GoDeep is an annotation‑free method for open‑vocabulary 3D scene understanding that uses a vision‑language model solely as a translator to generate structured, entity‑level descriptions of each image. These descriptions are projected and aggregated in a language‑only embedding space, eliminating the need for a 3D training corpus or domain‑specific encoder. The approach achieves competitive performance on ScanNet++ and a cultural heritage benchmark, accurately localizes out‑of‑vocabulary objects, and offers explainable, point‑level predictions.
By Thodoris Betsas, Anastasios Doulamis, Andreas Georgopoulos
The paper introduces a training‑free, target‑aware refinement method for 3D Gaussian language maps distilled from vision‑language features. By blending the visual embedding within user‑specified target regions with the corresponding text embedding using a convex combination, the method improves target localization while preserving open‑vocabulary queryability. The authors derive the optimal blending coefficient analytically, avoid hyperparameters, and demonstrate consistent gains across three benchmarks, including a real‑world robot navigation demo.
By Sibaek Lee, Kyeongsu Kang, Seongbo Ha, Yewon Jeon, Joonyeol Choi, Daegeol Ko, Jaehyuk Hur, Taemin Choi, Minsu Kim, Hyungjoo Ahn, Wonsok Yoo, Hyeonwoo Yu
The paper introduces Visual Retrieval Heads (VRHs), a small fraction of attention heads in vision‑language models that are causally responsible for grounding text descriptions to image regions. By recasting head‑scoring methods and evaluating across eleven VLMs and five benchmarks, the authors show that masking the top 20 VRHs can drop grounding accuracy by up to 80 percentage points, while random masking has little effect. VRHs generalize across various visual reference tasks, preserve output format while corrupting localization, and transfer causally across models sharing an LLM backbone.
By Chanho Park, Daehyeon Choi, Jihyun Lee, Minhyuk Sung
arXiv:2611.02356v2 Announce Type: replace-cross
Abstract: Cross-view object correspondence (CVOC) aims to establish associations of the same target between query and search videos captured from diffe...
By Bing Fan, Yunhe Feng, Yan Huang, Heng Fan
arXiv:2605.19410v2 Announce Type: replace
Abstract: Segmentation has become easy when the concept is known, requiring retrieval of a learned visual grounding from text. It remains hard for open ad-ho...
By Zilin Wang, Stella X. Yu
arXiv:2508.11256v2 Announce Type: replace-cross
Abstract: Dense visual perception tasks have been constrained by their reliance on predefined categories, limiting their applicability in real-world sc...
By Junjie Wang, Keyu Chen, Yulin Li, Bin Chen, Hengshuang Zhao, Xiaojuan Qi, Zhuotao Tian
STRAND is a new benchmark that tests multimodal large language models’ ability to track objects, their states, and relationships over time in videos. It evaluates intermediate reasoning by breaking queries into sub‑questions and uses Faithful Accuracy to ensure all parts of an answer are correct. The authors also propose an object‑centric framework that builds structured trajectories and shows reduced hallucinations and better temporal consistency compared to existing models.
By Thong Nguyen, Tri Cao, Khoi Le, Cong-Duy Nguyen, Quynh Vo, See-Kiong Ng, Bryan Hooi Kuen-Yew
arXiv:2606. 19733v1 Announce Type: cross Abstract: Efficiently retrieving specific 3D instances from large-scale scenes via natural language prompts remains a formidable challenge in multimedia analysis.
By Xiuyuan Zhu, Ke Lu, Zijie Yang, Chao Yue, Jian Xue, Dongming Zhang
The paper introduces MVLGeo, a unified framework for cross-view object geo-localization that combines multiple viewpoints into a single model. It employs Vision‑Language Reranking to use contextual text from the query view, a multi‑view Mixture‑of‑Experts architecture to share knowledge and reduce redundancy, and an adaptive elliptical prior for positional encoding. Experiments on CVOGL benchmarks show that MVLGeo achieves state‑of‑the‑art performance and robustness to input degradation.
By Xuyu Fan, Qi Ming, Zhu Han, Liuqian Wang, Siyuan Cao, Xiaohan Zhang, Xudong Zhao, Mingjing Zhao, Yuhan Zhang
arXiv:2601. 06550v3 Announce Type: replace-cross Abstract: Semantic Multi-Object Tracking (SMOT) is evolving from purely geometric localization toward comprehensive video understanding.
By Pan Liao, Feng Yang, Di Wu, Jinwen Yu, Wang Zhao, Dingwen Zhang
SplatScout is a new query‑driven, closed‑loop approach for part‑level referring segmentation in Gaussian scenes, placing perception under vision‑language model control. It actively selects views, directs segmentation, and refines predictions by inspecting rendered 3D results. On the newly introduced PartRef‑ScanNet++ benchmark, SplatScout achieves 0.538 mIoU, outperforming existing methods and demonstrating the benefits of iterative refinement and effective tool‑feedback utilization.
By Yihan Zang, Da Li, Shinkyu Park, Ivan Viola, Dominik Engel
arXiv:2606. 00987v1 Announce Type: cross Abstract: Large Vision-Language Models (LVLMs) have shown strong visual understanding and language-guided grounding abilities, yet their capacity for multi-temporal visual reasoning remains underexplored.
By Bingyu Li, Da Zhang, Tao Huo, Zhiyuan Zhao, Junyu Gao, Xuelong Li