Lei Jin

Yabo Xiao, Xiaojuan Wang, Dongdong Yu et al.

Multi-person pose estimation generally follows top-down and bottom-up paradigms. Both of them use an extra stage ($\boldsymbol{e.g.,}$ human detection in top-down paradigm or grouping process in bottom-up paradigm) to build the relationship between the human instance and corresponding keypoints, thus leading to the high computation cost and redundant two-stage pipeline. To address the above issue, we propose to represent the human parts as adaptive points and introduce a fine-grained body representation method. The novel body representation is able to sufficiently encode the diverse pose information and effectively model the relationship between the human instance and corresponding keypoints in a single-forward pass. With the proposed body representation, we further deliver a compact single-stage multi-person pose regression network, termed as AdaptivePose. During inference, our proposed network only needs a single-step decode operation to form the multi-person pose without complex post-processes and refinements. We employ AdaptivePose for both 2D/3D multi-person pose estimation tasks to verify the effectiveness of AdaptivePose. Without any bells and whistles, we achieve the most competitive performance on MS COCO and CrowdPose in terms of accuracy and speed. Furthermore, the outstanding performance on MuCo-3DHP and MuPoTS-3D further demonstrates the effectiveness and generalizability on 3D scenes. Code is available at https://github.com/buptxyb666/AdaptivePose.

Xuechao Zou, Kai Li, Junliang Xing et al.

Optical satellite images are a critical data source; however, cloud cover often compromises their quality, hindering image applications and analysis. Consequently, effectively removing clouds from optical satellite images has emerged as a prominent research direction. While recent advancements in cloud removal primarily rely on generative adversarial networks, which may yield suboptimal image quality, diffusion models have demonstrated remarkable success in diverse image-generation tasks, showcasing their potential in addressing this challenge. This paper presents a novel framework called DiffCR, which leverages conditional guided diffusion with deep convolutional networks for high-performance cloud removal for optical satellite imagery. Specifically, we introduce a decoupled encoder for conditional image feature extraction, providing a robust color representation to ensure the close similarity of appearance information between the conditional input and the synthesized output. Moreover, we propose a novel and efficient time and condition fusion block within the cloud removal model to accurately simulate the correspondence between the appearance in the conditional image and the target image at a low computational cost. Extensive experimental evaluations on two commonly used benchmark datasets demonstrate that DiffCR consistently achieves state-of-the-art performance on all metrics, with parameter and computational complexities amounting to only 5.1% and 5.4%, respectively, of those previous best methods. The source code, pre-trained models, and all the experimental results will be publicly available at https://github.com/XavierJiezou/DiffCR upon the paper's acceptance of this work.

Gen Luo, Yiyi Zhou, Lei Jin et al.

Semi-supervised object detection (SSOD) is a research hot spot in computer vision, which can greatly reduce the requirement for expensive bounding-box annotations. Despite great success, existing progress mainly focuses on two-stage detection networks like FasterRCNN, while the research on one-stage detectors is often ignored. In this paper, we focus on the semi-supervised learning for the advanced and popular one-stage detection network YOLOv5. Compared with Faster-RCNN, the implementation of YOLOv5 is much more complex, and the various training techniques used in YOLOv5 can also reduce the benefit of SSOD. In addition to this challenge, we also reveal two key issues in one-stage SSOD, which are low-quality pseudo-labeling and multi-task optimization conflict, respectively. To address these issues, we propose a novel teacher-student learning recipe called OneTeacher with two innovative designs, namely Multi-view Pseudo-label Refinement (MPR) and Decoupled Semi-supervised Optimization (DSO). In particular, MPR improves the quality of pseudo-labels via augmented-view refinement and global-view filtering, and DSO handles the joint optimization conflicts via structure tweaks and task-specific pseudo-labeling. In addition, we also carefully revise the implementation of YOLOv5 to maximize the benefits of SSOD, which is also shared with the existing SSOD methods for fair comparison. To validate OneTeacher, we conduct extensive experiments on COCO and Pascal VOC. The extensive experiments show that OneTeacher can not only achieve superior performance than the compared methods, e.g., 15.0% relative AP gains over Unbiased Teacher, but also well handle the key issues in one-stage SSOD. Our source code is available at: https://github.com/luogen1996/OneTeacher.

Yabo Xiao, Kai Su, Xiaojuan Wang et al.

We propose a sparse end-to-end multi-person pose regression framework, termed QueryPose, which can directly predict multi-person keypoint sequences from the input image. The existing end-to-end methods rely on dense representations to preserve the spatial detail and structure for precise keypoint localization. However, the dense paradigm introduces complex and redundant post-processes during inference. In our framework, each human instance is encoded by several learnable spatial-aware part-level queries associated with an instance-level query. First, we propose the Spatial Part Embedding Generation Module (SPEGM) that considers the local spatial attention mechanism to generate several spatial-sensitive part embeddings, which contain spatial details and structural information for enhancing the part-level queries. Second, we introduce the Selective Iteration Module (SIM) to adaptively update the sparse part-level queries via the generated spatial-sensitive part embeddings stage-by-stage. Based on the two proposed modules, the part-level queries are able to fully encode the spatial details and structural information for precise keypoint regression. With the bipartite matching, QueryPose avoids the hand-designed post-processes and surpasses the existing dense end-to-end methods with 73.6 AP on MS COCO mini-val set and 72.7 AP on CrowdPose test set. Code is available at https://github.com/buptxyb666/QueryPose.

Shufan Zhang, Ziyue Lin, Bairun Wang et al.

