13 papers
Beyond Accuracy: Evaluating Grounded Visual Evidence in Thinking with Images
Xuchen Li, Xuzhao Li, Renjie Pi +3
Despite the remarkable progress of Vision-Language Models (VLMs) in adopting "Thinking-with-Images" capabilities, accurately evaluating the authenticity of their reasoning process…
MATrack: Efficient Multiscale Adaptive Tracker for Real-Time Nighttime UAV Operations
Xuzhao Li, Xuchen Li, Shiyu Hu
Nighttime UAV tracking faces significant challenges in real-world robotics operations. Low-light conditions not only limit visual perception capabilities, but cluttered backgrounds…
Select Less, Reason More: Prioritizing Evidence Purity for Video Reasoning
Xuchen Li, Xuzhao Li, Shiyu Hu +1
Long-form video reasoning remains a major challenge for Video Large Language Models (Video LLMs), as static uniform frame sampling leads to information dilution and obscures critic…
Look Less, Reason More: Rollout-Guided Adaptive Pixel-Space Reasoning
Xuchen Li, Xuzhao Li, Jiahui Gao +3
Vision-Language Models (VLMs) excel at many multimodal tasks, yet they frequently struggle with tasks requiring precise understanding and handling of fine-grained visual elements.…
VS-LLM: Visual-Semantic Depression Assessment based on LLM for Drawing Projection Test
Meiqi Wu, Yaxuan Kang, Xuchen Li +5
The Drawing Projection Test (DPT) is an essential tool in art therapy, allowing psychologists to assess participants' mental states through their sketches. Specifically, through sk…
ATCTrack: Aligning Target-Context Cues with Dynamic Target States for Robust Vision-Language Tracking
X. Feng, S. Hu, X. Li +5
Vision-language tracking aims to locate the target object in the video sequence using a template patch and a language description provided in the initial frame. To achieve robust t…