Algorithm for fast TFT-LCD Mura defect image segmentation based on Chan-Vese model

Aiming at the issue of low efficiency of traditional Chan-Vese model(C-V model)method for uneven background TFT-LCD Mura defect segmentation,a new formulation which was the difference between the level set function and signed distance function was introduced to C-V model to completely eliminate the re-initialization procedure of signed distance function.In order to reduce the effect from non-uniformity brightness of image,the C-V model was proved by adding the brightness difference between inside and outside area of the contour curve to improve the accuracy of the segmentation.In the numerical implementation,the unconditionally stable semi-implicit scheme was used to accelerate the evolution velocity by increasing the time step appropriately.This semi-implicit scheme has higher segmentation speed than the finite difference scheme and AOS format.The results show that the proposed algorithm can segment the uneven background Mura defect accurately and rapidly.