Hierarchical spatiotemporal Feature Interaction Network for video saliency prediction

Open Webpage

Yingjie Jin, Xiaofei Zhou, Zhenjie Zhang, Hao Fang, Ran Shi, Xiaobin Xu

Published: 01 Jan 2025, Last Modified: 08 Apr 2025Image Vis. Comput. 2025EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: Highlights•Design a hierarchical spatiotemporal feature interaction network for video saliency prediction.•Multi-scale feature integration unit is proposed to interact with spatiotemporal features.•Hierarchical enhancement module (TFE, CFE) is proposed to strengthen the fused features at different levels.