The paper introduces MisVisAgentBench, a paired benchmark with 280 task instances that compare clean versus misleading chart evidence for web agents. It evaluates how deceptive visualizations affect agent execution, providing a diagnostic framework to separate misleading-induced failures from baseline difficulty. Results show that misleading charts reduce task success by over 34 percentage points and tend to redirect agents toward targeted misleading actions.
By Yisheng Li, Dongming Han, Hongkun Pan, Bingcheng Mao, Zijian Wang, Yunlong Mao, Wei Tong, Sheng Zhong
The paper introduces a multi‑task pre‑training strategy for German traffic law that integrates Knowledge Graph (KG) information through Next Sentence Prediction, KG‑Guided Clustering, and KG‑Aware Paragraph Classification. By embedding structured legal knowledge into Transformer weights, the model can perform zero‑shot legal reasoning without explicit graph traversal. Experiments on nine encoders and 324 runs show substantial gains—up to 155% over baselines on synthetic StVO QA and improved performance on four of seven LexGLUE benchmarks, with consistent zero‑shot transfer to Austrian driving regulations.
By Ibrahim Siddig, Munir Georges
The paper introduces AVA, a training‑free framework that improves interleaved multimodal in‑context learning for frozen large vision‑language models. AVA decomposes the interleaved context into per‑sample threads, selects influential demonstrations while removing redundancy, and calibrates consistency using a weak sibling model’s confidence. Across five vision‑language benchmarks, AVA raises performance by up to 2.4 points, narrows the visual performance gap on ScienceQA‑IMG, and remains robust even with 30% noisy demonstrations, all while keeping the LVLM frozen and adding only modest latency.
By Yigit Yilmaz, Elena Petrova, Mehmet Kaya, Lucia Rossi, Amir Rahman
The paper introduces TerraAgent, a training‑free framework that enhances remote sensing visual question answering by integrating three modules: GDR for demonstration retrieval, MACE for evidence‑typed sub‑question decomposition, and CVA for counterfactual evidence auditing. It demonstrates significant improvements on multiple benchmarks, raising accuracy, visual‑dependency scores, and reducing prior‑only answers and hallucinations without fine‑tuning.
By Jingwen Huang, Yiming Fang, Hanxiao Wu, Boyuan He, Keyi Luo
MemReg introduces a multi-frame outdoor LiDAR point cloud registration network that leverages hybrid memory buffers to capture temporal correlations in streaming sequences. The system uses an implicit memory feature buffer and an explicit memory pose buffer to store and update pose-related temporal features, while a dynamic history weighting module adaptively fuses current and historical information. Experiments on KITTI, nuScenes, and Apollo-Southbay datasets show MemReg surpasses previous pair-wise methods and multi-frame SLAM systems, and it also performs competitively on indoor datasets such as 3DMatch, 3DLoMatch, and ScanNet.
By Jiuming Liu, Mengmeng Liu, Guangming Wang, Lihao Liu, Michael Ying Yang, Yunpeng Zhang, Francesco Nex, Hao Cheng, Hesheng Wang
WoodBR-Cross is a new public dataset of 4,875 wood surface images grouped into 38 taxonomic classes and linked to 303 distinct trees, with a species-level subset of 3,406 images across 20 classes and 214 trees. The benchmark introduces three tasks that compare duplicate-aware image partitions, species subset partitions, and individual-tree-disjoint partitions, ensuring no tree appears in both training and test sets within a fold. Experiments with ResNet‑50, ConvNeXt‑Tiny, and ViT‑B/16 show that ConvNeXt‑Tiny and frozen DINOv2 features achieve macro‑F1 scores of 0.8621 and 0.9000, respectively, on the tree‑level disjoint split, providing realistic estimates of performance on unseen trees.
By Daniel Casanova, Alexandre Bahia Gontijo, Arnaldo Candido Junior, Silvana Nisgoski, Pedro Luiz de Paula Filho
The paper introduces Shift-Aware Identity-guided Latent (SAIL) refinement for Referring Audio‑Visual Segmentation (Ref‑AVS), a task that requires precise segmentation of a specific object using both visual and auditory cues grounded in natural language. SAIL employs a dual‑token strategy to separate frame‑level dynamics from a global semantic anchor, a shift‑aware refinement module to correct target‑shift drift, and a decoder with continuous memory updates to preserve spatio‑temporal coherence and identity consistency. Experiments on two Ref‑AVS benchmarks show that SAIL outperforms existing AVS, Ref‑VOS, and Ref‑AVS methods.
