This repo is a Tensorflow implementation of paper Full Resolution Image Compression with Recurrent Neural Networks. This repo also contains a trained Residual LSTM model (saved in save/) on a small dataset (doesn't include dataset). Here is one example shown in the paper
Original image:
Reconstruction:
- SciPy
- Tensorflow 1.4+
Training, put training data in imgs/ folder and run following code for default setting. The model parameter can be modified in model.py. Running time comparison of original and reconstructed images can be seen in eval/. Model file is saved in save/model.
python train.py
Encoding
python encode.py --model save/model --input kodim05.png --iters 10 --output compressed.npz
Decoding
python decode.py --model save/model --input compressed.npz --output compressed.png
Evaluation, code from Tensorflow's official repo
python msssim.py -o kodim05.png -c compressed.png