A New Framework for Canny Edge Detector in Hexagonal Lattice

Abstract

Feature extraction is widely used in image processing applications such as face recognition, character recognition, fingerprint identification and medicine. Edge features is among the most important features for such applications. Canny edge detector is the most popular one and has many benefits in comparison with other methods. Since pixels in hexagonal domain have many benefits in comparison with square domain, this paper presents an efficient Canny edge detector in hexagonal domain. The proposed method includes square to hexagonal transformation and edge detection based on a new algorithm. The proposed method has been evaluated on synthetic and real image datasets with different signal to noise ratios (SNRs). Detected edges in synthetic images show that the proposed hexagonal edge detector outperforms existing methods in 44 cases out of 60 cases with respect to figure of merit (FoM). Finally, results of real images demonstrate the superiority of the proposed method in qualitative analysis of sub-images.