Automatic Thalamus Segmentation from Magnetic Resonance Images Using Multiple Atlases Level Set Framework (MALSF)

Minghui Zhang,Zhentai Lu,Yu Zhang,Qianjin Feng

doi:10.1038/s41598-017-04276-6

Abstract

In this paper, we present an original multiple atlases level set framework (MALSF) for automatic, accurate and robust thalamus segmentation in magnetic resonance images (MRI). The contributions of the MALSF method are twofold. First, the main technical contribution is a novel label fusion strategy in the level set framework. Label fusion is achieved by seeking an optimal level set function that minimizes energy functional with three terms: label fusion term, image based term, and regularization term. This strategy integrates shape prior, image information and the regularity of the thalamus. Second, we use propagated labels from multiple registration methods with different parameters to take full advantage of the complementary information of different registration methods. Since different registration methods and different atlases can yield complementary information, multiple registration and multiple atlases can be incorporated into the level set framework to improve the segmentation performance. Experiments have shown that the MALSF method can improve the segmentation accuracy for the thalamus. Compared to ground truth segmentation, the mean Dice metrics of our method are 0.9239 and 0.9200 for left and right thalamus.

Highlights

Because of low image contrast, noise, and missing or diffuse boundaries, some discrete methods such as thresholding, FCM1, and region growing[2] are not reliable because they only use image intensity information
Since our method aims to improve segmentation accuracy in two ways, by incorporating multiple registration method into the conventional multiple atlas-based methods and by taking the advantage of level set framework, it is worthwhile to evaluate the contributions from these two different methods
We propose a novel method for segmenting the thalamus that combines multiple registration methods and a level set fusion strategy

Summary

Introduction

Because of low image contrast, noise, and missing or diffuse boundaries, some discrete methods such as thresholding, FCM1, and region growing[2] are not reliable because they only use image intensity information. Under the multi-atlas framework, auto-context model based classifiers are trained for all atlases to incorporate anatomical variability. It is time-consuming to extract image appearance features, texture features and context features. A natural extension of the MV method is to use adaptive weighted averaging Another popular approach is simultaneous truth and performance level estimation (STAPLE), which uses the expectation-maximization (EM) algorithm to achieve the best possible final segmentation[18]. This multi-atlas based segmentation approach reduces the effect of errors associated with individually propagated atlases. The main limitations of the multi-atlas segmentation methods are that they often lead to a compromise between the accuracy of the registration and the smoothness of the deformation, and that those methods are time consuming

Objectives

Methods

Findings

Discussion

Conclusion