Enhance Training Objectives for Image Captioning with Decomposed Sequence-level Metric.

Yehuan Wang, Jiefeng Long, Jian Hu, Lin Shang

Browse the full IJCNN paper archive.