Chen Feng

Yiming Li, Zhiding Yu, Christopher Choy et al.

Humans can easily imagine the complete 3D geometry of occluded objects and scenes. This appealing ability is vital for recognition and understanding. To enable such capability in AI systems, we propose VoxFormer, a Transformer-based semantic scene completion framework that can output complete 3D volumetric semantics from only 2D images. Our framework adopts a two-stage design where we start from a sparse set of visible and occupied voxel queries from depth estimation, followed by a densification stage that generates dense 3D voxels from the sparse ones. A key idea of this design is that the visual features on 2D images correspond only to the visible scene structures rather than the occluded or empty spaces. Therefore, starting with the featurization and prediction of the visible structures is more reliable. Once we obtain the set of sparse queries, we apply a masked autoencoder design to propagate the information to all the voxels by self-attention. Experiments on SemanticKITTI show that VoxFormer outperforms the state of the art with a relative improvement of 20.0% in geometry and 18.1% in semantics and reduces GPU memory during training to less than 16GB. Our code is available on https://github.com/NVlabs/VoxFormer.

Yifan Lu, Quanhao Li, Baoan Liu et al.

Collaborative 3D object detection exploits information exchange among multiple agents to enhance accuracy of object detection in presence of sensor impairments such as occlusion. However, in practice, pose estimation errors due to imperfect localization would cause spatial message misalignment and significantly reduce the performance of collaboration. To alleviate adverse impacts of pose errors, we propose CoAlign, a novel hybrid collaboration framework that is robust to unknown pose errors. The proposed solution relies on a novel agent-object pose graph modeling to enhance pose consistency among collaborating agents. Furthermore, we adopt a multi-scale data fusion strategy to aggregate intermediate features at multiple spatial resolutions. Comparing with previous works, which require ground-truth pose for training supervision, our proposed CoAlign is more practical since it doesn't require any ground-truth pose supervision in the training and makes no specific assumptions on pose errors. Extensive evaluation of the proposed method is carried out on multiple datasets, certifying that CoAlign significantly reduce relative localization error and achieving the state of art detection performance when pose errors exist. Code are made available for the use of the research community at https://github.com/yifanlu0227/CoAlign.

Chen Feng, Ioannis Patras

Self-supervised learning has recently achieved great success in representation learning without human annotations. The dominant method -- that is contrastive learning, is generally based on instance discrimination tasks, i.e., individual samples are treated as independent categories. However, presuming all the samples are different contradicts the natural grouping of similar samples in common visual datasets, e.g., multiple views of the same dog. To bridge the gap, this paper proposes an adaptive method that introduces soft inter-sample relations, namely Adaptive Soft Contrastive Learning (ASCL). More specifically, ASCL transforms the original instance discrimination task into a multi-instance soft discrimination task, and adaptively introduces inter-sample relations. As an effective and concise plug-in module for existing self-supervised learning frameworks, ASCL achieves the best performance on several benchmarks in terms of both performance and efficiency. Code is available at https://github.com/MrChenFeng/ASCL_ICPR2022.

Chen Feng, Ioannis Patras

Deep learning has achieved great success in recent years with the aid of advanced neural network structures and large-scale human-annotated datasets. However, it is often costly and difficult to accurately and efficiently annotate large-scale datasets, especially for some specialized domains where fine-grained labels are required. In this setting, coarse labels are much easier to acquire as they do not require expert knowledge. In this work, we propose a contrastive learning method, called $\textbf{Mask}$ed $\textbf{Con}$trastive learning~($\textbf{MaskCon}$) to address the under-explored problem setting, where we learn with a coarse-labelled dataset in order to address a finer labelling problem. More specifically, within the contrastive learning framework, for each sample our method generates soft-labels with the aid of coarse labels against other samples and another augmented view of the sample in question. By contrast to self-supervised contrastive learning where only the sample's augmentations are considered hard positives, and in supervised contrastive learning where only samples with the same coarse labels are considered hard positives, we propose soft labels based on sample distances, that are masked by the coarse labels. This allows us to utilize both inter-sample relations and coarse labels. We demonstrate that our method can obtain as special cases many existing state-of-the-art works and that it provides tighter bounds on the generalization error. Experimentally, our method achieves significant improvement over the current state-of-the-art in various datasets, including CIFAR10, CIFAR100, ImageNet-1K, Standford Online Products and Stanford Cars196 datasets. Code and annotations are available at https://github.com/MrChenFeng/MaskCon_CVPR2023.

Min Shi, Hao Lu, Chen Feng et al.

Class-agnostic counting (CAC) aims to count all instances in a query image given few exemplars. A standard pipeline is to extract visual features from exemplars and match them with query images to infer object counts. Two essential components in this pipeline are feature representation and similarity metric. Existing methods either adopt a pretrained network to represent features or learn a new one, while applying a naive similarity metric with fixed inner product. We find this paradigm leads to noisy similarity matching and hence harms counting performance. In this work, we propose a similarity-aware CAC framework that jointly learns representation and similarity metric. We first instantiate our framework with a naive baseline called Bilinear Matching Network (BMNet), whose key component is a learnable bilinear similarity metric. To further embody the core of our framework, we extend BMNet to BMNet+ that models similarity from three aspects: 1) representing the instances via their self-similarity to enhance feature robustness against intra-class variations; 2) comparing the similarity dynamically to focus on the key patterns of each exemplar; 3) learning from a supervision signal to impose explicit constraints on matching results. Extensive experiments on a recent CAC dataset FSC147 show that our models significantly outperform state-of-the-art CAC approaches. In addition, we also validate the cross-dataset generality of BMNet and BMNet+ on a car counting dataset CARPK. Code is at tiny.one/BMNet

Chen Feng, Hangning Zhou, Huadong Lin et al.

Predicting the future behavior of agents is a fundamental task in autonomous vehicle domains. Accurate prediction relies on comprehending the surrounding map, which significantly regularizes agent behaviors. However, existing methods have limitations in exploiting the map and exhibit a strong dependence on historical trajectories, which yield unsatisfactory prediction performance and robustness. Additionally, their heavy network architectures impede real-time applications. To tackle these problems, we propose Map-Agent Coupled Transformer (MacFormer) for real-time and robust trajectory prediction. Our framework explicitly incorporates map constraints into the network via two carefully designed modules named coupled map and reference extractor. A novel multi-task optimization strategy (MTOS) is presented to enhance learning of topology and rule constraints. We also devise bilateral query scheme in context fusion for a more efficient and lightweight network. We evaluated our approach on Argoverse 1, Argoverse 2, and nuScenes real-world benchmarks, where it all achieved state-of-the-art performance with the lowest inference latency and smallest model size. Experiments also demonstrate that our framework is resilient to imperfect tracklet inputs. Furthermore, we show that by combining with our proposed strategies, classical models outperform their baselines, further validating the versatility of our framework.

Howard Huang, Bharath Surianarayanan, Keifer Lee et al.

