and datasets will be made publicly available at https://github.com/lab206/EchoTrack.
Integrating information from multiple modalities enhances the robustness of scene perception systems in autonomous vehicles, providing a more comprehensive and reliable sensory framework. However, the modality incompleteness in multi-modal segmentation remains under-explored. In this work, we establish a task called Modality-Incomplete Scene Segmentation (MISS), which encompasses both system-level modality absence and sensor-level modality errors. To avoid the predominant modality reliance in multi-modal fusion, we introduce a Missing-aware Modal Switch (MMS) strategy to proactively manage missing modalities during training. Utilizing bit-level batch-wise sampling enhances the model's performance in both complete and incomplete testing scenarios. Furthermore, we introduce the Fourier Prompt Tuning (FPT) method to incorporate representative spectral information into a limited number of learnable prompts that maintain robustness against all MISS scenarios. Akin to fine-tuning effects but with fewer tunable parameters (1.1%). Extensive experiments prove the efficacy of our proposed approach, showcasing an improvement of 5.84% mIoU over the prior state-of-the-art parameter-efficient methods in modality missing. The source code will be publicly available at https://github.com/RuipingL/MISS.
Leveraging the rich information extracted from light field (LF) cameras is instrumental for dense prediction tasks. However, adapting light field data to enhance Salient Object Detection (SOD) still follows the traditional RGB methods and remains under-explored in the community. Previous approaches predominantly employ a custom two-stream design to discover the implicit angular feature within light field cameras, leading to significant information isolation between different LF representations. In this study, we propose an efficient paradigm (LF Tracy) to address this limitation. We eschew the conventional specialized fusion and decoder architecture for a dual-stream backbone in favor of a unified, single-pipeline approach. This comprises firstly a simple yet effective data augmentation strategy called MixLD to bridge the connection of spatial, depth, and implicit angular information under different LF representations. A highly efficient information aggregation (IA) module is then introduced to boost asymmetric feature-wise information fusion. Owing to this innovative approach, our model surpasses the existing state-of-the-art methods, particularly demonstrating a 23% improvement over previous results on the latest large-scale PKU dataset. By utilizing only 28.9M parameters, the model achieves a 10% increase in accuracy with 3M additional parameters compared to its backbone using RGB images and an 86% rise to its backbone using LF images. The source code will be made publicly available at https://github.com/FeiBryantkit/LF-Tracy.
3D human pose estimation captures the human joint points in three-dimensional space while keeping the depth information and physical structure. That is essential for applications that require precise pose information, such as human-computer interaction, scene understanding, and rehabilitation training. Due to the challenges in data collection, mainstream datasets of 3D human pose estimation are primarily composed of multi-view video data collected in laboratory environments, which contains rich spatial-temporal correlation information besides the image frame content. Given the remarkable self-attention mechanism of transformers, capable of capturing the spatial-temporal correlation from multi-view video datasets, we propose a multi-stage framework for 3D sequence-to-sequence (seq2seq) human pose detection. Firstly, the spatial module represents the human pose feature by intra-image content, while the frame-image relation module extracts temporal relationships and 3D spatial positional relationship features between the multi-perspective images. Secondly, the self-attention mechanism is adopted to eliminate the interference from non-human body parts and reduce computing resources. Our method is evaluated on Human3.6M, a popular 3D human pose detection dataset. Experimental results demonstrate that our approach achieves state-of-the-art performance on this dataset.
In real-world scenarios, human actions often fall outside the distribution of training data, making it crucial for models to recognize known actions and reject unknown ones. However, using pure skeleton data in such open-set conditions poses challenges due to the lack of visual background cues and the distinct sparse structure of body pose sequences. In this paper, we tackle the unexplored Open-Set Skeleton-based Action Recognition (OS-SAR) task and formalize the benchmark on three skeleton-based datasets. We assess the performance of seven established open-set approaches on our task and identify their limits and critical generalization issues when dealing with skeleton information. To address these challenges, we propose a distance-based cross-modality ensemble method that leverages the cross-modal alignment of skeleton joints, bones, and velocities to achieve superior open-set recognition performance. We refer to the key idea as CrossMax - an approach that utilizes a novel cross-modality mean max discrepancy suppression mechanism to align latent spaces during training and a cross-modality distance-based logits refinement method during testing. CrossMax outperforms existing approaches and consistently yields state-of-the-art results across all datasets and backbones. The benchmark, code, and models will be released at https://github.com/KPeng9510/OS-SAR.
Human pose estimation is a critical component in autonomous driving and parking, enhancing safety by predicting human actions. Traditional frame-based cameras and videos are commonly applied, yet, they become less reliable in scenarios under high dynamic range or heavy motion blur. In contrast, event cameras offer a robust solution for navigating these challenging contexts. Predominant methodologies incorporate event cameras into learning frameworks by accumulating events into event frames. However, such methods tend to marginalize the intrinsic asynchronous and high temporal resolution characteristics of events. This disregard leads to a loss in essential temporal dimension data, crucial for safety-critical tasks associated with dynamic human activities. To address this issue and to unlock the 3D potential of event information, we introduce two 3D event representations: the Rasterized Event Point Cloud (RasEPC) and the Decoupled Event Voxel (DEV). The RasEPC collates events within concise temporal slices at identical positions, preserving 3D attributes with statistical cues and markedly mitigating memory and computational demands. Meanwhile, the DEV representation discretizes events into voxels and projects them across three orthogonal planes, utilizing decoupled event attention to retrieve 3D cues from the 2D planes. Furthermore, we develop and release EV-3DPW, a synthetic event-based dataset crafted to facilitate training and quantitative analysis in outdoor scenes. On the public real-world DHP19 dataset, our event point cloud technique excels in real-time mobile predictions, while the decoupled event voxel method achieves the highest accuracy. Experiments reveal our proposed 3D representation methods' superior generalization capacities against traditional RGB images and event frame techniques. Our code and dataset are available at https://github.com/MasterHow/EventPointPose.
Human pose estimation is a critical component in autonomous driving and parking, enhancing safety by predicting human actions. Traditional frame-based cameras and videos are commonly applied, yet, they become less reliable in scenarios under high dynamic range or heavy motion blur. In contrast, event cameras offer a robust solution for navigating these challenging contexts. Predominant methodologies incorporate event cameras into learning frameworks by accumulating events into event frames. However, such methods tend to marginalize the intrinsic asynchronous and high temporal resolution characteristics of events. This disregard leads to a loss in essential temporal dimension data, crucial for safety-critical tasks associated with dynamic human activities. To address this issue and to unlock the 3D potential of event information, we introduce two 3D event representations: the Rasterized Event Point Cloud (RasEPC) and the Decoupled Event Voxel (DEV). The RasEPC collates events within concise temporal slices at identical positions, preserving 3D attributes with statistical cues and markedly mitigating memory and computational demands. Meanwhile, the DEV representation discretizes events into voxels and projects them across three orthogonal planes, utilizing decoupled event attention to retrieve 3D cues from the 2D planes. Furthermore, we develop and release EV-3DPW, a synthetic event-based dataset crafted to facilitate training and quantitative analysis in outdoor scenes. On the public real-world DHP19 dataset, our event point cloud technique excels in real-time mobile predictions, while the decoupled event voxel method achieves the highest accuracy. Experiments reveal our proposed 3D representation methods' superior generalization capacities against traditional RGB images and event frame techniques. Our code and dataset are available at https://github.com/MasterHow/EventPointPose.