×

Segmentation of ultrasound images based on fuzzy \(c\)-means clustering energy minimization. (Chinese. English summary) Zbl 1340.92026

Summary: A novel active contour model based on fuzzy \(C\)-means energy minimization is put forward. First, the proposed method uses fuzzy \(C\)-means clustering to segment images to obtain the fuzzy membership degree value of image foreground and background. Subsequently, the local pixel information of the target area and its membership degree value are used as the initial value of the level set function of the active contour model. The proposed method applies a fast algorithm to directly calculate the energy minimization of the fuzzy \(C\)-means membership degree value and drive the curve evolution of the traditional active contour model. The method avoids solving Euler-Lagrange equation. The proposed method is compared with the classic active contour model segmentation algorithms. The experiment is performed on simulated and the clinical ultrasonic images. It shows that the proposed algorithm can effectively segment ultrasound images with good segmentation performance and reasonable speed. The method can solve the problems of intensity inhomogeneous, fuzzy boundaries and speckle noise.

MSC:

92C55 Biomedical imaging and signal processing
68T10 Pattern recognition, speech recognition
90C70 Fuzzy and other nonstochastic uncertainty mathematical programming
PDFBibTeX XMLCite