SFMViT: SlowFast Meet ViT in Chaotic World
This work addresses the problem of video understanding in chaotic environments for applications like surveillance or autonomous systems, representing an incremental improvement through hybrid methods.
The paper tackles spatiotemporal action localization in chaotic scenes by proposing SFMViT, a dual-stream network combining ViT and SlowFast with an anchor pruning strategy, achieving a mAP of 26.62% on the Chaotic World dataset, significantly outperforming existing models.
The task of spatiotemporal action localization in chaotic scenes is a challenging task toward advanced video understanding. Paving the way with high-quality video feature extraction and enhancing the precision of detector-predicted anchors can effectively improve model performance. To this end, we propose a high-performance dual-stream spatiotemporal feature extraction network SFMViT with an anchor pruning strategy. The backbone of our SFMViT is composed of ViT and SlowFast with prior knowledge of spatiotemporal action localization, which fully utilizes ViT's excellent global feature extraction capabilities and SlowFast's spatiotemporal sequence modeling capabilities. Secondly, we introduce the confidence maximum heap to prune the anchors detected in each frame of the picture to filter out the effective anchors. These designs enable our SFMViT to achieve a mAP of 26.62% in the Chaotic World dataset, far exceeding existing models. Code is available at https://github.com/jfightyr/SlowFast-Meet-ViT.