Video reasoning aims to understand complex temporal events and causal relationships within videos. Recently, Chain-of-Thought (CoT) has been introduced to this field to enhance reasoning accuracy. However, existing CoT-based video reasoning methods primarily rely on text-only information for logical deduction, overlooking critical visual information during the inference process. Inspired by the human cognitive mechanism of reviewing visual segments during inference, we propose VTI-CoT, a Visual-Textual Interleaved CoT framework. VTI-CoT integrates textual reasoning steps with corresponding visual frames. Given the scarcity of visual-textual interleaved CoT in existing datasets, we develop an automated annotation pipeline to construct high-quality multimodal CoT data. Further, reasoning over long-form videos entails increasingly long CoT token sequences, which severely hinders training convergence and efficiency. To address this, we employ Optical Character Recognition (OCR)-based compression techniques to compress CoT supervision signals into a single canvas. Experimental results demonstrate that VTI-CoT achieves state-of-the-art performance among models of the same parameter scale while significantly improving training efficiency.

Xiaoyi Chen, Siyuan Tang, Rui Zhu et al.

The rapid advancements of large language models (LLMs) have raised public concerns about the privacy leakage of personally identifiable information (PII) within their extensive training datasets. Recent studies have demonstrated that an adversary could extract highly sensitive privacy data from the training data of LLMs with carefully designed prompts. However, these attacks suffer from the model's tendency to hallucinate and catastrophic forgetting (CF) in the pre-training stage, rendering the veracity of divulged PIIs negligible. In our research, we propose a novel attack, Janus, which exploits the fine-tuning interface to recover forgotten PIIs from the pre-training data in LLMs. We formalize the privacy leakage problem in LLMs and explain why forgotten PIIs can be recovered through empirical analysis on open-source language models. Based upon these insights, we evaluate the performance of Janus on both open-source language models and two latest LLMs, i.e., GPT-3.5-Turbo and LLaMA-2-7b. Our experiment results show that Janus amplifies the privacy risks by over 10 times in comparison with the baseline and significantly outperforms the state-of-the-art privacy extraction attacks including prefix attacks and in-context learning (ICL). Furthermore, our analysis validates that existing fine-tuning APIs provided by OpenAI and Azure AI Studio are susceptible to our Janus attack, allowing an adversary to conduct such an attack at a low cost.

Xue-Feng Zhu, Tianyang Xu, Jian Zhao et al.

Unmanned Aerial Vehicles (UAVs) have been widely used in many areas, including transportation, surveillance, and military. However, their potential for safety and privacy violations is an increasing issue and highly limits their broader applications, underscoring the critical importance of UAV perception and defense (anti-UAV). Still, previous works have simplified such an anti-UAV task as a tracking problem, where the prior information of UAVs is always provided; such a scheme fails in real-world anti-UAV tasks (i.e. complex scenes, indeterminate-appear and -reappear UAVs, and real-time UAV surveillance). In this paper, we first formulate a new and practical anti-UAV problem featuring the UAVs perception in complex scenes without prior UAVs information. To benchmark such a challenging task, we propose the largest UAV dataset dubbed AntiUAV600 and a new evaluation metric. The AntiUAV600 comprises 600 video sequences of challenging scenes with random, fast, and small-scale UAVs, with over 723K thermal infrared frames densely annotated with bounding boxes. Finally, we develop a novel anti-UAV approach via an evidential collaboration of global UAVs detection and local UAVs tracking, which effectively tackles the proposed problem and can serve as a strong baseline for future research. Extensive experiments show our method outperforms SOTA approaches and validate the ability of AntiUAV600 to enhance UAV perception performance due to its large scale and complexity. Our dataset, pretrained models, and source codes will be released publically.

Jiyao Liu, Jianghan Shen, Sida Song et al.

Recent advances in deep research systems enable large language models to retrieve, synthesize, and reason over large-scale external knowledge. In medicine, developing clinical guidelines critically depends on such deep evidence integration. However, existing benchmarks fail to evaluate this capability in realistic workflows requiring multi-step evidence integration and expert-level judgment. To address this gap, we introduce MedProbeBench, the first benchmark leveraging high-quality clinical guidelines as expert-level references. Medical guidelines, with their rigorous standards in neutrality and verifiability, represent the pinnacle of medical expertise and pose substantial challenges for deep research agents. For evaluation, we propose MedProbe-Eval, a comprehensive evaluation framework featuring: (1) Holistic Rubrics with 1,200+ task-adaptive rubric criteria for comprehensive quality assessment, and (2) Fine-grained Evidence Verification for rigorous validation of evidence precision, grounded in 5,130+ atomic claims. Evaluation of 17 LLMs and deep research agents reveals critical gaps in evidence integration and guideline generation, underscoring the substantial distance between current capabilities and expert-level clinical guideline development. Project: https://github.com/uni-medical/MedProbeBench

Jianqiang Xia, DianXi Shi, Ke Song et al.

