FluxMem: Adaptive Hierarchical Memory for Streaming Video Understanding
This addresses the challenge of real-time video processing for applications requiring low latency and memory usage, though it appears incremental as it builds on existing compression techniques.
The paper tackles the problem of efficient streaming video understanding by proposing FluxMem, a training-free framework that adaptively compresses redundant visual memory through hierarchical modules. The result shows state-of-the-art performance on benchmarks like 76.4 on StreamingBench and 67.2 on OVO-Bench, while reducing latency by 69.9% and GPU memory by 34.5%.
This paper presents FluxMem, a training-free framework for efficient streaming video understanding. FluxMem adaptively compresses redundant visual memory through a hierarchical, two-stage design: (1) a Temporal Adjacency Selection (TAS) module removes redundant visual tokens across adjacent frames, and (2) a Spatial Domain Consolidation (SDC) module further merges spatially repetitive regions within each frame into compact representations. To adapt effectively to dynamic scenes, we introduce a self-adaptive token compression mechanism in both TAS and SDC, which automatically determines the compression rate based on intrinsic scene statistics rather than manual tuning. Extensive experiments demonstrate that FluxMem achieves new state-of-the-art results on existing online video benchmarks, reaching 76.4 on StreamingBench and 67.2 on OVO-Bench under real-time settings, while reducing latency by 69.9% and peak GPU memory by 34.5% on OVO-Bench. Furthermore, it maintains strong offline performance, achieving 73.1 on MLVU while using 65% fewer visual tokens.