All rights to the pretrained models used in this project belong to their authors.
I do not vouch for their quality and do not claim to be licensed to use any model.
It is your responsibility to determine if you have permission to use the pretrained model under the license for the dataset it was trained on or licensed under.
Any use of the pretrained model is strictly regulated by the licenses under which the model is distributed.
If you own the model and want to update it (file, segmentation quality information, etc.) or don't want your model to be included in this tool, please get in touch through a GitHub issue.
The photos in the docs/imgs/input/ and docs/code_examples/python/input/ folders were taken from the Pexels website.
The original photos in the docs/imgs/compare folder were taken from the Unsplash site.
All images are copyrighted by their authors.
- https://pytorch.org/hub/pytorch_vision_deeplabv3_resnet101/
- https://github.com/NathanUA/U-2-Net
- https://github.com/NathanUA/BASNet
- https://github.com/MarcoForte/FBA_Matting
- https://arxiv.org/abs/1706.05587
- https://arxiv.org/pdf/2005.09007.pdf
- http://openaccess.thecvf.com/content_CVPR_2019/html/Qin_BASNet_Boundary-Aware_Salient_Object_Detection_CVPR_2019_paper.html
- https://arxiv.org/abs/2003.07711
- https://arxiv.org/abs/1506.01497
- https://arxiv.org/abs/1703.06870
- https://github.com/Karel911/TRACER
- https://arxiv.org/abs/2112.07380