Deep hashing for large-scale image retrieval

Abstract

In recent years, because of its ability to improve the retrieval efficiency and reduce the storage space, the image hashing methods based on deep learning have got widely attention in large-scale image retrieval system. At first, this paper introduces the development of the deep learning network, and analyzes the advantages and disadvantages of the deep hashing methods combined with convolutional network and hashing algorithm in current trend. And then it summarizes the research status of the traditional hashing, compares the deep hashing methods and the traditional hashing, points out the great potential of the large-scale image retrieval technology based on the deep hashing. Finally, combined with the existing hashing learning framework based on deep convolution neural network, we put forward some suggestions.