Brain Image Segmentation by Multiscale Analysis and Template Matching
Xiuquan Ji, Song Wang, Mark B. Skouson, Zhi‐Pei Liang · 1998
Introduction The primary goal of brain image segmentation is to partition a given brain image into non-intersecting regions representing true anatomical structures such as grey matter, white matter, etc. Over the last decade, many methods have been proposed to tackle this problem. A partial list includes edge-based methods [l], knowledge or rule-based methods [2], statistical model-based methods [3], neural network methods [4], and deformable model based methods [5]. In spite of this progress, automatic segmentation of brain structures remains a very challenging task. This paper presents a new hybrid method which integrates multiscale analysis, image normalization and elastic template deformation.