- python==3.7.5
- torch==1.13.1
- Torchvision==0.14.1
- Scikit_image==0.19.2
- Skimage==0.0
- timm==0.3.2
- tensorboard==2.11.2
- tensorboardX==2.5.1
- tqdm==4.64.1
- einops==0.4.1
- markdown==3.4.3
- markplotlib==3.5.2
- numpy==1.12.6
- opencv-python==4.7.0.72
- openpyxl==3.1.2
- pillow==9.5.0
- pysodmetrics==1.4.0
- PyYAML==6.0
- tabulate==0.9.0
- Point supervised PCOD: google
-
Generate square clusters (d*d) by point label, d*d cluster , After that training generates Pred_Map ,then , run Point2gt.py to generate the final supervisor labels .
-
Using the labels generated above for training ( "1" stands for foregrounds, "2" for backgrounds, and "0" for unlabeled regions. (The image is viewed as black because its range is 0-255)) . Put it in './CodDataset/train/Label'
-
Download training dataset and testing dataset . Put them in the right path, including './CodDataset/train/Imgs', '/CodDataset/test/CAMO/(GT+Imgs)', '/CodDataset/test/COD10K/(GT+Imgs)'.... and '/CodDataset/test/HCK4/(GT+Imgs)'
-
Download pretrain weight and put it in './Point supervised camouflaged object detection/xxx.pth'
-
Run python train.py.
- Modify path and filename.
- Run python test.py
Weakly-Supervised Camouflaged Object Detection with Scribble Annotations