Paper : Ordinal Constrained Binary Code Learning for Nearest Neighbor Search
Author : Hong Liu, Rongrong Ji, Yongjian Wu, Feiyue Huang
Published in AAAI 2017
Contact : [email protected]
Unfortunately, the original code cannot be shared, due to the Tencent Youtu Lab's IP policy.
Instead, another MATLAB implementation of OCH has been released, which details are similar to the original version, but it would be still useful and gives similar results.
This package contains cleaned up codes for the AAAI paper, including:
test_OCH.m: the demo test codes
trainOCH.m: function to train the OCH model
constructure_data.m: function to construct the data set
compressOCH.m: function to generate the hash code
LabelMe.mat: a random subset of the original LabelMe dataset