Title: Volumetric tumour detection using improved region grow algorithm

Authors: Shitala Prasad; Shikha Gupta

Addresses: Image Lab, GREYC, University of Caen Normandy, France ' Google Bangalore, India

Abstract: This paper works on segmentation of brain pathological tissues (tumour, edema and narcotic core) and visualise it in 3D for their better physiological understanding. We propose a novel approach which combines threshold and region grow algorithm for tumour detection. In this proposed system, FLAIR and T2 modalities of MRI are used due to their unique ability to detect the high and low contrast lesions with great accuracy. In this approach, first the tumour is segmented from an image which is a combination of FLAIR and T2 image using a threshold value, selected automatically based on the intensity variance of tumour and normal tissues in 3D MR images. Then the tumour part is extracted from the actual 3D MRI of brain by selecting the largest connected volume. To correctly detect tumour 26 connected neighbours are used. The method is evaluated using a publically available BRAT dataset of 80 different patients having gliomas tumours. The accuracy in terms of detection is reached to 97.5% which is best compared to other state-of-the-art in given time frame. The algorithm takes 4-5 minutes for generating the 3D visualisation for final output.

Keywords: 3D volumetric; brain tumour; region growing algorithm; thresholding; voxel seeding.

DOI: 10.1504/IJCSYSE.2018.091388

International Journal of Computational Systems Engineering, 2018 Vol.4 No.2/3, pp.127 - 139

Received: 19 Oct 2016
Accepted: 26 Mar 2017

Published online: 30 Apr 2018 *

Full-text access for editors Access for subscribers Purchase this article Comment on this article