ER-SAN: Enhanced-Adaptive Relation Self-Attention Network for Image Captioning.

Jingyu Li, Zhendong Mao, Shancheng Fang, Hao Li

Browse the full IJCAI paper archive.