Most existing RGB-T tracking networks extract modality features in a separate manner, which lacks interaction and mutual guidance between modalities. This limits the network's ability to adapt to the diverse dual-modality appearances of targets and the dynamic relationships between the modalities. Additionally, the three-stage fusion tracking paradigm followed by these networks significantly restricts the tracking speed. To overcome these problems, we propose a unified single-stage Transformer RGB-T tracking network, namely USTrack, which unifies the above three stages into a single ViT (Vision Transformer) backbone with a dual embedding layer through self-attention mechanism. With this structure, the network can extract fusion features of the template and search region under the mutual interaction of modalities. Simultaneously, relation modeling is performed between these features, efficiently obtaining the search region fusion features with better target-background discriminability for prediction. Furthermore, we introduce a novel feature selection mechanism based on modality reliability to mitigate the influence of invalid modalities for prediction, further improving the tracking performance. Extensive experiments on three popular RGB-T tracking benchmarks demonstrate that our method achieves new state-of-the-art performance while maintaining the fastest inference speed 84.2FPS. In particular, MPR/MSR on the short-term and long-term subsets of VTUAV dataset increased by 11.1$\%$/11.7$\%$ and 11.3$\%$/9.7$\%$.

Tao Wang, Lei Jin, Zhihua Wu et al.

Text-to-motion generation, which translates textual descriptions into human motions, faces the challenge that users often struggle to precisely convey their intended motions through text alone. To address this issue, this paper introduces DrawMotion, an efficient diffusion-based framework designed for multi-condition scenarios. DrawMotion generates motions based on both a conventional text condition and a novel hand-drawing condition, which provide semantic and spatial control over the generated motions, respectively. Specifically, we tackle the fine-grained motion generation task from three perspectives: 1) freehand drawing condition. To accurately capture users' intended motions without requiring tedious textual input, we develop an algorithm to automatically generate hand-drawn stickman sketches across different dataset formats; 2) multi-condition fusion. We propose a Multi-Condition Module (MCM) that is integrated into the diffusion process, enabling the model to exploit all possible condition combinations while reducing computational complexity compared to conventional approaches; and 3) training-free guidance. Notably, the MCM in DrawMotion ensures that its intermediate features lie in a continuous space, allowing classifier-guidance gradients to update the features and thereby aligning the generated motions with user intentions while preserving fidelity. Quantitative experiments and user studies demonstrate that the freehand drawing approach reduces user time by approximately 46.7% when generating motions aligned with their imagination. The code, demos, and relevant data are publicly available at https://github.com/InvertedForest/DrawMotion.

Jiaming Chu, Lei Jin, Junliang Xing et al.

This work studies the multi-human parsing problem. Existing methods, either following top-down or bottom-up two-stage paradigms, usually involve expensive computational costs. We instead present a high-performance Single-stage Multi-human Parsing (SMP) deep architecture that decouples the multi-human parsing problem into two fine-grained sub-problems, i.e., locating the human body and parts. SMP leverages the point features in the barycenter positions to obtain their segmentation and then generates a series of offsets from the barycenter of the human body to the barycenters of parts, thus performing human body and parts matching without the grouping process. Within the SMP architecture, we propose a Refined Feature Retain module to extract the global feature of instances through generated mask attention and a Mask of Interest Reclassify module as a trainable plug-in module to refine the classification results with the predicted segmentation. Extensive experiments on the MHPv2.0 dataset demonstrate the best effectiveness and efficiency of the proposed method, surpassing the state-of-the-art method by 2.1% in AP50p, 1.0% in APvolp, and 1.2% in PCP50. In particular, the proposed method requires fewer training epochs and a less complex model architecture. We will release our source codes, pretrained models, and online demos to facilitate further studies.

Xiaojian Lin, Yaomin Shen, Junyuan Ma et al.

Monocular vertex-level human-scene contact prediction is a fundamental capability for interactive systems such as assistive monitoring, embodied AI, and rehabilitation analysis. In this work, we study this task jointly with single-image 3D human mesh reconstruction, using reconstructed body geometry as a scaffold for contact reasoning. Existing approaches either focus on contact prediction without sufficiently exploiting explicit 3D human priors, or emphasize pose/mesh reconstruction without directly optimizing robust vertex-level contact inference under occlusion and perceptual noise. To address this gap, we propose GraphiContact, a pose-aware framework that transfers complementary human priors from two pretrained Transformer encoders and predicts per-vertex human-scene contact on the reconstructed mesh. To improve robustness in real-world scenarios, we further introduce a Single-Image Multi-Infer Uncertainty (SIMU) training strategy with token-level adaptive routing, which simulates occlusion and noisy observations during training while preserving efficient single-branch inference at test time. Experiments on five benchmark datasets show that GraphiContact achieves consistent gains on both contact prediction and 3D human reconstruction. Our code, based on the GraphiContact method, provides comprehensive 3D human reconstruction and interaction analysis, and will be publicly available at https://github.com/Aveiro-Lin/GraphiContact.

Jiaming Chu, Lei Jin, Junliang Xing et al.

Multi-human parsing is an image segmentation task necessitating both instance-level and fine-grained category-level information. However, prior research has typically processed these two types of information through separate branches and distinct output formats, leading to inefficient and redundant frameworks. This paper introduces UniParser, which integrates instance-level and category-level representations in three key aspects: 1) we propose a unified correlation representation learning approach, allowing our network to learn instance and category features within the cosine space; 2) we unify the form of outputs of each modules as pixel-level segmentation results while supervising instance and category features using a homogeneous label accompanied by an auxiliary loss; and 3) we design a joint optimization procedure to fuse instance and category representations. By virtual of unifying instance-level and category-level output, UniParser circumvents manually designed post-processing techniques and surpasses state-of-the-art methods, achieving 49.3% AP on MHPv2.0 and 60.4% AP on CIHP. We will release our source code, pretrained models, and online demos to facilitate future studies.

Yu Shang, Yinzhou Tang, Yiding Ma et al.

