Active contour model using fractional sinc wave function for medical image segmentation

Abstract

Intensity inhomogeneity occurs when pixels in medical images overlap due to anomalies in medical imaging devices. These anomalies lead to difficult medical image segmentation. This study proposes a new active contour model (ACM) with fractional sinc function to inexpensively segment medical images with intensity inhomogeneity. The method integrates a nonlinear fractional sinc function in its curve evolution and edge enhancement. The fractional sinc function contributes in giving a rapid contour movement where it improves the curve’s bending capability. Furthermore, the fractional sinc function enables the contour evolution to move toward the object based on the preserved edges. This study uses the proposed method to segment medical images with intensity inhomogeneity using five various image modalities. With improved speed, the proposed method more accurately segments medical images compared with other baseline methods.