1 paper · 1 filter
Runmin Cong, Qi Qin, Chen Zhang +4
Fully-supervised salient object detection (SOD) methods have made great progress, but such methods often rely on a large number of pixel-level annotations, which are time-consuming…