Active learning based segmentation of Crohns disease from abdominal MRI. Issue 128 (May 2016)

• Record Type: Journal Article
• Authors: Mahapatra, Dwarikanath; Vos, Franciscus M.; Buhmann, Joachim M.
• Is Part Of: Computer methods and programs in biomedicine. Issue 128(2016)
• Page Start: 75
• Page End: 85
• Publication Date: 2016-05
• Subjects: Crohns disease -- Segmentation -- Semi supervised classification -- Active learning -- Graph cuts -- Label query; Medicine -- Computer programs -- Periodicals; Biology -- Computer programs -- Periodicals; Computers -- Periodicals; Medicine -- Periodicals; Médecine -- Logiciels -- Périodiques; Biologie -- Logiciels -- Périodiques; Biology -- Computer programs; Medicine -- Computer programs; Periodicals; Electronic journals; 610.28
• DOI: 10.1016/j.cmpb.2016.01.014
• Languages: English
• ISSNs: 0169-2607

Abstract

Abstract : Highlights: We propose an interactive method combining semi supervised learning (SSL) and active learning (AL) for segmenting Crohns disease affected regions in MRI. A novel query strategy for AL has been proposed that makes use of context information to identify query samples. Compared to fully supervised methods we obtain high segmentation accuracy with fewer samples and lesser computation time. Our method has the potential to be used in scenarios which pose difficulties in obtaining large numbers of accurately labeled data. Abstract: This paper proposes a novel active learning (AL) framework, and combines it with semi supervised learning (SSL) for segmenting Crohns disease (CD) tissues from abdominal magnetic resonance (MR) images. Robust fully supervised learning (FSL) based classifiers require lots of labeled data of different disease severities. Obtaining such data is time consuming and requires considerable expertise. SSL methods use a few labeled samples, and leverage the information from many unlabeled samples to train an accurate classifier. AL queries labels of most informative samples and maximizes gain from the labeling effort. Our primary contribution is in designing a query strategy that combines novel context information with classification uncertainty and feature similarity. Combining SSL and AL gives a robust segmentation method that: (1) optimally uses few labeled samples and many unlabeled samples; and (2) requires lower training time. Experimental results show our method achieves higher segmentation accuracy than FSL methods with fewer samples and reduced training effort.