eduzhai > Applied Sciences > Engineering >

A survey on deep hashing for image retrieval

  • Save

... pages left unread,continue reading

Document pages: 14 pages

Abstract: Hashing has been widely used in approximate nearest search for large-scaledatabase retrieval for its computation and storage efficiency. Deep hashing,which devises convolutional neural network architecture to exploit and extractthe semantic information or feature of images, has received increasingattention recently. In this survey, several deep supervised hashing methods forimage retrieval are evaluated and I conclude three main different directionsfor deep supervised hashing methods. Several comments are made at the end.Moreover, to break through the bottleneck of the existing hashing methods, Ipropose a Shadow Recurrent Hashing(SRH) method as a try. Specifically, I devisea CNN architecture to extract the semantic features of images and design a lossfunction to encourage similar images projected close. To this end, I propose aconcept: shadow of the CNN output. During optimization process, the CNN outputand its shadow are guiding each other so as to achieve the optimal solution asmuch as possible. Several experiments on dataset CIFAR-10 show the satisfyingperformance of SRH.

Please select stars to rate!


0 comments Sign in to leave a comment.

    Data loading, please wait...