World models have emerged as a central paradigm for embodied intelligence, enabling agents to predict action-conditioned future and reason about environmental dynamics. However, existing embodied world model benchmarks are still largely confined to vision-only prediction, offline embodied applications, and simulator-based evaluation, making them insufficient for assessing increasingly comprehensive world models. In this work, we introduce WorldArena 2.0, an expanded benchmark that systematically broadens embodied world model evaluation along three dimensions: modality, functionality, and platform. Along the modality dimension, WorldArena 2.0 extends evaluation from vision-only to visuotactile modalities, enabling assessment of multimodal perception and prediction. Along the functionality dimension, it extends beyond policy evaluation and planning to assess world models as interactive RL environments for policy optimization. Along the platform dimension, it moves beyond simulator-only evaluation to a diverse suite of simulated and real-world robotic settings across multiple embodiments. Under a standardized protocol, WorldArena 2.0 comprehensively evaluates perceptual quality, interactive utility, and cross-platform performance, providing a comprehensive testbed for tracking progress toward embodied world models. The benchmark is available at: https://world-arena.ai.

Yu Shang, Xin Zhang, Yinzhou Tang et al.

World models have become indispensable tools for embodied intelligence, serving as powerful simulators capable of generating realistic robotic videos while addressing critical data scarcity challenges. However, current embodied world models exhibit limited physical awareness, particularly in modeling 3D geometry and motion dynamics, resulting in unrealistic video generation for contact-rich robotic scenarios. In this paper, we present RoboScape, a unified physics-informed world model that jointly learns RGB video generation and physics knowledge within an integrated framework. We introduce two key physics-informed joint training tasks: temporal depth prediction that enhances 3D geometric consistency in video rendering, and keypoint dynamics learning that implicitly encodes physical properties (e.g., object shape and material characteristics) while improving complex motion modeling. Extensive experiments demonstrate that RoboScape generates videos with superior visual fidelity and physical plausibility across diverse robotic scenarios. We further validate its practical utility through downstream applications including robotic policy training with generated data and policy evaluation. Our work provides new insights for building efficient physics-informed world models to advance embodied intelligence research. The code is available at: https://github.com/tsinghua-fib-lab/RoboScape.

Kejun Ren, Lei Jin, Tianxin Huang et al.

Streaming 3D reconstruction under a strict constant-memory budget hinges on how the recurrent state is updated as the stream evolves. We profile TTT3R-style per-token gates across five benchmarks and discover a structural bottleneck: the gate is intrinsically bounded in magnitude (median $0.31$; never exceeding $0.6$) and nearly frame-invariant, yielding an effective memory horizon of only $\sim$3 frames per state token, which serves as the structural origin of long-sequence drift. We trace this to a missing axis: existing inference-time methods modulate updates only at the per-token, intra-frame level, while the orthogonal frame-level question of \emph{how strongly each frame should contribute to the state} has been treated as content-independent. We close this gap with a scalar frame-level gate $α_t \in (0, 1]$ derived in closed form from frame-to-frame changes of internal features -- a continuous relaxation of classical Simultaneous Localization and Mapping (SLAM) keyframe selection that requires no parameters, no training, and no extra forward pass. Across six benchmarks spanning camera pose, video depth, and 3D reconstruction at sequence lengths up to $4,541$ frames, our gate cuts ATE by $51\%$ on long TUM-RGBD pose sequences, reduces AbsRel by $12.8\%$ on Bonn video depth, and on KITTI long-sequence pose estimation surpasses both LongStream and Keyframe-VO, while retaining strictly constant memory at zero training cost.

Tao Wang, Zhihua Wu, Qiaozhi He et al.

Text-to-motion generation, which translates textual descriptions into human motions, has been challenging in accurately capturing detailed user-imagined motions from simple text inputs. This paper introduces StickMotion, an efficient diffusion-based network designed for multi-condition scenarios, which generates desired motions based on traditional text and our proposed stickman conditions for global and local control of these motions, respectively. We address the challenges introduced by the user-friendly stickman from three perspectives: 1) Data generation. We develop an algorithm to generate hand-drawn stickmen automatically across different dataset formats. 2) Multi-condition fusion. We propose a multi-condition module that integrates into the diffusion process and obtains outputs of all possible condition combinations, reducing computational complexity and enhancing StickMotion's performance compared to conventional approaches with the self-attention module. 3) Dynamic supervision. We empower StickMotion to make minor adjustments to the stickman's position within the output sequences, generating more natural movements through our proposed dynamic supervision strategy. Through quantitative experiments and user studies, sketching stickmen saves users about 51.5% of their time generating motions consistent with their imagination. Our codes, demos, and relevant data will be released to facilitate further research and validation within the scientific community.

Xiaojian Lin, Wenxin Zhang, Yuchu Jiang et al. · tsinghua

Hierarchical feature representations play a pivotal role in computer vision, particularly in object detection for autonomous driving. Multi-level semantic understanding is crucial for accurately identifying pedestrians, vehicles, and traffic signs in dynamic environments. However, existing architectures, such as YOLO and DETR, struggle to maintain feature consistency across different scales while balancing detection precision and computational efficiency. To address these challenges, we propose Butter, a novel object detection framework designed to enhance hierarchical feature representations for improving detection robustness. Specifically, Butter introduces two key innovations: Frequency-Adaptive Feature Consistency Enhancement (FAFCE) Component, which refines multi-scale feature consistency by leveraging adaptive frequency filtering to enhance structural and boundary precision, and Progressive Hierarchical Feature Fusion Network (PHFFNet) Module, which progressively integrates multi-level features to mitigate semantic gaps and strengthen hierarchical feature learning. Through extensive experiments on BDD100K, KITTI, and Cityscapes, Butter demonstrates superior feature representation capabilities, leading to notable improvements in detection accuracy while reducing model complexity. By focusing on hierarchical feature refinement and integration, Butter provides an advanced approach to object detection that achieves a balance between accuracy, deployability, and computational efficiency in real-time autonomous driving scenarios. Our model and implementation are publicly available at https://github.com/Aveiro-Lin/Butter, facilitating further research and validation within the autonomous driving community.

