Deep Image Hashing

Description

Implementation of state-of-the-art deep hashing methods in PyTorch Lightning

How to run

First, install dependencies

# clone project   
git clone https://github.com/bgswaroop/deep-hashing

# install project   
cd deep-hashing 
pip install -e .   
pip install -r requirements.txt

Next, navigate to any file and run it.

# module folder
cd project

# run module
# Customize your run by overriding all the default arguments.
# To see list of possible arguments >> python module_name.py --help   
python sota_2016_CVPR_DSH.py
python sota_2017_NIPS_DSDH.py    

SOTA deep hashing

Implementation of other state-of-the-art deep hashing methods in PyTorch lightning

The following table summarizes the provided implementations of other deep hashing methods along with their performance on popular datasets. The evaluation metric is mAP (mean average precision).

Supervised Methods	CIFAR-10
	12-bit	48-bit
Deep supervised hashing for fast image retrieval DSH (paper, code) - CVPR 2016	0.6249 scratch	0.6948 scratch	0.7794 fine-tune
Deep supervised discrete hashing DSDH (paper, code) - NIPS 2017		- using CNN-F	0.9540 using vgg16