Deep Pairwise-supervised Hashing

With the explosive growing of data in real applications like image retrieval,ap-proximate nearest neighbor(ANN)search has become a hot research topic in recent years.Among existing ANN techniques,hashing has become one of the most popular and effective techniques due to its fast query speed and low memory cost,especially in image retrieval applications.However,most existing hashing methods are based on hand-crafted features which might not be optimally compatible with the hashing pro-cedure.Recently,deep hashing methods have been proposed to perform simultaneous feature learning and hash-code learning with deep neural networks,which have shown better performance than traditional hashing methods with hand-crafted features.Most of these deep hashing methods are supervised whose supervised information is given with triplet labels.For another common application scenario with pairwise labels,there have not existed methods for simultaneous feature learning and hash-code learning.In this thesis,I propose a novel deep hashing method,called deep pairwise-supervised hashing(DPSH),to perform simultaneous feature learning and hash-code learning for applications with pairwise labels.Experiments on real datasets show that our DPSH method can outperform other methods to achieve the state-of-the-art performance in image retrieval applications.Although our DPSH method has achieved promising performance in image re-trieval applications,the single-GPU implementation is time-consuming for large-scale datasets.So we design a multi-GPU implementation to parallelize the DPSH method,which achieves a good performance in time reduction without influencing the accuracy.