Precise 3D representations of industrial environments enable tasks such as robot localization and digital twin generation. We propose SAVMap, a method for generating a semantic wireframe map of warehouse shelf and light structures using only a panoramic video camera as the sensor input. Sequences of rectified images with shelf and ceiling-facing views are extracted from a panoramic video captured along the warehouse aisles. Using a semantic segmentation network front end, a set of sparse, semantic structure feature points (e.g., corners of shelf structures, centers of lights) are extracted from each image and tracked across the sequences. By accounting for real-world geometric relationships among the points such as Manhattan grids, a constrained structure-from-motion algorithm yields the 3D points that form a wireframe map. We demonstrate the scalability and accuracy of our proposal in a warehouse with 46 shelving rows, each with faces spanning 55\,m by 7\,m. From an hour of panoramic video content, we create wireframe maps for over 5000 shelf elements across the rows, achieving an aggregate mean absolute error of 4.8\,cm with respect to ground-truth.

Yiming Li, Qi Fang, Jiamu Bai et al.

Multiple robots could perceive a scene (e.g., detect objects) collaboratively better than individuals, although easily suffer from adversarial attacks when using deep learning. This could be addressed by the adversarial defense, but its training requires the often-unknown attacking mechanism. Differently, we propose ROBOSAC, a novel sampling-based defense strategy generalizable to unseen attackers. Our key idea is that collaborative perception should lead to consensus rather than dissensus in results compared to individual perception. This leads to our hypothesize-and-verify framework: perception results with and without collaboration from a random subset of teammates are compared until reaching a consensus. In such a framework, more teammates in the sampled subset often entail better perception performance but require longer sampling time to reject potential attackers. Thus, we derive how many sampling trials are needed to ensure the desired size of an attacker-free subset, or equivalently, the maximum size of such a subset that we can successfully sample within a given number of trials. We validate our method on the task of collaborative 3D object detection in autonomous driving scenarios.

Xinhao Liu, Moonjun Gong, Qi Fang et al.

Scene completion and forecasting are two popular perception problems in research for mobile agents like autonomous vehicles. Existing approaches treat the two problems in isolation, resulting in a separate perception of the two aspects. In this paper, we introduce a novel LiDAR perception task of Occupancy Completion and Forecasting (OCF) in the context of autonomous driving to unify these aspects into a cohesive framework. This task requires new algorithms to address three challenges altogether: (1) sparse-to-dense reconstruction, (2) partial-to-complete hallucination, and (3) 3D-to-4D prediction. To enable supervision and evaluation, we curate a large-scale dataset termed OCFBench from public autonomous driving datasets. We analyze the performance of closely related existing baseline models and our own ones on our dataset. We envision that this research will inspire and call for further investigation in this evolving and crucial area of 4D perception. Our code for data curation and baseline implementation is available at https://github.com/ai4ce/Occ4cast.

Sanbao Su, Yiming Li, Sihong He et al.

Sharing information between connected and autonomous vehicles (CAVs) fundamentally improves the performance of collaborative object detection for self-driving. However, CAVs still have uncertainties on object detection due to practical challenges, which will affect the later modules in self-driving such as planning and control. Hence, uncertainty quantification is crucial for safety-critical systems such as CAVs. Our work is the first to estimate the uncertainty of collaborative object detection. We propose a novel uncertainty quantification method, called Double-M Quantification, which tailors a moving block bootstrap (MBB) algorithm with direct modeling of the multivariant Gaussian distribution of each corner of the bounding box. Our method captures both the epistemic uncertainty and aleatoric uncertainty with one inference pass based on the offline Double-M training process. And it can be used with different collaborative object detectors. Through experiments on the comprehensive collaborative perception dataset, we show that our Double-M method achieves more than 4X improvement on uncertainty score and more than 3% accuracy improvement, compared with the state-of-the-art uncertainty quantification methods. Our code is public on https://coperception.github.io/double-m-quantification.

Sanbao Su, Songyang Han, Yiming Li et al.

Object detection and multiple object tracking (MOT) are essential components of self-driving systems. Accurate detection and uncertainty quantification are both critical for onboard modules, such as perception, prediction, and planning, to improve the safety and robustness of autonomous vehicles. Collaborative object detection (COD) has been proposed to improve detection accuracy and reduce uncertainty by leveraging the viewpoints of multiple agents. However, little attention has been paid to how to leverage the uncertainty quantification from COD to enhance MOT performance. In this paper, as the first attempt to address this challenge, we design an uncertainty propagation framework called MOT-CUP. Our framework first quantifies the uncertainty of COD through direct modeling and conformal prediction, and propagates this uncertainty information into the motion prediction and association steps. MOT-CUP is designed to work with different collaborative object detectors and baseline MOT algorithms. We evaluate MOT-CUP on V2X-Sim, a comprehensive collaborative perception dataset, and demonstrate a 2% improvement in accuracy and a 2.67X reduction in uncertainty compared to the baselines, e.g. SORT and ByteTrack. In scenarios characterized by high occlusion levels, our MOT-CUP demonstrates a noteworthy $4.01\%$ improvement in accuracy. MOT-CUP demonstrates the importance of uncertainty quantification in both COD and MOT, and provides the first attempt to improve the accuracy and reduce the uncertainty in MOT based on COD through uncertainty propagation. Our code is public on https://coperception.github.io/MOT-CUP/.

Yiming Li, Sihang Li, Xinhao Liu et al.

Monocular scene understanding is a foundational component of autonomous systems. Within the spectrum of monocular perception topics, one crucial and useful task for holistic 3D scene understanding is semantic scene completion (SSC), which jointly completes semantic information and geometric details from RGB input. However, progress in SSC, particularly in large-scale street views, is hindered by the scarcity of high-quality datasets. To address this issue, we introduce SSCBench, a comprehensive benchmark that integrates scenes from widely used automotive datasets (e.g., KITTI-360, nuScenes, and Waymo). SSCBench follows an established setup and format in the community, facilitating the easy exploration of SSC methods in various street views. We benchmark models using monocular, trinocular, and point cloud input to assess the performance gap resulting from sensor coverage and modality. Moreover, we have unified semantic labels across diverse datasets to simplify cross-domain generalization testing. We commit to including more datasets and SSC models to drive further advancements in this field.

Anbang Yang, Mahya Beheshti, Todd E Hudson et al.

