Research map / Vision-language and multimodal / Multimodal LLMs
WalkVLM: Aid Visually Impaired People Walking by Vision Language Model
Zhiqiang Yuan, Ting Zhang, Yeshuang Zhu, Jiapei Zhang, Ying Deng, Zexi Jia, Peixiang Luo, Xiaoyue Duan, Jie Zhou, Jinchao Zhang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add CoT-based hierarchical planning (ablation) | WalkVLM · WAD reminder task | TF-IDF: 0.094 -> 0.166 | Improved |
| add CoT-based hierarchical planning (ablation) | WalkVLM · WAD reminder task | ROUGE-L: 0.066 -> 0.173 | Improved |
| add CoT-based hierarchical planning (ablation) | WalkVLM · WAD reminder task | ROUGE-1: 0.073 -> 0.191 | Improved |
| add approximate position prior (POLM) (ablation) | WalkVLM · WAD reminder task | TF-IDF: 0.151 -> 0.166 | Improved |
| add approximate position prior (POLM) (ablation) | WalkVLM · WAD reminder task | ROUGE-2: 0.062 -> 0.062 | No clear effect |
| add exact POLM prior (location and category) (ablation) | WalkVLM · WAD reminder task | ROUGE-L: 0.164 -> 0.173 | Improved |
| add exact POLM prior (location and category) (ablation) | WalkVLM · WAD reminder task | TF-IDF: 0.152 -> 0.166 | Improved |
| fine-tune MiniCPM-V2.6 on WAD | MiniCPM-V2.6 (8B) · WAD reminder task | GPT Score: 0.025 -> 1.024 | Improved |
| fine-tune Qwen2-VL on WAD | Qwen2-VL (7B) · WAD QA task | GPT Score: 0.504 -> 3.246 | Improved |
| fine-tune DeepSeek-VL on WAD | DeepSeek-VL (7B) · WAD reminder task | ROUGE-L: 0.068 -> 0.141 | Improved |
| add temporal-aware adaptive prediction and WalkVLM design (vs GPT-4o) | WalkVLM · WAD reminder task (temporal redundancy) | TRF: 0.430 -> 0.505 | Improved |
| WalkVLM vs fine-tuned Qwen2-VL on QA TF-IDF | WalkVLM · WAD QA task | TF-IDF: 0.246 -> 0.189 | Hurt |
Closest papers by meaning
- mmWalk: Towards Multi-modal Multi-view Walking Assistance (NeurIPS 2025 · similarity 0.89)
- Evaluating Vision-Language Models as Evaluators in Path Planning (CVPR 2025 · similarity 0.85)
- EgoBlind: Towards Egocentric Visual Assistance for the Blind (NeurIPS 2025 · similarity 0.85)
- UniBench: Visual Reasoning Requires Rethinking Vision-Language Beyond Scaling (NeurIPS 2024 · similarity 0.84)
- Text to Blind Motion (NeurIPS 2024 · similarity 0.84)
- WalkGPT: Grounded Vision-Language Conversation with Depth-Aware Segmentation for Pedestrian Navigation (CVPR 2026 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
