ROMJIST Volume 19, No. 4, 2016, pp. 338-359, Paper no. 542/2016
Anand Prakash SHUKLA Training Cellular Automata for Image Edge Detection
ABSTRACT: Cellular automata can be significantly applied in image processing tasks. In this paper, a novel method to train two dimensional cellular automata for detection of edges in digital images has been proposed and experiments have been carried out for the same. Training of two dimensional cellular automata means selecting the optimum rule set from the given set of rules to perform a particular task. In order to train the cellular automata first, the size of rule set is reduced on the basis of symmetry. Then the sequential floating forward search method for rule selection is used to select the best rule set for edge detection. The misclassification error has been used as an objective function to train the cellular automata for edge detection. The whole experiment has been divided in two parts. First the training was performed for binary images then it is performed for gray scale images. A novel method of thresholding the image by Otsu's algorithm and then applying the cellular automata rules for the training purpose has been proposed. It has been observed that the proposed method significantly decreases the training time without affecting the results. Results are validated and compared with some standard edge detection methods both qualitatively and quantitatively and it is found better in terms of detecting the edges in digital images. Also the proposed method performs much better in corner detection as compared to the standard edge detection methods.KEYWORDS: Cellular Automata, Training of Cellular Automata,Sequential Float-ing Forward Search Algorithm, Misclassification Error, Otsu’s Algorithm, CornerDetectionRead full text (pdf)