A Framework of Vertebra Segmentation Using the Active Shape Model-Based Approach

Abstract

We propose a medical image segmentation approach based on the Active Shape Model theory. We apply this method for cervical vertebra detection. The main advantage of this approach is the application of a statistical model created after a training stage. Thus, the knowledge and interaction of the domain expert intervene in this approach. Our application allows the use of two different models, that is, a global one (with several vertebrae) and a local one (with a single vertebra). Two modes of segmentation are also proposed: manual and semiautomatic. For the manual mode, only two points are selected by the user on a given image. The first point needs to be close to the lower anterior corner of the last vertebra and the second near the upper anterior corner of the first vertebra. These two points are required to initialize the segmentation process. We propose to use the Harris corner detector combined with three successive filters to carry out the semiautomatic process. The results obtained on a large set of X-ray images are very promising.

Improving an Active Shape Model with Random Classification Forest for Segmentation of Cervical Vertebrae

“…An improved Mahalanobis distance-based search method has been introduced in [13]. This method has been used for vertebrae segmentation in [3,4]. …”

“…4. The axis (C1) and atlas (C2) of the cervical vertebrae have not been studied in this work due to their ambiguity in lateral X-ray images similar to other work in the literature [3,4]. …”

“…Due to the clinical importance of vertebrae segmentation, there is a large body of research in the literature [3][4][5][6][7][8][9][10][11]. Based on this literature, arguably the most successful segmentation method is the statistical shape model (SSM).…”

“…This method involves a training phase and an optimization step to find the amount of displacement needed to converge the mean shape on the actual object boundary. The method has been shown to work well on cervical vertebra X-ray images in [3,4]. In [15], a conventional binary classifier and a boosted regression predictor has been compared and used to improve the performance of ASM segmentation during image search phase.…”

“…Unlike the Mahalanobis distance-based method used in [3,4,13,14], this method predicts the displacement directly without a need of a sliding window-based search technique. Our method uses a multi-class forest in contrast with the binary classification method used in [15].…”