10 papers
Temporally Grounded Compositional Camera Motion Understanding via Geometric Knowledge Distillation
Dazhao Du, Shiyan Du, Jian Liu +8
Understanding camera motion is fundamental to video perception, with applications in spatial intelligence and controllable video generation. Multimodal large language models (MLLMs…
polyDAG: Polynomial Acyclicity Constraints for Efficient Continuous Causal Discovery in Visual Semantic Graphs
Wenhao Zhang, Ramin Ramezani, Tao Han +2
Modern image-analysis pipelines often convert images into structured semantic variables, such as facial attributes, object concepts, and scene descriptors. Learning directed depend…
AdaTok: Self-Budgeting Image Tokenization with Quality-Preserving Dynamic Tokens
Xiaocheng Lu, Yuxi Chen, Jie Zhang +5
Image tokenizers, from 2D grids to recent 1D sequences, typically encode every image with the same fixed number of tokens. Yet visual complexity is highly heterogeneous, so a unifo…
Train the Agent, Not the Expert: Learning to Harness Heterogeneous Experts for Multi-Turn Visual Reasoning
Yaowu Fan, Tao Han, Dazhao Du +2
Recent progress in computer vision has produced a wide range of powerful specialized models for detection, segmentation, counting, and other visual tasks. However, these models are…
Learning Spatiotemporal Sensitivity in Video LLMs via Counterfactual Reinforcement Learning
Dazhao Du, Jian Liu, Jialong Qin +7
Video large language models (Video LLMs) can achieve strong video-QA accuracy without reliably tracking spatiotemporal dynamics. A model may answer a motion question from static cu…
MLLMs Know When Before Speaking: Revealing and Recovering Temporal Grounding via Attention Cues
Dazhao Du, Liao Duan, Jian Liu +5
Video temporal grounding (VTG), which localizes the start and end times of a queried event in an untrimmed video, is a key test of whether multimodal large language models (MLLMs)…
What You Think is What You See: Driving Exploration in VLM Agents via Visual-Linguistic Curiosity
Haoxi Li, Qinglin Hou, Jianfei Ma +6
To navigate partially observable visual environments, recent VLM agents increasingly internalize world modeling capabilities into their policies via explicit CoT reasoning, enablin…
UNICBench: UNIfied Counting Benchmark for MLLM
Chenggang Rong, Tao Han, Zhiyuan Zhao +5
Counting is a core capability for multimodal large language models (MLLMs), yet there is no unified counting dataset to rigorously evaluate this ability across image, text, and aud…
FreeMesh: Boosting Mesh Generation with Coordinates Merging
Jian Liu, Haohan Weng, Biwen Lei +6
The next-coordinate prediction paradigm has emerged as the de facto standard in current auto-regressive mesh generation methods. Despite their effectiveness, there is no efficient…
Mesh-RFT: Enhancing Mesh Generation via Fine-grained Reinforcement Fine-Tuning
Jian Liu, Jing Xu, Song Guo +10
Existing pretrained models for 3D mesh generation often suffer from data biases and produce low-quality results, while global reinforcement learning (RL) methods rely on object-lev…