Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

UniComp: Rethinking Video Compression Through Informational Uniqueness

About

Distinct from attention-based compression methods, this paper presents an information uniqueness driven video compression framework, termed UniComp, which aims to maximize the information fidelity of video representations under constrained computational budgets. Starting from the information-theoretic perspective, we formulate the vision compression as an optimization problem that minimizes conditional entropy (reconstruction error) between retained and full tokens. To achieve this, we introduce the notion of information uniqueness to measure intrinsic redundancy among tokens to link with reconstruction error. Based on uniqueness, we design three modules-Frame Group Fusion, Token Allocation, and Spatial Dynamic Compression-that progressively perform semantic frame grouping, adaptive resource allocation, and fine-grained spatial compression. Extensive experiments demonstrate that UniComp consistently outperforms existing compression methods in preserving essential visual tokens under limited computational budgets, highlighting the pivotal role of information uniqueness in token compression efficacy.

Chao Yuan, Shimin Chen, Minliang Lin, Limeng Qiao, Guanglu Wan, Lin Ma• 2025

Related benchmarks

TaskDatasetResultRank
Video Question AnsweringVideoMME
Accuracy61
251
Video Question AnsweringMLVU
Accuracy61.6
194
Video Question AnsweringMVBench
Accuracy65.4
69
Video Temporal GroundingActivityNet
mIoU24.5
25
Video Question AnsweringLVB
Accuracy58.1
25
Video Question AnsweringMVBench, VideoMME, LVB, MLVU v1 (test val)
MVBench Score58.3
22
Video Temporal GroundingCharades
mIoU35
21
Video Question AnsweringVideoMME
Score60.2
16
Showing 8 of 8 rows

Other info

Follow for update