By Kun Li, Sami Sebastian Brandt, Michael Ying Yang
The paper introduces SAMF, a Structure-Aware Matching Framework designed to improve UAV thermal-to-satellite geo‑localization by exploiting stable cross‑modal structural cues. It comprises a Directional Structure Enhancement Module that captures directional and low‑frequency spatial patterns, and a Modality‑Aware Affine Remapping Module that aligns these enhanced features across modalities. Experiments on Boson‑nighttime and Thermal‑UAV‑Paired datasets demonstrate state‑of‑the‑art performance, with ablation studies confirming the complementary benefits of the two modules.
By Chun Liu, Yuxin Wang, Zhenxiang Huo, Linlin Zhu, Tao Xu, Henggang Han
MeshBind introduces an automatic annotation pipeline that generates per-point scene flow for non‑rigid objects such as pedestrians by reconstructing articulated mesh models from six surround cameras and registering them to LiDAR data. The method employs a local‑frame binding technique that attaches each point to a mesh via a per‑triangle tangent‑normal frame, and a three‑layer flow registration pipeline that compensates ego‑vehicle motion, performs box binding, and applies mesh binding to produce full‑scene flow. MeshBind is implemented in the SF‑Studio tool, and the authors release the SUScape‑Flow Benchmark (SFB), a large‑scale dataset with 41k frames of per‑point non‑rigid pedestrian annotations and a comprehensive evaluation protocol, demonstrating that fine‑tuning on SFB reduces endpoint error by 2‑3× and equalizes non‑rigid and rigid EPE.
By Yinqi Chen, E Li, Shuaijun Wang, Weiyuan Peng, Meiying Zhang, Qi Hao
UniDeblur introduces a unified motion deblurring framework that learns blur-adaptive global filters in Fourier space, addressing the limitations of pixel-level mapping in existing methods. By constructing a multi-source dataset and implementing a data pruning and quality control pipeline, the model can handle diverse motion blur kernels across different cameras and image contents. Experiments show that UniDeblur performs competitively on both in-domain and out-of-domain benchmarks, demonstrating its effectiveness for varied blur scenarios.
By Xintian Mao, Jiansheng Wang, Huaiyuan Wang, Qingli Li, Yan Wang
The paper introduces UniGVC, a unified generative video codec that supports both low‑delay streaming and high‑quality playback from a single bitstream. It offers two decoding modes: UniGVC‑BI, a bidirectional diffusion model with dense attention for quality‑oriented random‑access decoding, and UniGVC‑AR, a windowed autoregressive diffusion model for low‑delay streaming. Both modes share a hierarchical latent codec that produces ultra‑low‑bitrate conditions, enabling the BI model to act as a teacher for distilling the AR model, while a lightweight conditional decoder further reduces streaming latency.
By Tianyu Zhang, Shuai Huo, Dengchao Jin, Hewei Liu, Meng Lei, Tianwen Wu, Jiawen Gu, Chao Zhou, Li Li, Dong Liu
UniMergeTrack introduces a hierarchical model merging strategy for RGB-X visual object tracking, combining independently trained modality-specific trackers into a unified framework. The method merges transferable backbones using high‑pass‑filtered merging, aligns tracking heads with optimal transport‑guided channel alignment, and fine‑tunes with distribution alignment to integrate cross‑task knowledge while preserving modality cues. Experiments on multiple benchmarks show state‑of‑the‑art or competitive performance across RGB‑Thermal, RGB‑Event, and RGB‑Depth tracking tasks, with strong generalization to unseen datasets.
By Shiao Wang, Xiao Wang, Yuhe Ding, Jiarui Xu, Zikang Yan, Bo Jiang, Yonghong Tian, Jin Tang
The paper identifies a new inference-time issue called FGVR Laziness in Multimodal Large Language Models (MLLMs), where visual cues are insufficiently linked to category concepts during fine‑grained visual recognition. To address this, the authors propose LIVE, a training‑free framework that refines candidate lists and applies category‑specific guidance based on recognition uncertainty, thereby enhancing MLLM performance across 13 benchmarks and 24 state‑of‑the‑art methods. LIVE demonstrates improved effectiveness and generalizability across various MLLM backbones without requiring additional training.
