Fast thresholding computation by searching for zero derivatives of image between-class variance

Many bilevel thresholding methods are extendable for multi-level thresholding applications as well. However, the amount of thresholding computation will consequently increase significantly. This study is specifically devoted to improving the efficiency in Otsu´s (1979) bi-level and multi-level thresholding computations. Zero (partial) derivatives of image between-class variance with respect to gray levels are derived and it results in a set of nonlinear equations to solve for optimal thresholds using a root finder. The proposed method is more efficient than Otsu´s in computation, especially for multi-level thresholding applications. The adequacy of the proposed method has been proven through extensive tests. Included are two examples to illustrate the feasibility of the proposed method and its outstanding performance in computation