1 paper
Hengyue Pan, Xin Niu, Rongchun Li +2
In this paper, we propose a fixed-size object encoding method (FOE-VRD) to improve performance of visual relationship detection tasks. Comparing with previous methods, FOE-VRD has…