Vision-based localization approaches now underpin newly emerging navigation pipelines for myriad use cases from robotics to assistive technologies. Compared to sensor-based solutions, vision-based localization does not require pre-installed sensor infrastructure, which is costly, time-consuming, and/or often infeasible at scale. Herein, we propose a novel vision-based localization pipeline for a specific use case: navigation support for end-users with blindness and low vision. Given a query image taken by an end-user on a mobile application, the pipeline leverages a visual place recognition (VPR) algorithm to find similar images in a reference image database of the target space. The geolocations of these similar images are utilized in downstream tasks that employ a weighted-average method to estimate the end-user's location and a perspective-n-point (PnP) algorithm to estimate the end-user's direction. Additionally, this system implements Dijkstra's algorithm to calculate a shortest path based on a navigable map that includes trip origin and destination. The topometric map used for localization and navigation is built using a customized graphical user interface that projects a 3D reconstructed sparse map, built from a sequence of images, to the corresponding a priori 2D floor plan. Sequential images used for map construction can be collected in a pre-mapping step or scavenged through public databases/citizen science. The end-to-end system can be installed on any internet-accessible device with a camera that hosts a custom mobile application. For evaluation purposes, mapping and localization were tested in a complex hospital environment. The evaluation results demonstrate that our system can achieve localization with an average error of less than 1 meter without knowledge of the camera's intrinsic parameters, such as focal length.

Yiming Li, Ziang Cao, Andrew Liang et al.

We are interested in anticipating as early as possible the target location of a person's object manipulation action in a 3D workspace from egocentric vision. It is important in fields like human-robot collaboration, but has not yet received enough attention from vision and learning communities. To stimulate more research on this challenging egocentric vision task, we propose a large multimodality dataset of more than 1 million frames of RGB-D and IMU streams, and provide evaluation metrics based on our high-quality 2D and 3D labels from semi-automatic annotation. Meanwhile, we design baseline methods using recurrent neural networks and conduct various ablation studies to validate their effectiveness. Our results demonstrate that this new task is worthy of further study by researchers in robotics, vision, and learning communities.

Chao Chen, Xinhao Liu, Yiming Li et al.

LiDAR mapping is important yet challenging in self-driving and mobile robotics. To tackle such a global point cloud registration problem, DeepMapping converts the complex map estimation into a self-supervised training of simple deep networks. Despite its broad convergence range on small datasets, DeepMapping still cannot produce satisfactory results on large-scale datasets with thousands of frames. This is due to the lack of loop closures and exact cross-frame point correspondences, and the slow convergence of its global localization network. We propose DeepMapping2 by adding two novel techniques to address these issues: (1) organization of training batch based on map topology from loop closing, and (2) self-supervised local-to-global point consistency loss leveraging pairwise registration. Our experiments and ablation studies on public datasets (KITTI, NCLT, and Nebula) demonstrate the effectiveness of our method.

Chen Feng, Georgios Tzimiropoulos, Ioannis Patras

Learning with Noisy labels (LNL) poses a significant challenge for the Machine Learning community. Some of the most widely used approaches that select as clean samples for which the model itself (the in-training model) has high confidence, e.g., `small loss', can suffer from the so called `self-confirmation' bias. This bias arises because the in-training model, is at least partially trained on the noisy labels. Furthermore, in the classification case, an additional challenge arises because some of the label noise is between classes that are visually very similar (`hard noise'). This paper addresses these challenges by proposing a method (\textit{CLIPCleaner}) that leverages CLIP, a powerful Vision-Language (VL) model for constructing a zero-shot classifier for efficient, offline, clean sample selection. This has the advantage that the sample selection is decoupled from the in-training model and that the sample selection is aware of the semantic and visual similarities between the classes due to the way that CLIP is trained. We provide theoretical justifications and empirical evidence to demonstrate the advantages of CLIP for LNL compared to conventional pre-trained models. Compared to current methods that combine iterative sample selection with various techniques, \textit{CLIPCleaner} offers a simple, single-step approach that achieves competitive or superior performance on benchmark datasets. To the best of our knowledge, this is the first time a VL model has been used for sample selection to address the problem of Learning with Noisy Labels (LNL), highlighting their potential in the domain.

Shaojie Zhuo, Hongyu Chen, Ramchalam Kinattinkara Ramakrishnan et al.

Tiny machine learning (tinyML) has emerged during the past few years aiming to deploy machine learning models to embedded AI processors with highly constrained memory and computation capacity. Low precision quantization is an important model compression technique that can greatly reduce both memory consumption and computation cost of model inference. In this study, we focus on post-training quantization (PTQ) algorithms that quantize a model to low-bit (less than 8-bit) precision with only a small set of calibration data and benchmark them on different tinyML use cases. To achieve a fair comparison, we build a simulated quantization framework to investigate recent PTQ algorithms. Furthermore, we break down those algorithms into essential components and re-assembled a generic PTQ pipeline. With ablation study on different alternatives of components in the pipeline, we reveal key design choices when performing low precision quantization. We hope this work could provide useful data points and shed lights on the future research of low precision quantization.

Chao Chen, Zegang Cheng, Xinhao Liu et al.

Visual place recognition (VPR) using deep networks has achieved state-of-the-art performance. However, most of them require a training set with ground truth sensor poses to obtain positive and negative samples of each observation's spatial neighborhood for supervised learning. When such information is unavailable, temporal neighborhoods from a sequentially collected data stream could be exploited for self-supervised training, although we find its performance suboptimal. Inspired by noisy label learning, we propose a novel self-supervised framework named TF-VPR that uses temporal neighborhoods and learnable feature neighborhoods to discover unknown spatial neighborhoods. Our method follows an iterative training paradigm which alternates between: (1) representation learning with data augmentation, (2) positive set expansion to include the current feature space neighbors, and (3) positive set contraction via geometric verification. We conduct auto-labeling and generalization tests on both simulated and real datasets, with either RGB images or point clouds as inputs. The results show that our method outperforms self-supervised baselines in recall rate, robustness, and heading diversity, a novel metric we propose for VPR. Our code and datasets can be found at https://ai4ce.github.io/TF-VPR/

Yuting Wang, Hangning Zhou, Zhigang Zhang et al.

This technical report presents an effective method for motion prediction in autonomous driving. We develop a Transformer-based method for input encoding and trajectory prediction. Besides, we propose the Temporal Flow Header to enhance the trajectory encoding. In the end, an efficient K-means ensemble method is used. Using our Transformer network and ensemble method, we win the first place of Argoverse 2 Motion Forecasting Challenge with the state-of-the-art brier-minFDE score of 1.90.

Wenyu Han, Congcong Wen, Lazarus Chok et al.

City modeling and generation have attracted an increased interest in various applications, including gaming, urban planning, and autonomous driving. Unlike previous works focused on the generation of single objects or indoor scenes, the huge volumes of spatial data in cities pose a challenge to the generative models. Furthermore, few publicly available 3D real-world city datasets also hinder the development of methods for city generation. In this paper, we first collect over 3,000,000 geo-referenced objects for the city of New York, Zurich, Tokyo, Berlin, Boston and several other large cities. Based on this dataset, we propose AETree, a tree-structured auto-encoder neural network, for city generation. Specifically, we first propose a novel Spatial-Geometric Distance (SGD) metric to measure the similarity between building layouts and then construct a binary tree over the raw geometric data of building based on the SGD metric. Next, we present a tree-structured network whose encoder learns to extract and merge spatial information from bottom-up iteratively. The resulting global representation is reversely decoded for reconstruction or generation. To address the issue of long-dependency as the level of the tree increases, a Long Short-Term Memory (LSTM) Cell is employed as a basic network element of the proposed AETree. Moreover, we introduce a novel metric, Overlapping Area Ratio (OAR), to quantitatively evaluate the generation results. Experiments on the collected dataset demonstrate the effectiveness of the proposed model on 2D and 3D city generation. Furthermore, the latent features learned by AETree can serve downstream urban planning applications.

