Atmegal / Comprehensive-Distance-Preserving-Autoencoders-for-Cross-Modal-Retrieval

The code of Comprehensive Distance-Preserving Autoencoders for Cross-Modal Retrieval

This is the source code of our ACM Multimedia Conference 2018 paper "Comprehensive Distance-Preserving Autoencoders for Cross-Modal Retrieval", Please cite the following paper if you use our code.

Yibing Zhan,Jun Yu,Zhou Yu,Rong Zhang,Dacheng Tao,Qi Tian,"Comprehensive Distance-Preserving Autoencoders for Cross-Modal Retrieval", ACM MM 2018


Usage
For Wikipedia-Multiple dataset:

1.Generate the Comprehensive Distance-Preserving Autoencoders by using the code:Auto1.py
2. Get the best data after training by using the code:decoder.py
3. Unsupervised Cross-Modal Similarity Measurement is used to further improve the retrieval performance by using the code:Knn.py


For more information, please refer to our ACM MM paper.