Lightweight Face Hash Retrieval via Knowledge Distillation and Optimal Transport

Proposing a Face Hash Retrieval Algorithm based on Knowledge Distillation and Optimal Transport Technology. To address the issues of excessive visual neural network model parameters, high computational complexity, and inaccurate fine-grained retrieval, this paper proposes a lightweight hash retrieval algorithm for face images using knowledge distillation and optimal transport technology. An important contribution of this algorithm is the design of a triplet knowledge distillation based on attention mechanism, where the loss function consists of three parts: attention loss, Kullback-Leibler loss, and identity loss. This knowledge distillation scheme can focus on the salient areas of the face and reduce network parameters. Another contribution is the design of a hash quantization scheme based on optimal transport technology. This scheme partitions the face feature space by computing class centers and utilizes optimal transport technology to achieve binary quantization, effectively improving hash retrieval performance. Additionally, an alternate training strategy is designed to fine-tune the parameters of the lightweight hash network. Experimental results show that this lightweight hash retrieval algorithm performs better than some well-known hash retrieval algorithms on two benchmark face datasets.