Ear Identification Method Based On Gray Surface Matching

Ear recognition is a new technology of biologic recognition, which has been paid attention to by more and more scientists because of ear's unique features and applying range. At present, the domestic research institutions and staff are few in this area and other countries are still be in preliminary and exploratory stage. So it has tremendous potential and broad application and development of space. Ear recognition not only a beneficial supplement for other biometric recognitions such as fingerprint recognition and face recognition, but also solely used at some occasion.Now ear recognition doesn't form an integrated theory system. It depends on traditional recognition technology and theory belonging to recognition field. But ear is a special biological body, has the inherent physical structure and location, will have its own unique approach and with suitable identification technology. Here the paper chiefly studies and researches pretreatment of images, the edge detection of ear recognition and gray surface matching.Outer edge of ear detection is an important step in the ear recognition. According to the intrinsic figure of the ear, this paper proposes a method based on edge-tracking for human outer edge of ear. Pretreated the collection of the images, binary images were generated using an iterative method which can get a divided threshold. Then ear images can be divided into four regions for tracking according as its shape. Through an association of tracking in binary image and in gray image get an integrated outer edge of ear. Finally, separated region of the ear from the gray image.Since extracted the region of the ear, an ear recognition method based on gray surface matching has been proposed. Firstly, gray and size of the ear images are adjusted. Secondly, it calculates the gray difference between two ear images and gets the gray difference surface. Finally, it computes the variance of the gray difference surface and decides whether these two ear images are in the same ear or not. This method abandons the traditional operations including the feature extracting and encoding, but uses the theory of gray surface matching based on the whole feature analysis. Simultaneously it meets the need of invariance to translation, scaling and rotation.The results indicate that the algorithm of edge-tracking for human ear outer edge is feasible, and it achieves a remarkable effect using the gray surface matching algorithm on human ear recognition.