1 paper
Jun Fu, Jing Liu, Haijie Tian +4
In this paper, we address the scene segmentation task by capturing rich contextual dependencies based on the selfattention mechanism. Unlike previous works that capture contexts by…