New Method for Automatic and Precise Edge Extraction from Slimi Images

ABSTRACT: Slimi is a kind of schemes made of arcs that was ripped in Islamic Era. All of the Slimi curves have inside vectors and outside vectors. Theses vectors are essence of Slimi and are the sign of infinity and eternity. Slimi is a design including of beautiful rounds and these rounds will be completed with appropriate subsets. Slimi is one of the seven Iranian traditional paintings. In extraction of schemes boundaries from Slimi images, we were faced to edge extraction. The problem was quality of images that were taken by typical cameras and images had inseparable amount of noises. Noise affectedthe boundaries between regions and so edge detection methods did not work very well. In this article, we developed a region-based segmentation method for region boundary extraction. We used Level Set Method to extract every kind of schemes region and to eliminate of noises. After using proposed method on test images, a comparison between this method and usual methods for edge extraction was done in this research. Results showed that the proposed method had a higher completeness and precision in comparison to edge detection methods like Canny Method.