Wavelet Energy Embedded into a Level Set Method for Medical Images Segmentation in the Presence of Highly Similar Regions

Abstract

This paper is motivated by present a new segmentation method that integrates a novel feature, which is able to enhance the dissimilarity between regions. This feature is integrated to formulate a new level set based active contour model, which addresses the segmentation of regions with highly similar intensities, which do not have clear boundaries between them. The power of wavelet transform is adapted to formulate the new feature, named as wavelet energy. In this formulation, the two terms that guide the contour are the wavelet energy incorporated region term and the contour smoothness term. With this formulation, the equations for evolving the contour are derived and implemented in MATLAB. This segmentation method is named Wavelet energy Embedded into a level set method. The experimental results show that the proposed method is able to segment the region of interest that have high similarity in intensities with their background.

Fully automatic multisegmentation approach for magnetic resonance imaging brain tumor detection using improved region‐growing and quasi‐Monte Carlo‐expectation maximization algorithm

“…The accuracy of this technique depends on the doctor's knowledge. 7 Many researchers have developed hybridized semi-and fully automatic MRI image segmentation models. Medical imageprocessing methods, used for fully automatic segmentation, are classified into four main categories.…”

“…1 The last category is contour detections, including Active Contour Models, Gradient Vector Flow, Vector Field Convolution 6 and level set. 7 Many researchers have developed hybridized semi-and fully automatic MRI image segmentation models. Lu et al 4 proposed an improved region-growing algorithm initialized by the quasi-Monte Carlo (QMC) method for liver segmentation.…”