[News] This paper is accepted for IJCNN'24!!

This repository provide the model for compressive sensing-based reconstruction.

Training images are cropped to 96*96 pixel size and 89600 images are randomly selected for training the network. You can download it from here.

Please set the training dataset under ./Datasets/ folder. You can train HFFDNet with following script:

python train.py --sensing_rate 0.1 --layer_num 8

Please set the test dataset under ./Datasets/ folder. You can test HFFDNet with following script:

python test.py --sensing_rate 0.1 --layer_num 8