metadata
library_name: keras
language:
- pt
- en
tags:
- art
- image
- remove-background
- u^2-net
- residual
U²-Net
Input Shape:(512, 512, 3)
This is an implementation of U²-Net for image background removing purposes based on this paper. The U²-Net architecture is designed to capture both fine details and global context, making it highly effective for image segmentation tasks such as background removal.
Dataset
Dichotomous Image Segmentation (DIS) is a dataset wich contains around 5k images (images + masks) and it's used in image segmenation model training.
You can find more about this dataset here.
Me:
Find me at GitHub: https://github.com/reidn3r
Find me at Linkedin: https://www.linkedin.com/in/reidner-adnan/
Project repository: GitHub repository