Interactive Grabcut with Flask

Description

This is an interactive implementation of GrabCut that runs in Flask. The Flask app allows users to first draw a rectangle to indicate the foreground and then refine the GrabCut output by drawing lines to indicate background to be removed on the first canvas and indicate foreground to be included on the second canvas.

Read more about the app and its implementation!

Running the App

• Use python __init__.py to run the flask server.
• Go to http://127.0.0.1:5000/ on your browser.

Adding your own images

• Make a new directory with a project name you want under /static/images.
• In this directory, make a directory for the original images called "_original".
• Add your images to the directory names _original.
• Go to /static/js/grabcut_demo.js and call makeGrabCutRow(appending_container,image_path, project_name) with the id of the container you want to append to (you can also use the appending_container variable already defined), the path to the image, and the project name.
• The path to the image should be ../static/images/<project_name>/_original/<image_file_name>

Dependencies

• To run the app you will need Flask, opencv-python and numpy, use the following to install them:

pip install flask

pip install opencv-python

pip install numpy