This is the source code of our TMM 2020 paper "Deep Reinforcement Learning for Image Hashing". Please cite the following paper if you use our code.
Yuxin Peng, Jian Zhang and Zhaoda Ye, "Deep Reinforcement Learning for Image Hashing", IEEE Transactions on Multimedia (TMM), Vol. 22, No. 8, pp. 2061-2073, Aug. 2020.(SCI, EI)
This code is implemented with pytorch.
The codes takes the images of cifar10 as the input.
The data and the pre-trained model which can be download from the links:
https://pan.baidu.com/s/1BWKRWJPQ5r-rP3H9MRXe6g
password: 94a2
The images are organized as follows:
./train/class_id/name_of_image
./test/class_id/name_of_image
Start training and tesing by executiving the following commands. This will train and test the model on Cifar10 datatset.
python train.py 0 32 16 cifar
train.py [gpu_id,bit_length,batch_size,dataset]