MAFH: Multilabel aware framework for bit-scalable cross-modal hashing

Xue Li, Jiong Yu, Hongchun Lu, Shaochen Jiang, Ziyang Li, Peiyun Yao

Abstract: Highlights•Robustness and flexibility is a key challenge in cross modal retrieval Models.•Label network optimizes the use of data labels across modalities.•Coding strategy allows flexible control of hash length based on scenario demands.•Layered multilabel similarity can preserve the highly discriminative information.•The IAPR TC-12 dataset saw a 6.74% improvement in retrieval accuracy.