Triage: Hierarchical Visual Budgeting for Efficient Video Reasoning in Vision-Language Models
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866918315745607680 |
|---|---|
| author | Wang, Anmin Zhang, Nan Tao, Wei Qu, Xiaoyang Li, Guokuan Wan, Jiguang Wang, Jianzong |
| author_facet | Wang, Anmin Zhang, Nan Tao, Wei Qu, Xiaoyang Li, Guokuan Wan, Jiguang Wang, Jianzong |
| contents | Vision-Language Models (VLMs) face significant computational challenges in video processing due to massive data redundancy, which creates prohibitively long token sequences. To address this, we introduce Triage, a training-free, plug-and-play framework that reframes video reasoning as a resource allocation problem via hierarchical visual budgeting. Its first stage, Frame-Level Budgeting, identifies keyframes by evaluating their visual dynamics and relevance, generating a strategic prior based on their importance scores. Guided by this prior, the second stage, Token-Level Budgeting, allocates tokens in two phases: it first secures high-relevance Core Tokens, followed by diverse Context Tokens selected with an efficient batched Maximal Marginal Relevance (MMR) algorithm. Extensive experiments demonstrate that Triage improves inference speed and reduces memory footprint, while maintaining or surpassing the performance of baselines and other methods on various video reasoning benchmarks. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_22959 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Triage: Hierarchical Visual Budgeting for Efficient Video Reasoning in Vision-Language Models Wang, Anmin Zhang, Nan Tao, Wei Qu, Xiaoyang Li, Guokuan Wan, Jiguang Wang, Jianzong Computer Vision and Pattern Recognition Vision-Language Models (VLMs) face significant computational challenges in video processing due to massive data redundancy, which creates prohibitively long token sequences. To address this, we introduce Triage, a training-free, plug-and-play framework that reframes video reasoning as a resource allocation problem via hierarchical visual budgeting. Its first stage, Frame-Level Budgeting, identifies keyframes by evaluating their visual dynamics and relevance, generating a strategic prior based on their importance scores. Guided by this prior, the second stage, Token-Level Budgeting, allocates tokens in two phases: it first secures high-relevance Core Tokens, followed by diverse Context Tokens selected with an efficient batched Maximal Marginal Relevance (MMR) algorithm. Extensive experiments demonstrate that Triage improves inference speed and reduces memory footprint, while maintaining or surpassing the performance of baselines and other methods on various video reasoning benchmarks. |
| title | Triage: Hierarchical Visual Budgeting for Efficient Video Reasoning in Vision-Language Models |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2601.22959 |