By Yutong Yang, Zhuqi Liu, Mouxing Yang, Xi Peng
The paper introduces STITCH, a self‑supervised compatibility measure that replaces traditional boundary‑dependent metrics for solving visual puzzles with eroded fragments. By comparing frozen ViT patch tokens from piece interiors using a lightweight cross‑attention head, STITCH can be plugged into existing solvers and achieves near‑perfect reconstruction on the JPwLEG‑5 benchmark, outperforms prior methods on a new ImageNet 7×7 benchmark, and remains robust across varying erosion levels and high missing‑piece scenarios. The authors demonstrate that improving the matching metric, rather than the solver itself, yields significant performance gains across multiple datasets.
By Nadav Alali, Ohad Ben-Shahar
The paper introduces DSG‑GI, a method that adapts pretrained latent video diffusion models for generative inbetweening by using Destination‑Conditioned Adaptation and Semantic‑Geometric Alignment. Destination‑Conditioned Adaptation treats the target frame as the sequence’s endpoint, providing progressive guidance along the generated trajectory. Semantic‑Geometric Alignment aligns token‑wise spatio‑temporal geometry of backbone features with a frozen video encoder and enforces a flow‑matching latent flow constraint near the data endpoint, thereby improving temporal stability, target‑frame faithfulness, and perceptual quality on large‑motion benchmarks.
By Zhiyu Xu (Tianjin University), Wei Shang (Singapore Management University), Yushen Xu (Tianjin University), Wenwen Li (Tianjin Medical University), Zhiyan Xu (Southeast University), Pengfei Zhu (Tianjin University), Dongwei Ren (Tianjin University)
The paper introduces a two-stage framework for detecting cartographic signs in historical maps and generating interpretable embeddings that capture both semantics and stylistic variations. It presents CaSiDD, an open benchmark dataset covering multiple sign categories and diverse cartographic styles, and demonstrates the effectiveness of YOLOv10-X for automated sign detection. The authors also develop a method based on an adapted RADIOv4 backbone to encode sign form and semantics while suppressing confounding factors, providing a foundation for scalable evidence extraction in cartographic culture and geohistorical research.
By Remi Petitpierre
Jag is a method that transforms a compact pretrained multimodal model into a direct box predictor for visual grounding. It adds a lightweight regression head that, conditioned on the image and referring expression, outputs all four box coordinates in a single forward pass without sequential decoding. Geometric supervision jointly fine‑tunes the language backbone, visual merger, and regression head, leading to improved accuracy and lower latency compared to existing specialized grounding models.
By Xiuyuan Zhu, Ke Lu, Hao Wu, Zijin Du, Dongming Zhang, Jian Xue
The paper introduces ContextReID, a framework that generates player embeddings by incorporating contextual information from co-occurring players in the same frame and from multiple frames in a broadcast sequence. Unlike traditional single-crop methods, this approach uses a vision‑language model to produce context‑aware embeddings that capture cross‑frame identity relationships. Experiments on three multi‑object tracking benchmarks demonstrate that ContextReID outperforms conventional single‑crop embeddings in Rank‑1 and mAP metrics for player re‑identification.
By Ge Luo, Changhuo Yang, Weisi Fan, Rui Yang, Shaokai Wang
ConClick introduces a contrastive learning framework that improves GUI grounding by teaching models to distinguish target interface elements from plausible distractors. The authors built ConClick-Data, a dataset of 4,292 challenging target-distractor pairs, and trained the model with listwise ranking and pairwise margin objectives, achieving significant accuracy gains on multiple benchmarks. An inference-time strategy called ReClick further boosts performance by revisiting promising regions and comparing competing predictions.
By Ruotian Zhang, Yuhui Zhang, Junzhe Huang, Yu Xin, Hehe Fan, Xiaoxiao Sun
The paper introduces ELMA, a new episodic long‑term motion dataset designed to benchmark anaphoric compositions in text‑to‑motion generation. ELMA contains densely annotated 30‑second clips extracted from in‑the‑wild videos, preserving cross‑segment dependencies that current models struggle to handle. Fine‑tuning state‑of‑the‑art models on ELMA improves both global physical realism and semantic alignment, establishing a robust baseline for long‑term motion synthesis.
By Shun Kato, Dixin Yang, Kotaro Amaya, Yuiga Wada, Kazumi Fukuda, Takashi Shibuya, Yuki Mitsufuji, Mariko Isogawa