Yu Shang, Lei Jin, Yiding Ma et al.

Video-based world models hold significant potential for generating high-quality embodied manipulation data. However, current video generation methods struggle to achieve stable long-horizon generation: classical diffusion-based approaches often suffer from temporal inconsistency and visual drift over multiple rollouts, while autoregressive methods tend to compromise on visual detail. To solve this, we introduce LongScape, a hybrid framework that adaptively combines intra-chunk diffusion denoising with inter-chunk autoregressive causal generation. Our core innovation is an action-guided, variable-length chunking mechanism that partitions video based on the semantic context of robotic actions. This ensures each chunk represents a complete, coherent action, enabling the model to flexibly generate diverse dynamics. We further introduce a Context-aware Mixture-of-Experts (CMoE) framework that adaptively activates specialized experts for each chunk during generation, guaranteeing high visual quality and seamless chunk transitions. Extensive experimental results demonstrate that our method achieves stable and consistent long-horizon generation over extended rollouts. Our code is available at: https://github.com/tsinghua-fib-lab/Longscape.

Hui Dou, Lei Jin, Yuxuan Zhou et al.

The performance of modern DBMSs such as MySQL and PostgreSQL heavily depends on the configuration of performance-critical knobs. Manual tuning these knobs is laborious and inefficient due to the complex and high-dimensional nature of the configuration space. Among the automated tuning methods, reinforcement learning (RL)-based methods have recently sought to improve the DBMS knobs tuning process from several different perspectives. However, they still encounter challenges with slow convergence speed during offline training. In this paper, we mainly focus on how to leverage the valuable tuning hints contained in various textual documents such as DBMS manuals and web forums to improve the offline training of RL-based methods. To this end, we propose an efficient DBMS knobs tuning framework named DemoTuner via a novel LLM-assisted demonstration reinforcement learning method. Specifically, to comprehensively and accurately mine tuning hints from documents, we design a structured chain of thought prompt to employ LLMs to conduct a condition-aware tuning hints extraction task. To effectively integrate the mined tuning hints into RL agent training, we propose a hint-aware demonstration reinforcement learning algorithm HA-DDPGfD in DemoTuner. As far as we know, DemoTuner is the first work to introduce the demonstration reinforcement learning algorithm for DBMS knobs tuning. Experimental evaluations conducted on MySQL and PostgreSQL across various workloads demonstrate the significant advantages of DemoTuner in both performance improvement and online tuning cost reduction over three representative baselines including DB-BERT, GPTuner and CDBTune. Additionally, DemoTuner also exhibits superior adaptability to application scenarios with unknown workloads.

Yuchu Jiang, Jiaming Chu, Jian Zhao et al.

The proliferation of generative models has raised serious concerns about visual content forgery. Existing deepfake detection methods primarily target either image-level classification or pixel-wise localization. While some achieve high accuracy, they often suffer from limited generalization across manipulation types or rely on complex architectures. In this paper, we propose Loupe, a lightweight yet effective framework for joint deepfake detection and localization. Loupe integrates a patch-aware classifier and a segmentation module with conditional queries, allowing simultaneous global authenticity classification and fine-grained mask prediction. To enhance robustness against distribution shifts of test set, Loupe introduces a pseudo-label-guided test-time adaptation mechanism by leveraging patch-level predictions to supervise the segmentation head. Extensive experiments on the DDL dataset demonstrate that Loupe achieves state-of-the-art performance, securing the first place in the IJCAI 2025 Deepfake Detection and Localization Challenge with an overall score of 0.846. Our results validate the effectiveness of the proposed patch-level fusion and conditional query design in improving both classification accuracy and spatial localization under diverse forgery patterns. The code is available at https://github.com/Kamichanw/Loupe.

Zehao Yu, Lei Jin, Shenghua Gao

This paper tackles the unsupervised depth estimation task in indoor environments. The task is extremely challenging because of the vast areas of non-texture regions in these scenes. These areas could overwhelm the optimization process in the commonly used unsupervised depth estimation framework proposed for outdoor environments. However, even when those regions are masked out, the performance is still unsatisfactory. In this paper, we argue that the poor performance suffers from the non-discriminative point-based matching. To this end, we propose P$^2$Net. We first extract points with large local gradients and adopt patches centered at each point as its representation. Multiview consistency loss is then defined over patches. This operation significantly improves the robustness of the network training. Furthermore, because those textureless regions in indoor scenes (e.g., wall, floor, roof, \etc) usually correspond to planar regions, we propose to leverage superpixels as a plane prior. We enforce the predicted depth to be well fitted by a plane within each superpixel. Extensive experiments on NYUv2 and ScanNet show that our P$^2$Net outperforms existing approaches by a large margin. Code is available at \url{https://github.com/svip-lab/Indoor-SfMLearner}.

Jian Tan, Fan Bu, Yuqing Gao et al.

