Contrast Constrained Local Binary Fitting for Image Segmentation

Abstract

This paper presents a more robust and efficient level set method than the original Local binary fitting (LBF) model in [6] for image segmentation under a constrained energy minimization framework. Image segmentation is formulated as a problem of seeking an optimal contour and two fitting functions that best approximate local intensities on the two sides of the contour. The contribution in this paper is twofold. First, we introduce a contrast constraint on the fitting functions to effectively prevent the contour from being stuck in spurious local minima, which thereby makes our model more robust to the initialization of contour. Second, we provide an efficient narrow band implementation to greatly reduce the computational cost of the original LBF algorithm. The proposed algorithm is validated on synthetic and real images with desirable performance in the presence of intensity inhomogeneities and weak object boundaries. Comparisons with the LBF model and the piecewise smooth (PS) model demonstrate the superior performance of our model in terms of robustness, accuracy, and efficiency.