--The code for paper:"Stochastic Multiview Hashing for Large-scale Near-duplicate Video Retrieval"-- —-Version 1.0—- --Updated: 10/07/2017--
CC_WEB_VIDEO dataset: http://vireo.cs.cityu.edu.hk/webvideo/ UQ_VIDEO dataset: http://staff.itee.uq.edu.au/shenht/UQ_VIDEO/
-
'SMVH_main.m' is the main function. Please find the video datasets from providers' websites. We provide a pretrained weight matrix 'W_pretrained.mat' using HSV and LBP features, in which the code length is 320.
-
'MAP_CC.m' and 'PR_CC.m' are evaluation codes for CC_WEB_VIDEO dataset, 'MAP_UQ.m' and 'PR_UQ.m' are for UQ_VIDEO dataset.
-
We do not provide the code for the other hashing method, e.g., spectral hashing (SPH), self-taught hashing (STH) and multiple feature hashing (MFH), due to the copyright problem.
If you have any question, please directly contact the author by email. Email: haoyanbin@hotmail.com