PROJECT TITLE :

Image Retrieval Based On Deep Convolutional Neural Networks And Binary Hashing Learning - 2017

ABSTRACT:

With the increasing quantity of image information, the image retrieval strategies have several drawbacks, such as the low expression ability of visual feature, high dimension of feature, low precision of image retrieval and so on. To solve these problems, a learning methodology of binary hashing based on deep convolutional neural networks is proposed. The basic plan is to add a hash layer into the deep learning framework and simultaneously learn image features and hash functions that should satisfy independence and quantization error minimized. 1st, convolutional neural network is used to learn the intrinsic implications of training pictures thus as to boost the distinguish ability and expression ability of visual feature. Second, the visual feature is putted into the hash layer, in which hash functions are learned. And therefore the learned hash functions ought to satisfy the classification error and quantization error minimized and therefore the independence constraint. Finally, given an input image, hash codes are generated by the output of the hash layer of the proposed framework and giant scale image retrieval can be accomplished in low-dimensional hamming area. Experimental results on the 3 benchmark datasets show that the binary hash codes generated by the proposed method has superior performance gains over alternative state-of-the-art hashing methods.