基于CNN的医学内窥镜除雾算法.zip

# EndoscopySmokeRemoval A small but interesting research on medical endoscope Version-1 Our method compared with other methods  Original: original haze image [1] DC: Dark Channel[2] AOD-Net: AOD-Net[3] R: Training with real data S: Training with synthesized data S+R: Training with synthesized and real data DarkChannel.py is the dehaze method using dark channel. This program is obtained from internet: HazeGen.py is the program for synthesizing haze images The training program is modified from AOD-Net: https://github.com/TheFairBear/PyTorch-Image-Dehazing Todo: (1) Version-2: Codes and results will be optimized (2) Documents for more details (3) Add a module to determine whether smoke is produced Reference: [1] Real images in partial nephrectomy in da Vinci surgery. Website: hamlyn.doc.ic.ac.uk/vision/ [2] Single Image Haze Removal Using Dark Channel Prior [3] AOD-Net: An All-in-One Network for Dehazing and Beyond