An adaptive-scale active contour model for inhomogeneous image segmentation and bias field estimation

Highlights：

• A novel adaptive scale operator is proposed based on image entropy and Gaussian kernel function, which not only realizes the adaptive adjustment of scale for the ASACM according to the different degrees of intensity inhomogeneity but also improves the segmentation speed and robustness to initial contours and to noise.

• By distributing a dependent-membership function for each pixel, we construct an improved bias field estimation term to estimate the bias field of severe intensity inhomogeneous image that further improves the segmentation accuracy.

• To avoid the time-consuming re-initialization process and to solve the instability in traditional penalty term, we define an improved penalty term using piecewise polynomial.