Image subject isolation/background removal using CNNs on-device (WinForms frontend)
Backsolate is a frontend for several models specializing in detecting subjects in images. This app allows you to use those models to remove backgrounds from the images you input.
- Download the latest version from the releases.
- Open the app, choose an image by clicking on the left box, and click Remove Background.
- The app will now download a model of approximately 160MB.
- Wait until it finishes processing the image.
- After the processing is complete, you can save the image by clicking on the right box.
- If the background removal doesn't work or does not produce desirable results, try using a different model by selecting one from the dropdown on the bottom right.
Experiment with different models until you find the one that works best for you.
- u2net, u2netp, u2net_human_seg, u2net_cloth_seg, silueta, isnet-general-use, isnet-anime by Xuebin Qin (https://github.com/xuebinqin/U-2-Net)
- birefnet-general, birefnet-general-lite, birefnet-portrait, birefnet-dis, birefnet-hrsod, birefnet-cod, birefnet-massive by ZhengPeng7 (https://github.com/ZhengPeng7/BiRefNet)
- ONNX Model Runtime by Microsoft (https://github.com/microsoft/onnxruntime)
- Model ONNX files from: https://github.com/danielgatis/rembg
This project is licensed under the MIT License.