Cross-modal face matching: Tackling visual abstraction using fine-grained attributes | IEEE Conference Publication | IEEE Xplore