Iterative Triclass Thresholding Technique using Otsu's Method

A new Iterative Triclass Thresholding method in image segmentation which uses the Otsu-s method to segment the image. This method iteratively searches for the sub-regions of the images, instead of treating the full image as a whole region for segmentation. First the Otsu-s method is applied on the input image to find the Otsu-s threshold and compute the mean values of two classes as separated by the threshold and based on this the image is divided into three sub-region instead of two as the standard Otsu-s method does. The three classes are determined as foreground, background and -to-be-determined- [TBD] region. The foreground and background are not included for further processing. The Otsu-s method is again applied on the TBD region to find the new threshold value and the two class mean, based on this the TBD region is again divided into three sub-regions. The iteration stops when the Otsu-s threshold calculated between two successive iterations is less than that of the pre-set threshold. Finally all the previous foreground and background is combined to form the final output. The obtained output will achieve better result than the standard Otsu-s method.