CONTENT BASED IMAGE RETRIEVAL WITH HASH CODES
Sri Vasavi College, Erode Self-Finance Wing 3rd February 2017 National Conference on Computer and Communication NCCC’17
This paper proposes a new CBIR(Content Based Image Retrieval) system using Hash codes performs few major tasks. The first one is feature extraction (FE), where a set of features or feature vector, is generated for accurate representation of content for each image in the Wang dataset comprising of 499 images. A feature vector is much smaller in size compared to the original image. The second is similarity measurement(SM), where a distance between the query image and each image in the dataset using their features is computed so that the top images can be retrieved. The third is hash functions produce hash values based on the image visual appearance. Hash function calculates similar values for similar images also for dissimilar images dissimilar hash values are calculated. Using similarity function to compare two hash values, it can be decided whether two images are different or not. Another task is hamming distance to calculate how many bits are different in hash.
Y-G Jiang, J Wang, X Due and S-F Chang, “Query adaptive Image Search with hash codes,” IEEE Trans. on Multimedia, vol. 15, no.2, Feb 2013. Y.-G. Jiang, J compact codes,” in Proc. ACM Int. Conf. Multimedia, 2011 C. Y. Mu, J. Shen, and S. Yan, “Weakly- Recognition, 2010.supervised hashing in kernel Space,” in Proc. IEEE Conf. Computer Vision and Pattern Hao Xu , JD Wang , Zhu Li , G Zeng , S Li and Nenghai Yu ,” Complementary hashing for approximate Nearest Neighbor Search ,”in Proc IEEE transaction ,2011. J. Wang, S. Kumar and S.-F. Chang, “Sequential projection and Pattern
Hash function, Hamming Distance, CBIR, FE, SM