Machine data is central to observability and diagnosis in modern computing systems, appearing in logs, metrics, telemetry traces, and configuration snapshots. When provided to large language models (LLMs), this data typically arrives as a mixture of natural language and structured payloads such as JSON or Python/AST literals. Yet LLMs remain brittle on such inputs, particularly when they are long, deeply nested, and dominated by repetitive structure. We present HYVE (HYbrid ViEw), a framework for LLM context engineering for inputs containing large machine-data payloads, inspired by database management principles. HYVE surrounds model invocation with coordinated preprocessing and postprocessing, centered on a request-scoped datastore augmented with schema information. During preprocessing, HYVE detects repetitive structure in raw inputs, materializes it in the datastore, transforms it into hybrid columnar and row-oriented views, and selectively exposes only the most relevant representation to the LLM. During postprocessing, HYVE either returns the model output directly, queries the datastore to recover omitted information, or performs a bounded additional LLM call for SQL-augmented semantic synthesis. We evaluate HYVE on diverse real-world workloads spanning knowledge QA, chart generation, anomaly detection, and multi-step network troubleshooting. Across these benchmarks, HYVE reduces token usage by 50-90% while maintaining or improving output quality. On structured generation tasks, it improves chart-generation accuracy by up to 132% and reduces latency by up to 83%. Overall, HYVE offers a practical approximation to an effectively unbounded context window for prompts dominated by large machine-data payloads.

Chen Min, Dawei Zhao, Liang Xiao et al.

Vision-centric autonomous driving has recently raised wide attention due to its lower cost. Pre-training is essential for extracting a universal representation. However, current vision-centric pre-training typically relies on either 2D or 3D pre-text tasks, overlooking the temporal characteristics of autonomous driving as a 4D scene understanding task. In this paper, we address this challenge by introducing a world model-based autonomous driving 4D representation learning framework, dubbed \emph{DriveWorld}, which is capable of pre-training from multi-camera driving videos in a spatio-temporal fashion. Specifically, we propose a Memory State-Space Model for spatio-temporal modelling, which consists of a Dynamic Memory Bank module for learning temporal-aware latent dynamics to predict future changes and a Static Scene Propagation module for learning spatial-aware latent statics to offer comprehensive scene contexts. We additionally introduce a Task Prompt to decouple task-aware features for various downstream tasks. The experiments demonstrate that DriveWorld delivers promising results on various autonomous driving tasks. When pre-trained with the OpenScene dataset, DriveWorld achieves a 7.5% increase in mAP for 3D object detection, a 3.0% increase in IoU for online mapping, a 5.0% increase in AMOTA for multi-object tracking, a 0.1m decrease in minADE for motion forecasting, a 3.0% increase in IoU for occupancy prediction, and a 0.34m reduction in average L2 error for planning.

Jiaming Chu, Tao Wang, Lei Jin

Variational Autoencoder (VAE) encoders play a critical role in modern generative models, yet their computational cost often motivates the use of knowledge distillation or quantification to obtain compact alternatives. Existing studies typically believe that the model work better on the samples closed to their training data distribution than unseen data distribution. In this work, we report a counter-intuitive phenomenon in VAE encoder distillation: a compact encoder distilled only at low resolutions exhibits poor reconstruction performance at its native resolution, but achieves dramatically improved results when evaluated at higher, unseen input resolutions. Despite never being trained beyond $256^2$ resolution, the distilled encoder generalizes effectively to $512^2$ resolution inputs, partially inheriting the teacher model's resolution preference.We further analyze latent distributions across resolutions and find that higher-resolution inputs produce latent representations more closely aligned with the teacher's manifold. Through extensive experiments on ImageNet-256, we show that simple resolution remapping-upsampling inputs before encoding and downsampling reconstructions for evaluation-leads to substantial gains across PSNR, MSE, SSIM, LPIPS, and rFID metrics. These findings suggest that VAE encoder distillation learns resolution-consistent latent manifolds rather than resolution-specific pixel mappings. This also means that the high training cost on memory, time and high-resolution datasets are not necessary conditions for distilling a VAE with high-resolution image reconstruction capabilities. On low resolution datasets, the distillation model still could learn the detailed knowledge of the teacher model in high-resolution image reconstruction.

Xuechao Zou, Shun Zhang, Kai Li et al.

Cloud segmentation is a critical challenge in remote sensing image interpretation, as its accuracy directly impacts the effectiveness of subsequent data processing and analysis. Recently, vision foundation models (VFM) have demonstrated powerful generalization capabilities across various visual tasks. In this paper, we present a parameter-efficient adaptive approach, termed Cloud-Adapter, designed to enhance the accuracy and robustness of cloud segmentation. Our method leverages a VFM pretrained on general domain data, which remains frozen, eliminating the need for additional training. Cloud-Adapter incorporates a lightweight spatial perception module that initially utilizes a convolutional neural network (ConvNet) to extract dense spatial representations. These multi-scale features are then aggregated and serve as contextual inputs to an adapting module, which modulates the frozen transformer layers within the VFM. Experimental results demonstrate that the Cloud-Adapter approach, utilizing only 0.6% of the trainable parameters of the frozen backbone, achieves substantial performance gains. Cloud-Adapter consistently achieves state-of-the-art performance across various cloud segmentation datasets from multiple satellite sources, sensor series, data processing levels, land cover scenarios, and annotation granularities. We have released the code and model checkpoints at https://xavierjiezou.github.io/Cloud-Adapter/ to support further research.

Chuni Liu, Hongjie Li, Jiaqi Du et al.

