We present the code for our paper Interactive Scribble Segmentation (https://doi.org/10.7557/18.6823) presented at the Northen Lights Deep Learning conference 2023.
By drawing scribbles on an image our network will segment the class, which is being drawn, in real time. As such it is easy to correct the segmentation by simply adding more or less scribble.
To run the annotation tool run the command
python demo_Qt.py --path /path/to/images
If no path is specified the default is /images. In the tool use the mouse to draw, and the network wil segment at each mouse movement. The keyboard shortcuts are:
If you use any of our work please cite the following
@inproceedings{lowes2023interactive,
title = {Interactive Scribble Segmentation},
author = {Lowes, Mathias M. and Christensen, Jakob L. and Hansen, Bj{\o}rn Schreblowski and Hannemose, Morten Rieger and Dahl, Anders B. and Dahl, Vedrana},
booktitle = {Proceedings of the Northern Lights Deep Learning Workshop},
volume = {4},
year = {2023},
doi = {https://doi.org/10.7557/18.6823},
}