Dataset

/storage/datasets/dustdataset

256.8bit - 8bit images with ground truth dust + masks

256.clean - 8bit image with ground truth no dust

Film dust and scratch prediction

Based on modified PyTorch implementation of U-Net: Convolutional Networks for Biomedical Image Segmentation (Ronneberger et al., 2015).

python3 infer.py --weights_path /storage/vg_models/dust_weights/dust8_19.pth path/to/img.png

python3 train.py

Name		Name	Last commit message	Last commit date
Latest commit History 31 Commits
weights		weights
.gitignore		.gitignore
LICENSE		LICENSE
R1A.jpg		R1A.jpg
R3B.jpg		R3B.jpg
README.md		README.md
dataset.py		dataset.py
finetune.py		finetune.py
ga.py		ga.py
gaga.py		gaga.py
hackdataset.py		hackdataset.py
infer.py		infer.py
test.py		test.py
train.py		train.py
unet.py		unet.py
util.py		util.py