Algorithm for automatic threshold determination for image segmentation

This paper presents a new algorithm for automatic determination of threshold for image segmentation. The algorithm is based on the histogram of extremum (maximum or minimum) gray levels in a collection of subimage areas. The image needs to be corrected for uneven illumination prior to the application of the algorithm. The algorithm was tested on a large class of images and performed well for all cases tested. The algorithm is robust, flexible, and computationally efficient. >