Towards developing a segmentation method for cerebral aneurysm using a statistical multiresolution approach

Abstract

The computer aided diagnosis (CAD) algorithms are considered crucial during the treatment planning of cerebral aneurysms (CA), where segmentation is the first and foremost step. This paper presents a segmentation algorithm in two-dimensional domain combining a multiresolution and a statistical approach. Precisely, Contourlet transform (CT) extracts the image features, while Hidden Markov Random Field with Expectation Maximization (HMRF-EM) segments the image, based on the spatial contextual constraints. The proposed algorithm is tested on Three-Dimensional Rotational Angiography (3DRA) datasets; the average values of accuracy, DSC, FPR, FNR, specificity, and sensitivity, are found to be 99.64%, 92.44%, 0.09%, 5.81%, 99.84%, and 93.22%, respectively. Both qualitative and quantitative results obtained show the potential of the proposed method.