Two-dimensional extension of minimum error threshold segmentation method for gray-level images

One-dimensional minimum error thresholding method assumed that the histogram distributions of object and background are governed by a mixture Gaussian distribution. Considering the affects of noise and other factors on image quality, based on the assumption of a two-dimensional mixture Gaussian distribution, a two-dimensional expression of the minimum error thresholding method on the two-dimensional gray-level histogram is proposed. In order to improve the running speed, the fast recursive formulas are also given. Experimental results show that the two-dimensional minimum error thresholding method is a valuable image segmentation method, and can be well adapted to the images with noises and large variances between object and background. © 2009 Acta Automatica Sinica. All rights reserved.