Level set segmentation of intensity inhomogeneous images based on local linear approximation of difference image

This study presents a new local image approximation active contour model in a variational level set formulation for image segmentation. The model consists of two parts: a local linear approximation term and the regularisation terms. The local linear approximation term is more accurate than the local average in the local binary fitting (LBF) model. The authors´ proposed model achieves sub-pixel segmentation result. Comparative experiments show that the proposed model achieves satisfying results, and demonstrates its efficiency.