Deep Self-Enhancement Hashing For Robust Multi-label Cross-modal Retrieval -- Tensorflow Implementation

Deep Self-Enhancement Hashing (DSEH)

The details can be found in the paper Deep Self-Enhancement Hashing For Robust Multi-label Cross-modal Retrieval (submitted to Pattern Recognition)

Implementation platform:

• python 3.7
• tensorflow 2.5.0
• matlab 2016b

Datasets

We use the following datasets.

• MIRFlickr-25k
• MSCOCO
• NUSWIDE

Pre-extracted features by ResNet-152:

• MIRFlickr-25k download lmt3

Training

The command for training is

• python3 train_DSEH.py
• our trained model (on MIRFlickr-25k) can be download ve27

Evaluation

The command for evaluation is

• extract hash codes: python3 DSEH_encoding.py
• run evaluation.m