VideoMamba: Spatio-Temporal Selective State Space Model

TL;DR

VideoMamba introduces a resource-efficient spatio-temporal state space model for video recognition, capturing complex dependencies with linear complexity, outperforming transformer-based methods in efficiency and effectiveness.

Contribution

It presents a novel spatio-temporal SSM architecture for video recognition that is more efficient than transformers and effectively models long-range dependencies.

Findings

Achieves competitive performance on video benchmarks.

Demonstrates superior efficiency compared to transformer models.

Effectively captures long-range spatio-temporal dependencies.

Abstract

We introduce VideoMamba, a novel adaptation of the pure Mamba architecture, specifically designed for video recognition. Unlike transformers that rely on self-attention mechanisms leading to high computational costs by quadratic complexity, VideoMamba leverages Mamba's linear complexity and selective SSM mechanism for more efficient processing. The proposed Spatio-Temporal Forward and Backward SSM allows the model to effectively capture the complex relationship between non-sequential spatial and sequential temporal information in video. Consequently, VideoMamba is not only resource-efficient but also effective in capturing long-range dependency in videos, demonstrated by competitive performance and outstanding efficiency on a variety of video understanding benchmarks. Our work highlights the potential of VideoMamba as a powerful tool for video understanding, offering a simple yet effective baseline for future research in video analysis.