Ruoyu Wang, Yanfei Xue, Bharath Surianarayanan et al.

We propose Concavity-induced Distance (CID) as a novel way to measure the dissimilarity between a pair of points in an unoriented point cloud. CID indicates the likelihood of two points or two sets of points belonging to different convex parts of an underlying shape represented as a point cloud. After analyzing its properties, we demonstrate how CID can benefit point cloud analysis without the need for meshing or normal estimation, which is beneficial for robotics applications when dealing with raw point cloud observations. By randomly selecting very few points for manual labeling, a CID-based point cloud instance segmentation via label propagation achieves comparable average precision as recent supervised deep learning approaches, on S3DIS and ScanNet datasets. Moreover, CID can be used to group points into approximately convex parts whose convex hulls can be used as compact scene representations in robotics, and it outperforms the baseline method in terms of grouping quality. Our project website is available at: https://ai4ce.github.io/CID/

Zeyu Jiang, Sihang Li, Siqi Tan et al.

Most existing 3D assembly methods treat the problem as pure pose estimation, rearranging observed parts via rigid transformations. In contrast, human assembly naturally couples structural reasoning with holistic shape inference. Inspired by this intuition, we reformulate 3D assembly as a joint problem of assembly and generation. We show that these two processes are mutually reinforcing: assembly provides part-level structural priors for generation, while generation injects holistic shape context that resolves ambiguities in assembly. Unlike prior methods that cannot synthesize missing geometry, we propose CRAG, which simultaneously generates plausible complete shapes and predicts poses for input parts. Extensive experiments demonstrate state-of-the-art performance across in-the-wild objects with diverse geometries, varying part counts, and missing pieces. Our code and models will be released.

Zheng Gao, Chen Feng, Ioannis Patras

Whilst contrastive learning yields powerful representations by matching different augmented views of the same instance, it lacks the ability to capture the similarities between different instances. One popular way to address this limitation is by learning global features (after the global pooling) to capture inter-instance relationships based on knowledge distillation, where the global features of the teacher are used to guide the learning of the global features of the student. Inspired by cross-modality learning, we extend this existing framework that only learns from global features by encouraging the global features and intermediate layer features to learn from each other. This leads to our novel self-supervised framework: cross-context learning between global and hypercolumn features (CGH), that enforces the consistency of instance relations between low- and high-level semantics. Specifically, we stack the intermediate feature maps to construct a hypercolumn representation so that we can measure instance relations using two contexts (hypercolumn and global feature) separately, and then use the relations of one context to guide the learning of the other. This cross-context learning allows the model to learn from the differences between the two contexts. The experimental results on linear classification and downstream tasks show that our method outperforms the state-of-the-art methods.

Jiawei Ge, Jiuxin Cao, Xinyi Li et al.

Weakly-Supervised Camouflaged Object Detection (WSCOD) aims to locate and segment objects that are visually concealed within their surrounding scenes, relying solely on sparse supervision such as scribble annotations. Despite recent progress, existing WSCOD methods still lag far behind fully supervised ones due to two major limitations: (1) the pseudo masks generated by general-purpose segmentation models (e.g., SAM) and filtered via rules are often unreliable, as these models lack the task-specific semantic understanding required for effective pseudo labeling in COD; and (2) the neglect of inherent annotation bias in scribbles, which hinders the model from capturing the global structure of camouflaged objects. To overcome these challenges, we propose ${D}^{3}$ETOR, a two-stage WSCOD framework consisting of Debate-Enhanced Pseudo Labeling and Frequency-Aware Progressive Debiasing. In the first stage, we introduce an adaptive entropy-driven point sampling method and a multi-agent debate mechanism to enhance the capability of SAM for COD, improving the interpretability and precision of pseudo masks. In the second stage, we design FADeNet, which progressively fuses multi-level frequency-aware features to balance global semantic understanding with local detail modeling, while dynamically reweighting supervision strength across regions to alleviate scribble bias. By jointly exploiting the supervision signals from both the pseudo masks and scribble semantics, ${D}^{3}$ETOR significantly narrows the gap between weakly and fully supervised COD, achieving state-of-the-art performance on multiple benchmarks.

Zihao Qi, Chen Feng, Fan Zhang et al.

In recent years, user-generated content (UGC) has become one of the major video types consumed via streaming networks. Numerous research contributions have focused on assessing its visual quality through subjective tests and objective modeling. In most cases, objective assessments are based on a no-reference scenario, where the corresponding reference content is assumed not to be available. However, full-reference video quality assessment is also important for UGC in the delivery pipeline, particularly associated with the video transcoding process. In this context, we present a new UGC video quality database, BVI-UGC, for user-generated content transcoding, which contains 60 (non-pristine) reference videos and 1,080 test sequences. In this work, we simulated the creation of non-pristine reference sequences (with a wide range of compression distortions), typical of content uploaded to UGC platforms for transcoding. A comprehensive crowdsourced subjective study was then conducted involving more than 3,500 human participants. Based on this collected subjective data, we benchmarked the performance of 10 full-reference and 11 no-reference quality metrics. Our results demonstrate the poor performance (SROCC values are lower than 0.6) of these metrics in predicting the perceptual quality of UGC in two different scenarios (with or without a reference).

Ruixuan Zhang, Beichen Wang, Juexiao Zhang et al.

The increasing availability of traffic videos functioning on a 24/7/365 time scale has the great potential of increasing the spatio-temporal coverage of traffic accidents, which will help improve traffic safety. However, analyzing footage from hundreds, if not thousands, of traffic cameras in a 24/7/365 working protocol remains an extremely challenging task, as current vision-based approaches primarily focus on extracting raw information, such as vehicle trajectories or individual object detection, but require laborious post-processing to derive actionable insights. We propose SeeUnsafe, a new framework that integrates Multimodal Large Language Model (MLLM) agents to transform video-based traffic accident analysis from a traditional extraction-then-explanation workflow to a more interactive, conversational approach. This shift significantly enhances processing throughput by automating complex tasks like video classification and visual grounding, while improving adaptability by enabling seamless adjustments to diverse traffic scenarios and user-defined queries. Our framework employs a severity-based aggregation strategy to handle videos of various lengths and a novel multimodal prompt to generate structured responses for review and evaluation and enable fine-grained visual grounding. We introduce IMS (Information Matching Score), a new MLLM-based metric for aligning structured responses with ground truth. We conduct extensive experiments on the Toyota Woven Traffic Safety dataset, demonstrating that SeeUnsafe effectively performs accident-aware video classification and visual grounding by leveraging off-the-shelf MLLMs. Source code will be available at \url{https://github.com/ai4ce/SeeUnsafe}.

