Complementary linear biases in spatial derivative estimation for improving geometry-driven diffusion processes

摘要: This paper introduces a broadly applicable technique for visibly improving the digitized, grey-level outputs produced by host of iterative geometric diffusion methods. By replacing standard, central-difference estimates discrete spatial gradients with alternating image derivative that are offset known, complementary biases, errors accumulated during iteration reduced and quality diffusions is improved. unexpected synergy occurs at no added computational cost over central difference Very simple to implement, innovation introduced level derivatives; hence, given process, any derived higher mathematical properties-for example, group invariance or scale-space properties-can be preserved.