https://github.com/dnanhkhoa/rust-background-removal

Remove backgrounds from images with ease
https://github.com/dnanhkhoa/rust-background-removal

background-removal image-matting image-processing image-segmentation onnx onnxruntime open-source rust

Last synced: about 1 month ago
JSON representation

Remove backgrounds from images with ease

• Host: GitHub
• URL: https://github.com/dnanhkhoa/rust-background-removal
• Owner: dnanhkhoa
• License: mit
• Created: 2023-09-13T14:34:07.000Z (over 1 year ago)
• Default Branch: main
• Last Pushed: 2024-08-13T23:41:04.000Z (6 months ago)
• Last Synced: 2024-08-14T02:06:41.069Z (6 months ago)
• Topics: background-removal, image-matting, image-processing, image-segmentation, onnx, onnxruntime, open-source, rust
• Language: Rust
• Homepage:
• Size: 4.53 MB
• Stars: 24
• Watchers: 1
• Forks: 9
• Open Issues: 1
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# 🦀 Background Removal

![Example](images/example.png)

This Rust implementation is a minimal port of the original JavaScript library [@imgly/background-removal](https://github.com/imgly/background-removal-js). It aims to explore **the use of ONNX models in Rust** for background removal tasks. If you're interested in the full capabilities and details of the background removal process, I highly recommend checking out the original JavaScript library's [README](https://github.com/imgly/background-removal-js) here.

**Update:** For the technical details of the model, please check out the paper **Highly Accurate Dichotomous Image Segmentation** mentioned in [this repository](https://github.com/xuebinqin/DIS).

## Usage

```sh

# Download all pretrained models

./download_models.sh

# Update variable input_img_file in src/main.rs

# Run the program

cargo run --release

```

## Limitations

The model resolution is limited to a maximum of 1024x1024 pixels.

## License

[MIT](https://choosealicense.com/licenses/mit/)

## Related sites

[pykeio/ort: A Rust wrapper for ONNX Runtime](https://github.com/pykeio/ort)

## Credits

Khoa Duong, David Horner