Chao Chen, Mingzhi Zhu, Ankush Pratap Singh et al.

Humans are remarkably efficient at forming spatial understanding from just a few visual observations. When browsing real estate or navigating unfamiliar spaces, they intuitively select a small set of views that summarize the spatial layout. Inspired by this ability, we introduce scene summarization, the task of condensing long, continuous scene videos into a compact set of spatially diverse keyframes that facilitate global spatial reasoning. Unlike conventional video summarization-which focuses on user-edited, fragmented clips and often ignores spatial continuity-our goal is to mimic how humans abstract spatial layout from sparse views. We propose SceneSum, a two-stage self-supervised pipeline that first clusters video frames using visual place recognition to promote spatial diversity, then selects representative keyframes from each cluster under resource constraints. When camera trajectories are available, a lightweight supervised loss further refines clustering and selection. Experiments on real and simulated indoor datasets show that SceneSum produces more spatially informative summaries and outperforms existing video summarization baselines.

Chao Chen, Ruoyu Wang, Yuliang Guo et al.

Autonomous driving in complex urban scenarios requires 3D perception to be both comprehensive and precise. Traditional 3D perception methods focus on object detection, resulting in sparse representations that lack environmental detail. Recent approaches estimate 3D occupancy around vehicles for a more comprehensive scene representation. However, dense 3D occupancy prediction increases computational demands, challenging the balance between efficiency and resolution. High-resolution occupancy grids offer accuracy but demand substantial computational resources, while low-resolution grids are efficient but lack detail. To address this dilemma, we introduce AdaOcc, a novel adaptive-resolution, multi-modal prediction approach. Our method integrates object-centric 3D reconstruction and holistic occupancy prediction within a single framework, performing highly detailed and precise 3D reconstruction only in regions of interest (ROIs). These high-detailed 3D surfaces are represented in point clouds, thus their precision is not constrained by the predefined grid resolution of the occupancy map. We conducted comprehensive experiments on the nuScenes dataset, demonstrating significant improvements over existing methods. In close-range scenarios, we surpass previous baselines by over 13% in IOU, and over 40% in Hausdorff distance. In summary, AdaOcc offers a more versatile and effective framework for delivering accurate 3D semantic occupancy prediction across diverse driving scenarios.

Chen Feng, Tianhao Peng, Fan Zhang et al.

Recent works in video quality assessment (VQA) typically employ monolithic models that typically predict a single quality score for each test video. These approaches cannot provide diagnostic, interpretable feedback, offering little insight into why the video quality is degraded. Most of them are also specialized, format-specific metrics rather than truly ``generic" solutions, as they are designed to learn a compromised representation from disparate perceptual domains. To address these limitations, this paper proposes Unified-VQA, a framework that provides a single, unified quality model applicable to various distortion types within multiple video formats by recasting generic VQA as a Diagnostic Mixture-of-Experts (MoE) problem. Unified-VQA employs multiple ``perceptual experts'' dedicated to distinct perceptual domains. A novel multi-proxy expert training strategy is designed to optimize each expert using a ranking-inspired loss, guided by the most suitable proxy metric for its domain. We also integrated a diagnostic multi-task head into this framework to generate a global quality score and an interpretable multi-dimensional artifact vector, which is optimized using a weakly-supervised learning strategy, leveraging the known properties of the large-scale training database generated for this work. With static model parameters (without retraining or fine-tuning), Unified-VQA demonstrates consistent and superior performance compared to over 18 benchmark methods for both generic VQA and diagnostic artifact detection tasks across 17 databases containing diverse streaming artifacts in HD, UHD, HDR and HFR formats. This work represents an important step towards practical, actionable, and interpretable video quality assessment.

Diwei Sheng, Vijayraj Gohil, Satyam Gaba et al.

Scene change detection (SCD) is crucial for urban monitoring and navigation but remains challenging in real-world environments due to lighting variations, seasonal shifts, viewpoint differences, and complex urban layouts. Existing methods rely primarily on low-level visual features, limiting their ability to accurately identify changed objects amid the visual complexity of urban scenes. In this paper, we propose LangSCD, a vision-language framework for scene change detection that overcomes this single-modal limitation by incorporating semantic reasoning through language. Our approach introduces a modular language component that leverages vision-language models (VLMs) to generate textual descriptions of scene changes, which are fused with visual features through a cross-modal feature enhancer. We further introduce a geometric-semantic matching module that refines the predicted masks by enforcing semantic consistency and spatial completeness. Existing real-world scene change detection benchmarks provide only binary change annotations, which are insufficient for downstream applications requiring fine-grained understanding of scene dynamics. To address this limitation, we introduce NYC-CD, a large-scale dataset of 8,122 real-world image pairs collected in New York City with multiclass change annotations generated through a semi-automatic pipeline. Extensive experiments across multiple street-view benchmarks demonstrate that our language and matching modules consistently improve existing change-detection architectures, achieving state-of-the-art performance and highlighting the value of integrating linguistic reasoning with visual representations for robust scene change detection.

Jisang Han, Sunghwan Hong, Jaewoo Jung et al.

Reliable 3D reconstruction from in-the-wild image collections is often hindered by "noisy" images-irrelevant inputs with little or no view overlap with others. While traditional Structure-from-Motion pipelines handle such cases through geometric verification and outlier rejection, feed-forward 3D reconstruction models lack these explicit mechanisms, leading to degraded performance under in-the-wild conditions. In this paper, we discover that the existing feed-forward reconstruction model, e.g., VGGT, despite lacking explicit outlier-rejection mechanisms or noise-aware training, can inherently distinguish distractor images. Through an in-depth analysis under varying proportions of synthetic distractors, we identify a specific layer that naturally exhibits outlier-suppressing behavior. Further probing reveals that this layer encodes discriminative internal representations that enable an effective noise-filtering capability, which we simply leverage to perform outlier-view rejection in feed-forward 3D reconstruction without any additional fine-tuning or supervision. Extensive experiments on both controlled and in-the-wild datasets demonstrate that this implicit filtering mechanism is consistent and generalizes well across diverse scenarios.

Chao Chen, Shunyu Yao, Yuanwu He et al.

Precise parking requires an end-to-end system where perception adaptively provides policy-relevant details-especially in critical areas where fine control decisions are essential. End-to-end learning offers a unified framework by directly mapping sensor inputs to control actions, but existing approaches lack effective synergy between perception and control. We find that transformer-based self-attention, when used alone, tends to produce unstable and temporally inconsistent spatial attention, which undermines the reliability of downstream policy decisions over time. Instead, we propose CAA-Policy, an end-to-end imitation learning system that allows control signal to guide the learning of visual attention via a novel Control-Aided Attention (CAA) mechanism. For the first time, we train such an attention module in a self-supervised manner, using backpropagated gradients from the control outputs instead of from the training loss. This strategy encourages the attention to focus on visual features that induce high variance in action outputs, rather than merely minimizing the training loss-a shift we demonstrate leads to a more robust and generalizable policy. To further enhance stability, CAA-Policy integrates short-horizon waypoint prediction as an auxiliary task, and introduces a separately trained motion prediction module to robustly track the target spot over time. Extensive experiments in the CARLA simulator show that \titlevariable~consistently surpasses both the end-to-end learning baseline and the modular BEV segmentation + hybrid A* pipeline, achieving superior accuracy, robustness, and interpretability. Code is released at https://github.com/Joechencc/CAAPolicy.

