[2603.26737] Beyond Static Visual Tokens: Structured Sequential Visual Chain-of-Thought Reasoning

Computer Science > Computer Vision and Pattern Recognition arXiv:2603.26737 (cs) [Submitted on 21 Mar 2026] Title:Beyond Static Visual Tokens: Structured Sequential Visual Chain-of-Thought Reasoning Authors:Guangfu Guo, Xiaoqian Lu, Yue Feng, Mingming Sun View a PDF of the paper titled Beyond Static Visual Tokens: Structured Sequential Visual Chain-of-Thought Reasoning, by Guangfu Guo and Xiaoqian Lu and Yue Feng and Mingming Sun View PDF HTML (experimental) Abstract:Current multimodal LLMs encode images as static visual prefixes and rely on text-based reasoning, lacking goal-driven and adaptive visual access. Inspired by human visual perception-where attention is selectively and sequentially shifted from the most informative regions to secondary cues-we propose Structural Sequential Visual CoT SSV-CoT. First, a question-relevant saliency map identifies and organizes key visual regions, explicitly modeling the spatial distribution of visual importance. Second, reasoning is performed following this discriminative order, inducing a curriculum-like semantic progression from primary to secondary cues. This method is trained end-to-end, using text cot and answer supervision, without relying on region-level annotations or specialized external tools. Experiments on diverse visual reasoning benchmarks show gains, validating structured and sequential visual cognition. Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI) Cite as: arXiv:2603.2673...