P071 - Red-GAN: Attacking class imbalance via conditioned generation. Yet another medical imaging perspective.

Ahmad B Qasim, Ivan Ezhov, Suprosanna Shit, Oliver Schoppe, Johannes Paetzold, Anjany Sekuboyina, Florian Kofler, Jana Lipkova, Hongwei Li, Bjoern Menze

Show abstract - Show schedule - Proceedings - PDF - Reviews - Teaser

Exploiting learning algorithms under scarce data regimes is a limitation and a reality of the medical imaging field. In an attempt to mitigate the problem, we propose a data augmentation protocol based on generative adversarial networks. The networks are conditioned at a pixel-level (segmentation mask) and at a global-level information (acquisition environment or lesion type). Such conditioning provides immediate access to the image-label pairs while controlling class specific appearance of the synthesized images. To stimulate synthesis of the features relevant for the segmentation task, an additional passive player in a form of segmentor is introduced into the the adversarial game. We validate the approach on two medical datasets: BraTS, ISIC. By controlling the class distribution through injection of synthetic images into the training set we achieve control over the accuracy levels of the datasets\' classes.
Hide abstract

Poster Session #3 - 9:30 - 11:00 UTC-4 (Tuesday)
Hide schedule

Access paper channel


Spotlight presentation


Can't display slides, your browser doesn't support embedding PDFs.

Download slides