cloth-segmentation / README.md
wildoctopus's picture
Update README.md
4a4859d
|
raw
history blame
2.19 kB
metadata
title: Cloth Segmentation
emoji: 📉
colorFrom: indigo
colorTo: gray
sdk: gradio
sdk_version: 3.36.0
app_file: app.py
pinned: false
license: mit

Huggingface cloth segmentation using U2NET

Python 3.8 License: MIT Open In Colab

This repo contains inference code and gradio demo script using pre-trained U2NET model for Cloths Parsing from human portrait.
Here clothes are parsed into 3 category: Upper body(red), Lower body(green) and Full body(yellow). The provided script also generates alpha images for each class.

Inference

  • clone the repo git clone https://github.com/wildoctopus/huggingface-cloth-segmentation.git.
  • Install dependencies pip install -r requirements.txt
  • Run `python process.py --image 'input/03615_00.jpg' . Script will automatically download the pretrained model.
  • Outputs will be saved in output folder.
  • output/alpha/.. contains alpha images corresponding to each class.
  • output/cloth_seg contains final segmentation.

Gradio Demo

  • Run `python app.py'
  • Navigate to local or public url provided by app on successfull execution.

OR

  • Inference in colab from here Open In Colab

Huggingface Demo

Output samples

Sample 000 Sample 024

This model works well with any background and almost all poses.

Acknowledgements