Christopher Blake, Chen Feng, Xuechao Wang et al.

A proof of the security of the Bitcoin protocol is made rigorous, and simplified in certain parts. A computational model in which an adversary can delay transmission of blocks by time $Δ$ is considered. The protocol is generalized to allow blocks of different scores and a proof within this more general model is presented. An approach used in a previous paper that used random walk theory is shown through a counterexample to be incorrect; an approach involving a punctured block arrival process is shown to remedy this error. Thus, it is proven that with probability one, the Bitcoin protocol will have infinitely many honest blocks so long as the fully-delayed honest mining rate exceeds the adversary mining rate. This means that an adversary cannot censor future transactions of a user in perpetuity, which would render the protocol useless.

Heyang Yu, Yinan Han, Xiangyu Zhang et al.

Humans rely on the synergistic control of head (cephalomotor) and eye (oculomotor) to efficiently search for visual information in 360°. However, prior approaches to visual search are limited to a static image, neglecting the physical embodiment and its interaction with the 3D world. How can we develop embodied visual search agents as efficient as humans while bypassing the constraints imposed by real-world hardware? To this end, we propose humanoid visual search where a humanoid agent actively rotates its head to search for objects or paths in an immersive world represented by a 360° panoramic image. To study visual search in visually-crowded real-world scenarios, we build H* Bench, a new benchmark that moves beyond household scenes to challenging in-the-wild scenes that necessitate advanced visual-spatial reasoning capabilities, such as transportation hubs, large-scale retail spaces, urban streets, and public institutions. Our experiments first reveal that even top-tier proprietary models falter, achieving only ~30% success in object and path search. We then use post-training techniques to enhance the open-source Qwen2.5-VL, increasing its success rate by over threefold for both object search (14.83% to 47.38%) and path search (6.44% to 24.94%). Notably, the lower ceiling of path search reveals its inherent difficulty, which we attribute to the demand for sophisticated spatial commonsense. Our results not only show a promising path forward but also quantify the immense challenge that remains in building MLLM agents that can be seamlessly integrated into everyday human life.

Daxiang Dong, Mingming Zheng, Dong Xu et al.

We present Qianfan-VL, a series of multimodal large language models ranging from 3B to 70B parameters, achieving state-of-the-art performance through innovative domain enhancement techniques. Our approach employs multi-stage progressive training and high-precision data synthesis pipelines, which prove to be critical technologies for enhancing domain-specific capabilities while maintaining strong general performance. Qianfan-VL achieves comparable results to leading open-source models on general benchmarks, with state-of-the-art performance on benchmarks such as CCBench, SEEDBench IMG, ScienceQA, and MMStar. The domain enhancement strategy delivers significant advantages in OCR and document understanding, validated on both public benchmarks (OCRBench 873, DocVQA 94.75%) and in-house evaluations. Notably, Qianfan-VL-8B and 70B variants incorporate long chain-of-thought capabilities, demonstrating superior performance on mathematical reasoning (MathVista 78.6%) and logical inference tasks. All models are trained entirely on Baidu's Kunlun P800 chips, validating the capability of large-scale AI infrastructure to train SOTA-level multimodal models with over 90% scaling efficiency on 5000 chips for a single task. This work establishes an effective methodology for developing domain-enhanced multimodal models suitable for diverse enterprise deployment scenarios.

Zhenyang Ni, Zixing Lei, Yifan Lu et al.

Collaborative perception has garnered considerable attention due to its capacity to address several inherent challenges in single-agent perception, including occlusion and out-of-range issues. However, existing collaborative perception systems heavily rely on precise localization systems to establish a consistent spatial coordinate system between agents. This reliance makes them susceptible to large pose errors or malicious attacks, resulting in substantial reductions in perception performance. To address this, we propose~$\mathtt{CoBEVGlue}$, a novel self-localized collaborative perception system, which achieves more holistic and robust collaboration without using an external localization system. The core of~$\mathtt{CoBEVGlue}$ is a novel spatial alignment module, which provides the relative poses between agents by effectively matching co-visible objects across agents. We validate our method on both real-world and simulated datasets. The results show that i) $\mathtt{CoBEVGlue}$ achieves state-of-the-art detection performance under arbitrary localization noises and attacks; and ii) the spatial alignment module can seamlessly integrate with a majority of previous methods, enhancing their performance by an average of $57.7\%$. Code is available at https://github.com/VincentNi0107/CoBEVGlue

Jing Zhang, Irving Fang, Juexiao Zhang et al.

Lithic Use-Wear Analysis (LUWA) using microscopic images is an underexplored vision-for-science research area. It seeks to distinguish the worked material, which is critical for understanding archaeological artifacts, material interactions, tool functionalities, and dental records. However, this challenging task goes beyond the well-studied image classification problem for common objects. It is affected by many confounders owing to the complex wear mechanism and microscopic imaging, which makes it difficult even for human experts to identify the worked material successfully. In this paper, we investigate the following three questions on this unique vision task for the first time:(i) How well can state-of-the-art pre-trained models (like DINOv2) generalize to the rarely seen domain? (ii) How can few-shot learning be exploited for scarce microscopic images? (iii) How do the ambiguous magnification and sensing modality influence the classification accuracy? To study these, we collaborated with archaeologists and built the first open-source and the largest LUWA dataset containing 23,130 microscopic images with different magnifications and sensing modalities. Extensive experiments show that existing pre-trained models notably outperform human experts but still leave a large gap for improvements. Most importantly, the LUWA dataset provides an underexplored opportunity for vision and learning communities and complements existing image classification problems on common objects.

Yiming Li, Dekun Ma, Ziyan An et al.

Vehicle-to-everything (V2X) communication techniques enable the collaboration between vehicles and many other entities in the neighboring environment, which could fundamentally improve the perception system for autonomous driving. However, the lack of a public dataset significantly restricts the research progress of collaborative perception. To fill this gap, we present V2X-Sim, a comprehensive simulated multi-agent perception dataset for V2X-aided autonomous driving. V2X-Sim provides: (1) \hl{multi-agent} sensor recordings from the road-side unit (RSU) and multiple vehicles that enable collaborative perception, (2) multi-modality sensor streams that facilitate multi-modality perception, and (3) diverse ground truths that support various perception tasks. Meanwhile, we build an open-source testbed and provide a benchmark for the state-of-the-art collaborative perception algorithms on three tasks, including detection, tracking and segmentation. V2X-Sim seeks to stimulate collaborative perception research for autonomous driving before realistic datasets become widely available. Our dataset and code are available at \url{https://ai4ce.github.io/V2X-Sim/}.

