Improving Face Recognition from Caption Supervision with Multi-Granular Contextual Feature Aggregation | IEEE Conference Publication | IEEE Xplore