/JPSH

There is the official implementation of our TOMM paper "Binary Representation via Jointly Personalized Sparse Hashing"

Primary LanguageMATLAB

Binary Representation via Jointly Personalized Sparse Hashing (TOMM 2022)

There is the official implementation of our paper "Binary Representation via Jointly Personalized Sparse Hashing".

Overview

Existing hashing methods lack satisfactory performance in dealing with real-world scenarios that produce similar features with different semantic information. To address this challenge, we proposed an unsupervised method, namely Jointly Personalized Sparse Hashing (JPSH), for binary representation learning. It constructs a seamless hash function, which consists of twofold properties: semantic and pairwise similarities. JPSH accommodated the proposed Personalized Sparse Hashing (PSH) module to maintain semantic similarity and preserves pairwise similarity using a manifold-based hashing method. Thus, we learn discriminative binary codes by combining the two similarities. The framework is shown in the following figure.

image

Description and Usage

demo.m--run demo.m to verify the JPSH model.

test_model.m--return evaluation values of the JPSH model.

train_model.m--train the JPSH model

Please put the training data in the data folder, and then run demo.m. Finally, the results will be saved in the results folder.

Support

MATLAB 2017 on a PC with 3.6GHz and 64G RAM.

Citation

If you are interested in our JPSH model, please consider citing our paper:

@article{xiaoqin_tomm_2022,
author = {Wang, Xiaoqin and Chen, Chen and Lan, Rushi and Liu, Licheng and Liu, Zhenbing and Zhou, Huiyu and Luo, Xiaonan},
title = {Binary Representation via Jointly Personalized Sparse Hashing},
year = {2022},
publisher = {Association for Computing Machinery},
address = {New York, NY, USA},
issn = {1551-6857},
url = {https://doi.org/10.1145/3558769},
doi = {10.1145/3558769},
journal = {ACM Transactions on Multimedia Computing, Communications, and Applications}}