Brain MR image segmentation using multiphase active contours based on local and global fitted images.

Akram, Farhan; Singh, Vivek Kumar; Sarker, Md. Mostafa Kamal; Garcia, Miguel Angel; Puig, Domenec

Abstract

The study of white matter (WM), grey matter (GM) and cerebrospinal fluid (CSF) regions in the brain magnetic resonance (MR) images can be useful for determining different brain disorders, assisting brain surgery, post-surgical analysis, saliency detection and for studying regions of interest. In this paper, a novel hybrid region-based multiphase (four-phase) active contours method is proposed to partition a brain MR image into three distinct regions i.e., GM, WM and CSF. The proposed energy functional is formulated by combining local and global fitted images in a multiplicative fashion. Both fitted images are defined by integrating two-phase global and local intensity from Chan-Vese (CV) and local binary fitted (LBF) models, respectively. In this paper, a post processing (pixel correction) method is also devised which improves the accuracy of the segmented WM, GM and CSF regions in a brain MR image. Different thresholds are decided based on averages of all three regions. According to the computed thresholds, a binary value (0 or 1) is then assigned to each pixel. Experimental results using both two-dimensional (2D) and three-dimensional (3D) brain MR images show that the proposed method outperforms the state-of-the-art both qualitatively and quantitatively.

Citation

AKRAM, F., SINGH, V.K., SARKER, M.M.K., GARCIA, M.A. and PUIG, D. 2018. Brain MR image segmentation using multiphase active contours based on local and global fitter images. In Falomir, Z., Gilbert, K. and Plaza, E. (eds.). Artificial intelligence research and development: current challenges, new trends and applications; contributions from 21st international conference of Catalan Association for Artificial Intelligence 2018 (CCIA 2018), 8-10 October 2018, Alt Empordà, Spain. Frontiers in artificial intelligence and applications, 308. Amsterdam: IOP Press [online], pages 325-334. Available from: https://doi.org/10.3233/978-1-61499-918-8-325

• Conference Name: 21st international conference of Catalan Association for Artificial Intelligence 2018 (CCIA 2018)
• Conference Location: Alt Empordà, Catalonia
• Start Date: Oct 8, 2018
• End Date: Oct 10, 2018
• Acceptance Date: Jun 4, 2018
• Online Publication Date: Oct 10, 2018
• Publication Date: Oct 31, 2018
• Deposit Date: Dec 4, 2021
• Publicly Available Date: Jan 31, 2022
• Publisher: IOS Press
• Pages: 325-334
• Series Title: Frontiers in artificial intelligence and applications
• Series Number: 308
• Series ISSN: 0922-6389
• Book Title: Artificial intelligence research and development: current challenges, new trends and applications
• ISBN: 9781614999171
• DOI: https://doi.org/10.3233/978-1-61499-918-8-325
• Keywords: Active contours; Multiphase; Local and global fitted images; Brain segmentation; Pixel correction; Thresholding
• Public URL: https://rgu-repository.worktribe.com/output/1542075