The pretraining-finetuning paradigm is a crucial strategy in metallic surface defect detection for mitigating the challenges posed by data scarcity. However, its implementation presents a critical dilemma. Pretraining on natural image datasets such as ImageNet, faces a significant domain gap. Meanwhile, naive self-supervised pretraining on in-domain industrial data is often ineffective due to the inability of existing learning objectives to distinguish subtle defect patterns from complex background noise and textures. To resolve this, we introduce Anomaly-Guided Self-Supervised Pretraining (AGSSP), a novel paradigm that explicitly guides representation learning through anomaly priors. AGSSP employs a two-stage framework: (1) it first pretrains the model's backbone by distilling knowledge from anomaly maps, encouraging the network to capture defect-salient features; (2) it then pretrains the detector using pseudo-defect boxes derived from these maps, aligning it with localization tasks. To enable this, we develop a knowledge-enhanced method to generate high-quality anomaly maps and collect a large-scale industrial dataset of 120,000 images. Additionally, we present two small-scale, pixel-level labeled metallic surface defect datasets for validation. Extensive experiments demonstrate that AGSSP consistently enhances performance across various settings, achieving up to a 10\% improvement in mAP@0.5 and 11.4\% in mAP@0.5:0.95 compared to ImageNet-based models. All code, pretrained models, and datasets are publicly available at https://clovermini.github.io/AGSSP-Dev/.

Xin Zhang, Jiaming Chu, Jian Zhao et al.

Deepfake detection is a critical task in identifying manipulated multimedia content. In real-world scenarios, deepfake content can manifest across multiple modalities, including audio and video. To address this challenge, we present ERF-BA-TFD+, a novel multimodal deepfake detection model that combines enhanced receptive field (ERF) and audio-visual fusion. Our model processes both audio and video features simultaneously, leveraging their complementary information to improve detection accuracy and robustness. The key innovation of ERF-BA-TFD+ lies in its ability to model long-range dependencies within the audio-visual input, allowing it to better capture subtle discrepancies between real and fake content. In our experiments, we evaluate ERF-BA-TFD+ on the DDL-AV dataset, which consists of both segmented and full-length video clips. Unlike previous benchmarks, which focused primarily on isolated segments, the DDL-AV dataset allows us to assess the model's performance in a more comprehensive and realistic setting. Our method achieves state-of-the-art results on this dataset, outperforming existing techniques in terms of both accuracy and processing speed. The ERF-BA-TFD+ model demonstrated its effectiveness in the "Workshop on Deepfake Detection, Localization, and Interpretability," Track 2: Audio-Visual Detection and Localization (DDL-AV), and won first place in this competition.

Xiaowei Yuan, Lei Jin, Haoxin Zhang et al.

Retrieval-augmented generation (RAG) plays a critical role in user-generated content (UGC) platforms, but its effectiveness depends heavily on accurate relevance assessment of query-document pairs. Despite recent advances in applying large language models (LLMs) to relevance modeling, UGC platforms present unique challenges: 1) ambiguous user intent due to sparse user feedback in RAG scenarios, and 2) substantial noise introduced by informal and unstructured language. To address these issues, we propose the Reinforced Reasoning Model for Relevance Assessment (R3A), which introduces a decomposed reasoning framework over queries and candidate documents before scoring. R3A first leverages auxiliary high-ranked documents within the platform to infer latent query intent. It then performs verbatim fragment extraction to justify relevance decisions, thereby reducing errors caused by noisy UGC. Based on a reinforcement learning framework, R3A is optimized to mitigate distortions arising from ambiguous queries and unstructured content. Experimental results show that R3A significantly outperforms existing baseline methods in terms of relevance accuracy, across both offline benchmarks and online experiments.

Jiaming Chu, Lei Jin, Tao Wang et al.

The rapid development of image generation and editing algorithms in recent years has enabled ordinary user to produce realistic images. However, the current AI painting ecosystem predominantly relies on text-driven diffusion models (T2I), which pose challenges in accurately capturing user requirements. Furthermore, achieving compatibility with other modalities incurs substantial training costs. To this end, we introduce DiffBrush, which is compatible with T2I models and allows users to draw and edit images. By manipulating and adapting the internal representation of the diffusion model, DiffBrush guides the model-generated images to converge towards the user's hand-drawn sketches for user's specific needs without additional training. DiffBrush achieves control over the color, semantic, and instance of objects in images by continuously guiding the latent and instance-level attention map during the denoising process of the diffusion model. Besides, we propose a latent regeneration, which refines the randomly sampled noise in the diffusion model, obtaining a better image generation layout. Finally, users only need to roughly draw the mask of the instance (acceptable colors) on the canvas, DiffBrush can naturally generate the corresponding instance at the corresponding location.

Jian Zhao, Jianan Li, Lei Jin et al.

The 3rd Anti-UAV Workshop & Challenge aims to encourage research in developing novel and accurate methods for multi-scale object tracking. The Anti-UAV dataset used for the Anti-UAV Challenge has been publicly released. There are two main differences between this year's competition and the previous two. First, we have expanded the existing dataset, and for the first time, released a training set so that participants can focus on improving their models. Second, we set up two tracks for the first time, i.e., Anti-UAV Tracking and Anti-UAV Detection & Tracking. Around 76 participating teams from the globe competed in the 3rd Anti-UAV Challenge. In this paper, we provide a brief summary of the 3rd Anti-UAV Workshop & Challenge including brief introductions to the top three methods in each track. The submission leaderboard will be reopened for researchers that are interested in the Anti-UAV challenge. The benchmark dataset and other information can be found at: https://anti-uav.github.io/.

