MHSAN: Multi-Head Self-Attention Network for Visual Semantic Embedding | IEEE Conference Publication | IEEE Xplore