Chen Feng, Georgios Tzimiropoulos, Ioannis Patras

Despite the large progress in supervised learning with neural networks, there are significant challenges in obtaining high-quality, large-scale and accurately labelled datasets. In such a context, how to learn in the presence of noisy labels has received more and more attention. As a relatively complex problem, in order to achieve good results, current approaches often integrate components from several fields, such as supervised learning, semi-supervised learning, transfer learning and resulting in complicated methods. Furthermore, they often make multiple assumptions about the type of noise of the data. This affects the model robustness and limits its performance under different noise conditions. In this paper, we consider a novel problem setting, Learning with Unknown Label Noise}(LULN), that is, learning when both the degree and the type of noise are unknown. Under this setting, unlike previous methods that often introduce multiple assumptions and lead to complex solutions, we propose a simple, efficient and robust framework named Sample Selection and Relabelling(SSR), that with a minimal number of hyperparameters achieves SOTA results in various conditions. At the heart of our method is a sample selection and relabelling mechanism based on a non-parametric KNN classifier~(NPK) $g_q$ and a parametric model classifier~(PMC) $g_p$, respectively, to select the clean samples and gradually relabel the noisy samples. Without bells and whistles, such as model co-training, self-supervised pre-training and semi-supervised learning, and with robustness concerning the settings of its few hyper-parameters, our method significantly surpasses previous methods on both CIFAR10/CIFAR100 with synthetic noise and real-world noisy datasets such as WebVision, Clothing1M and ANIMAL-10N. Code is available at https://github.com/MrChenFeng/SSR_BMVC2022.

Yiming Li, Shunli Ren, Pengxiang Wu et al.

To promote better performance-bandwidth trade-off for multi-agent perception, we propose a novel distilled collaboration graph (DiscoGraph) to model trainable, pose-aware, and adaptive collaboration among agents. Our key novelties lie in two aspects. First, we propose a teacher-student framework to train DiscoGraph via knowledge distillation. The teacher model employs an early collaboration with holistic-view inputs; the student model is based on intermediate collaboration with single-view inputs. Our framework trains DiscoGraph by constraining post-collaboration feature maps in the student model to match the correspondences in the teacher model. Second, we propose a matrix-valued edge weight in DiscoGraph. In such a matrix, each element reflects the inter-agent attention at a specific spatial region, allowing an agent to adaptively highlight the informative regions. During inference, we only need to use the student model named as the distilled collaboration network (DiscoNet). Attributed to the teacher-student framework, multiple agents with the shared DiscoNet could collaboratively approach the performance of a hypothetical teacher model with a holistic view. Our approach is validated on V2X-Sim 1.0, a large-scale multi-agent perception dataset that we synthesized using CARLA and SUMO co-simulation. Our quantitative and qualitative experiments in multi-agent 3D object detection show that DiscoNet could not only achieve a better performance-bandwidth trade-off than the state-of-the-art collaborative perception methods, but also bring more straightforward design rationale. Our code is available on https://github.com/ai4ce/DiscoNet.

Ruoyu Wang, Xuchu Xu, Li Ding et al.

PoseNet can map a photo to the position where it is taken, which is appealing in robotics. However, training PoseNet requires full supervision, where ground truth positions are non-trivial to obtain. Can we train PoseNet without knowing the ground truth positions for each observation? We show that this is possible via constraint-based weak-supervision, leading to the proposed framework: DeepGPS. Particularly, using wheel-encoder-estimated distances traveled by a robot along random straight line segments as constraints between PoseNet outputs, DeepGPS can achieve a relative positioning error of less than 2%. Moreover, training DeepGPS can be done as auto-calibration with almost no human attendance, which is more attractive than its competing methods that typically require careful and expert-level manual calibration. We conduct various experiments on simulated and real datasets to demonstrate the general applicability, effectiveness, and accuracy of DeepGPS, and perform a comprehensive analysis of its robustness. Our code is available at https://ai4ce.github.io/DeepGPS/.

Yiming Li, Congcong Wen, Felix Juefei-Xu et al.

LiDAR point clouds collected from a moving vehicle are functions of its trajectories, because the sensor motion needs to be compensated to avoid distortions. When autonomous vehicles are sending LiDAR point clouds to deep networks for perception and planning, could the motion compensation consequently become a wide-open backdoor in those networks, due to both the adversarial vulnerability of deep learning and GPS-based vehicle trajectory estimation that is susceptible to wireless spoofing? We demonstrate such possibilities for the first time: instead of directly attacking point cloud coordinates which requires tampering with the raw LiDAR readings, only adversarial spoofing of a self-driving car's trajectory with small perturbations is enough to make safety-critical objects undetectable or detected with incorrect positions. Moreover, polynomial trajectory perturbation is developed to achieve a temporally-smooth and highly-imperceptible attack. Extensive experiments on 3D object detection have shown that such attacks not only lower the performance of the state-of-the-art detectors effectively, but also transfer to other detectors, raising a red flag for the community. The code is available on https://ai4ce.github.io/FLAT/.

Haozhe Qi, Chen Feng, Zhiguo Cao et al.

Towards 3D object tracking in point clouds, a novel point-to-box network termed P2B is proposed in an end-to-end learning manner. Our main idea is to first localize potential target centers in 3D search area embedded with target information. Then point-driven 3D target proposal and verification are executed jointly. In this way, the time-consuming 3D exhaustive search can be avoided. Specifically, we first sample seeds from the point clouds in template and search area respectively. Then, we execute permutation-invariant feature augmentation to embed target clues from template into search area seeds and represent them with target-specific features. Consequently, the augmented search area seeds regress the potential target centers via Hough voting. The centers are further strengthened with seed-wise targetness scores. Finally, each center clusters its neighbors to leverage the ensemble power for joint 3D target proposal and verification. We apply PointNet++ as our backbone and experiments on KITTI tracking dataset demonstrate P2B's superiority (~10%'s improvement over state-of-the-art). Note that P2B can run with 40FPS on a single NVIDIA 1080Ti GPU. Our code and model are available at https://github.com/HaozheQi/P2B.

Ruoyu Wang, Shiheng Wang, Songyu Du et al.

