Eigen value for image segmentation

Abstract EN

In this paper eigenvalue of a gray image is used for image segmentation by calculating the eigenvalue for each segment and to be compared with a threshold value [which will be calculated dynamically by the system] and an iteration procedure will continue running until saturation arrived.

By using the quadtree technique the gray image is divide into small segments depending on the value of the threshold and then the segmented image will pass into low pass filter in order to produce more enhancement in the image.

The algorithm was applied on different types of images [natural images, text images] and both of them give high quality of segmentation.