2
views
0
recommends
+1 Recommend
0 collections
    0
    shares
      • Record: found
      • Abstract: found
      • Article: found
      Is Open Access

      ReTaKe: Reducing Temporal and Knowledge Redundancy for Long Video Understanding

      Preprint
      , , , , ,

      Read this article at

      Bookmark
          There is no author summary for this article yet. Authors can add summaries to their articles on ScienceOpen to make them more accessible to a non-specialist audience.

          Abstract

          Video Large Language Models (VideoLLMs) have achieved remarkable progress in video understanding. However, existing VideoLLMs often inherit the limitations of their backbone LLMs in handling long sequences, leading to challenges for long video understanding. Common solutions either simply uniformly sample videos' frames or compress visual tokens, which focus primarily on low-level temporal visual redundancy, overlooking high-level knowledge redundancy. This limits the achievable compression rate with minimal loss. To this end. we introduce a training-free method, ReTaKe, containing two novel modules DPSelect and PivotKV, to jointly model and reduce both temporal visual redundancy and knowledge redundancy for long video understanding. Specifically, DPSelect identifies keyframes with local maximum peak distance based on their visual features, which are closely aligned with human video perception. PivotKV employs the obtained keyframes as pivots and conducts KV-Cache compression for the non-pivot tokens with low attention scores, which are derived from the learned prior knowledge of LLMs. Experiments on benchmarks VideoMME, MLVU, and LVBench, show that ReTaKe can support 4x longer video sequences with minimal performance loss (<1%) and outperform all similar-size VideoLLMs with 3%-5%, even surpassing or on par with much larger ones. Our code is available at https://github.com/SCZwangxiao/video-ReTaKe

          Related collections

          Author and article information

          Journal
          29 December 2024
          Article
          2412.20504
          bdad0b4c-a622-4991-b907-8c89763cae3d

          http://creativecommons.org/licenses/by/4.0/

          History
          Custom metadata
          cs.CV cs.CL cs.MM

          Computer vision & Pattern recognition,Theoretical computer science,Graphics & Multimedia design

          Comments

          Comment on this article