Ordinal Embedding Hashing
Matlab
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
EuDist2.m
LabelMe.mat
ObtainAffinityMatrix.m
README.md
area_RP.m
compactbit.m
compressOEH.m
construct_dataset.m
construct_dataset_cifar.m
distMat.m
hammingDist.m
litekmeans.m
normalize.m
project.m
recall_precision.m
test_OEH.m
test_OEH_cifar.m
trainOEH.m

README.md

Ordinal Embedding Hashing (OEH)

Paper : Towards Optimal Binary Code Learning via Ordinal Embedding
Author : Hong Liu, Rongrong Ji, Yongjian Wu, Wei Liu
Published in AAAI 2016
Contact : lynnliu.xmu@gmail.com

Unfortunately, the original code cannot be shared, due to the Tencent Youtu Lab's IP policy.
Instead, another MATLAB implementation of OEH 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_OEH.m: the demo test codes on LabelMe dataset
test_OEH_cifar.m: the demo test codes on CIFAR10 (dataset can be download from https://pan.baidu.com/s/1o877VXC)
trainOEH.m: function to train the OEH model
constructure_data.m/constructure_data_cifar.m: function to construct the data set
compressOEH.m: function to generate the hash code
LabelMe.mat: a random subset of the original LabelMe dataset