clipx 0.5.0

Image background removal tool supporting U2Net and CascadePSP models

clipx

English | 简体中文

Introduction

clipx is an open-source Python library designed for quick and flexible image background removal. It supports both command-line interface (CLI) and Python API usage. Currently, it integrates two image segmentation models, U2Net and CascadePSP, which can be used individually or combined for enhanced performance. Additionally, clipx has an extensible architecture, making it easy to incorporate additional models in the future.

---

Installation

pip install clipx

---

CLI Examples

• Generate transparent image using combined mode (U2Net + CascadePSP):

clipx -i input.jpg

• Specify output path:

clipx -i input.jpg -o output.png

• Use fast mode for CascadePSP (faster but less accurate):

clipx -i input.jpg -o output.png --fast

---

Python API Example

from PIL import Image
from clipx import remove

# Testing the API
img = Image.open("tests/leaves-8273504_1920.jpg")
result = remove(img)
result.save("api_test_result.png")

# Get only the mask
mask = remove(img, only_mask=True)
mask.save("mask_only.png")

---

Acknowledgements and Code Sources

This project uses code from the following open source projects:

• U2Net: U^2-Net: Going Deeper with Nested U-Structure for Salient Object Detection
• CascadePSP: CascadePSP: Toward Class-Agnostic and Very High-Resolution Segmentation via Global and Local Refinement
• Special thanks to rembg project for inspiration on project structure and implementation approach.

We greatly appreciate the original authors' work and contributions.

---

License

This project is licensed under the MIT License.