Yabo Xiao, Dongdong Yu, Xiaojuan Wang et al.

Off-the-shelf single-stage multi-person pose regression methods generally leverage the instance score (i.e., confidence of the instance localization) to indicate the pose quality for selecting the pose candidates. We consider that there are two gaps involved in existing paradigm:~1) The instance score is not well interrelated with the pose regression quality.~2) The instance feature representation, which is used for predicting the instance score, does not explicitly encode the structural pose information to predict the reasonable score that represents pose regression quality. To address the aforementioned issues, we propose to learn the pose regression quality-aware representation. Concretely, for the first gap, instead of using the previous instance confidence label (e.g., discrete {1,0} or Gaussian representation) to denote the position and confidence for person instance, we firstly introduce the Consistent Instance Representation (CIR) that unifies the pose regression quality score of instance and the confidence of background into a pixel-wise score map to calibrates the inconsistency between instance score and pose regression quality. To fill the second gap, we further present the Query Encoding Module (QEM) including the Keypoint Query Encoding (KQE) to encode the positional and semantic information for each keypoint and the Pose Query Encoding (PQE) which explicitly encodes the predicted structural pose information to better fit the Consistent Instance Representation (CIR). By using the proposed components, we significantly alleviate the above gaps. Our method outperforms previous single-stage regression-based even bottom-up methods and achieves the state-of-the-art result of 71.7 AP on MS COCO test-dev set.

Jian Zhao, Gang Wang, Jianan Li et al.

The 2nd Anti-UAV Workshop \& Challenge aims to encourage research in developing novel and accurate methods for multi-scale object tracking. The Anti-UAV dataset used for the Anti-UAV Challenge has been publicly released. There are two subsets in the dataset, $i.e.$, the test-dev subset and test-challenge subset. Both subsets consist of 140 thermal infrared video sequences, spanning multiple occurrences of multi-scale UAVs. Around 24 participating teams from the globe competed in the 2nd Anti-UAV Challenge. In this paper, we provide a brief summary of the 2nd Anti-UAV Workshop \& Challenge including brief introductions to the top three methods.The submission leaderboard will be reopened for researchers that are interested in the Anti-UAV challenge. The benchmark dataset and other information can be found at: https://anti-uav.github.io/.

Bo Zhang, Yuqi Cui, Meng Wang et al.

Tens of millions of women suffer from infertility worldwide each year. In vitro fertilization (IVF) is the best choice for many such patients. However, IVF is expensive, time-consuming, and both physically and emotionally demanding. The first question that a patient usually asks before the IVF is how likely she will conceive, given her basic medical examination information. This paper proposes three approaches to predict the cumulative pregnancy rate after multiple oocyte pickup cycles. Experiments on 11,190 patients showed that first clustering the patients into different groups and then building a support vector machine model for each group can achieve the best overall performance. Our model could be a quick and economic approach for reliably estimating the cumulative pregnancy rate for a patient, given only her basic medical examination information, well before starting the actual IVF procedure. The predictions can help the patient make optimal decisions on whether to use her own oocyte or donor oocyte, how many oocyte pickup cycles she may need, whether to use embryo frozen, etc. They will also reduce the patient's cost and time to pregnancy, and improve her quality of life.

Zihan Liu, Bo Huang, Yuqi Cui et al.

Time-lapse is a technology used to record the development of embryos during in-vitro fertilization (IVF). Accurate classification of embryo early development stages can provide embryologists valuable information for assessing the embryo quality, and hence is critical to the success of IVF. This paper proposes a multi-task deep learning with dynamic programming (MTDL-DP) approach for this purpose. It first uses MTDL to pre-classify each frame in the time-lapse video to an embryo development stage, and then DP to optimize the stage sequence so that the stage number is monotonically non-decreasing, which usually holds in practice. Different MTDL frameworks, e.g., one-to-many, many-to-one, and many-to-many, are investigated. It is shown that the one-to-many MTDL framework achieved the best compromise between performance and computational cost. To our knowledge, this is the first study that applies MTDL to embryo early development stage classification from time-lapse videos.

Lei Jin, Xuelian Long, James Joshi

In an undirected social graph, a friendship link involves two users and the friendship is visible in both the users' friend lists. Such a dual visibility of the friendship may raise privacy threats. This is because both users can separately control the visibility of a friendship link to other users and their privacy policies for the link may not be consistent. Even if one of them conceals the link from a third user, the third user may find such a friendship link from another user's friend list. In addition, as most users allow their friends to see their friend lists in most social network systems, an adversary can exploit the inconsistent policies to launch privacy attacks to identify and infer many of a targeted user's friends. In this paper, we propose, analyze and evaluate such an attack which is called Friendship Identification and Inference (FII) attack. In a FII attack scenario, we assume that an adversary can only see his friend list and the friend lists of his friends who do not hide the friend lists from him. Then, a FII attack contains two attack steps: 1) friend identification and 2) friend inference. In the friend identification step, the adversary tries to identify a target's friends based on his friend list and those of his friends. In the friend inference step, the adversary attempts to infer the target's friends by using the proposed random walk with restart approach. We present experimental results using three real social network datasets and show that FII attacks are generally efficient and effective when adversaries and targets are friends or 2-distant neighbors. We also comprehensively analyze the attack results in order to find what values of parameters and network features could promote FII attacks. Currently, most popular social network systems with an undirected friendship graph, such as Facebook, LinkedIn and Foursquare, are susceptible to FII attacks.