Soft bodies made from flexible and deformable materials are popular in many robotics applications, but their proprioceptive sensing has been a long-standing challenge. In other words, there has hardly been a method to measure and model the high-dimensional 3D shapes of soft bodies with internal sensors. We propose a framework to measure the high-resolution 3D shapes of soft bodies in real-time with embedded cameras. The cameras capture visual patterns inside a soft body, and a convolutional neural network (CNN) produces a latent code representing the deformation state, which can then be used to reconstruct the body's 3D shape using another neural network. We test the framework on various soft bodies, such as a Baymax-shaped toy, a latex balloon, and some soft robot fingers, and achieve real-time computation ($\leq$2.5ms/frame) for robust shape estimation with high precision ($\leq$1% relative error) and high resolution. We believe the method could be applied to soft robotics and human-robot interaction for proprioceptive shape sensing. Our code is available at https://ai4ce.github.io/Deep-Soft-Prorioception/

Li Ding, Chen Feng

We propose DeepMapping, a novel registration framework using deep neural networks (DNNs) as auxiliary functions to align multiple point clouds from scratch to a globally consistent frame. We use DNNs to model the highly non-convex mapping process that traditionally involves hand-crafted data association, sensor pose initialization, and global refinement. Our key novelty is that "training" these DNNs with properly defined unsupervised losses is equivalent to solving the underlying registration problem, but less sensitive to good initialization than ICP. Our framework contains two DNNs: a localization network that estimates the poses for input point clouds, and a map network that models the scene structure by estimating the occupancy status of global coordinates. This allows us to convert the registration problem to a binary occupancy classification, which can be solved efficiently using gradient-based optimization. We further show that DeepMapping can be readily extended to address the problem of Lidar SLAM by imposing geometric constraints between consecutive point clouds. Experiments are conducted on both simulated and real datasets. Qualitative and quantitative comparisons demonstrate that DeepMapping often enables more robust and accurate global registration of multiple point clouds than existing techniques. Our code is available at https://ai4ce.github.io/DeepMapping/.

Yaoqing Yang, Chen Feng, Yiru Shen et al.

Recent deep networks that directly handle points in a point set, e.g., PointNet, have been state-of-the-art for supervised learning tasks on point clouds such as classification and segmentation. In this work, a novel end-to-end deep auto-encoder is proposed to address unsupervised learning challenges on point clouds. On the encoder side, a graph-based enhancement is enforced to promote local structures on top of PointNet. Then, a novel folding-based decoder deforms a canonical 2D grid onto the underlying 3D object surface of a point cloud, achieving low reconstruction errors even for objects with delicate structures. The proposed decoder only uses about 7% parameters of a decoder with fully-connected neural networks, yet leads to a more discriminative representation that achieves higher linear SVM classification accuracy than the benchmark. In addition, the proposed decoder structure is shown, in theory, to be a generic architecture that is able to reconstruct an arbitrary point cloud from a 2D grid. Our code is available at http://www.merl.com/research/license#FoldingNet

Yiru Shen, Chen Feng, Yaoqing Yang et al.

Unlike on images, semantic learning on 3D point clouds using a deep network is challenging due to the naturally unordered data structure. Among existing works, PointNet has achieved promising results by directly learning on point sets. However, it does not take full advantage of a point's local neighborhood that contains fine-grained structural information which turns out to be helpful towards better semantic learning. In this regard, we present two new operations to improve PointNet with a more efficient exploitation of local structures. The first one focuses on local 3D geometric structures. In analogy to a convolution kernel for images, we define a point-set kernel as a set of learnable 3D points that jointly respond to a set of neighboring data points according to their geometric affinities measured by kernel correlation, adapted from a similar technique for point cloud registration. The second one exploits local high-dimensional feature structures by recursive feature aggregation on a nearest-neighbor-graph computed from 3D positions. Experiments show that our network can efficiently capture local information and robustly achieve better performances on major datasets. Our code is available at http://www.merl.com/research/license#KCNet

Anbang Yang, Ge Jin, Junjie Huang et al.

Real-time visual localization often utilizes online computing, for which query images or videos are transmitted to remote servers for visual place recognition (VPR). However, limited network bandwidth necessitates image-quality reduction and thus the degradation of global image descriptors, reducing VPR accuracy. We address this issue at the descriptor extraction level with a knowledge-distillation methodology that learns feature representations from high-quality images to extract more discriminative descriptors from low-quality images. Our approach includes the Inter-channel Correlation Knowledge Distillation (ICKD) loss, Mean Squared Error (MSE) loss, and Triplet loss. We validate the proposed losses on multiple VPR methods and datasets subjected to JPEG compression, resolution reduction, and video quantization. We obtain significant improvements in VPR recall rates under all three tested modalities of lowered image quality. Furthermore, we fill a gap in VPR literature on video-based data and its influence on VPR performance. This work contributes to more reliable place recognition in resource-constrained environments.

Jiahao Huang, Fengyan Lin, Xuechao Yang et al.

The development of affective multimodal language models (MLMs) has long been constrained by a gap between low-level perception and high-level interaction, leading to fragmented affective capabilities and limited generalization. To bridge this gap, we propose a cognitively inspired three-level hierarchy that organizes affective tasks according to their cognitive depth-perception, understanding, and interaction-and provides a unified conceptual foundation for advancing affective modeling. Guided by this hierarchy, we introduce Nano-EmoX, a small-scale multitask MLM, and P2E (Perception-to-Empathy), a curriculum-based training framework. Nano-EmoX integrates a suite of omni-modal encoders, including an enhanced facial encoder and a fusion encoder, to capture key multimodal affective cues and improve cross-task transferability. The outputs are projected into a unified language space via heterogeneous adapters, empowering a lightweight language model to tackle diverse affective tasks. Concurrently, P2E progressively cultivates emotional intelligence by aligning rapid perception with chain-of-thought-driven empathy. To the best of our knowledge, Nano-EmoX is the first compact MLM (2.2B) to unify six core affective tasks across all three hierarchy levels, achieving state-of-the-art or highly competitive performance across multiple benchmarks, demonstrating excellent efficiency and generalization.

Jiawei Ge, Xintian Zhang, Jiuxin Cao et al.

Cross-view Referring Multi-Object Tracking (CRMOT) aims to track multiple objects specified by natural language across multiple camera views, with globally consistent identities. Despite recent progress, existing methods rely heavily on costly frame-level spatial annotations and cross-view identity supervision. To reduce such reliance, we explore CRMOT under weak supervision by leveraging the capabilities of foundation models. However, our empirical study shows that directly applying foundation models such as SAM2 and SAM3, even with task-specific modifications, fails to accurately understand referring expressions and maintain consistent identities across views. Yet, they remain effective at producing reliable object tracklets that can serve as pseudo supervision. We therefore repurpose foundation models as pseudo-label generators and propose a two-stage framework for weakly supervised CRMOT, using only object category labels as coarse-grained supervision. In the first stage, we design an Affinity-guided Cross-view Re-prompting strategy to refine and associate SAM3-generated tracklets across cameras, producing reliable cross-view pseudo labels for subsequent training. In the second stage, we introduce ViewSAM, a CRMOT model built upon SAM2 that explicitly models view-aware cross-modal semantics. By formulating view-induced variations as learnable conditions, ViewSAM bridges the gap between view-variant visual observations and view-invariant textual expressions, enabling robust cross-view referring tracking with only approximately 10% additional parameters. Extensive experiments demonstrate that ViewSAM achieves SOTA performance under weak supervision and remains competitive with fully supervised methods.