TOMM2020 Dual-Path Convolutional Image-Text Embedding :feet: https://arxiv.org/abs/1711.05535 - View it on GitHub