A new effective hybrid segmentation method based on C-V and LGDF

Abstract

Image segmentation is a significant research topic in image processing and computer vision. Active contour methods (ACMs) are widely used in image segmentation. In this paper, a new hybrid ACM segmentation model based on Chan-Vese (C-V) and Local Gaussian Distribution Fitting (LGDF) methods is proposed for the images with intensity inhomogeneity. In this model, new gradient descent flow equations are proposed and applied for the energy minimization of C-V and LGDF methods. Firstly, the proposed C-V method is applied to the image to effectively and quickly find the homogeneous regions of the image. Then, the proposed LGDF method is performed in these regions to detect inhomogeneous areas of the image. Thus, more effective and successful segmentation is obtained for inhomogeneous images. Experimental results show that the satisfactory segmentation results have been obtained by the proposed method for MRI and real images. Also, the proposed method is compared with the local binary fitting, LGDF, adaptive local-fitting-based, global and local weighted signed pressure ACMs, and convolutional neural network-based methods.