Label-Anything is a proof of concept project developed for the '2023 IGC Research Showcase'. This project demonstrates an image labeling tool that leverages Meta's SAM (Segment Anything Model) and Gradio framework. The aim is to streamline the image labeling process by using an auto-segmentation model, thereby eliminating the need for laborious polygon labeling.
- Uses SAM (Segment Anything Model) from Meta
- Built with Gradio for an interactive interface
- Automatically segments images to simplify labeling
- [Dependency 1]
- [Dependency 2]
- [Dependency 3]
-
Clone the repository:
git clone [email protected]:kyleok/Label-Anything.git cd Label-Anything
-
Install the dependencies:
[installation command]
To run the application, simply execute the following command:
python click_grad.py