Abstract: A novel edge detection method for medical image is proposed for the characteristics of the medical images and drawback of classical method. The characters of ideal image edge of 3×3 mask are considered and two pixel sets in the mask are used to construct an objective function. By distance measure, the values of the objective function corresponding to four directions can be determined and the edge intensity is computed. Then a max value of each pixel in the mask can be obtained. Through comparing with threshold given and by applying nonmaxima suppression method, the pixel can be judged that it is or not an edge pixel. At last, the edge points of the image are extracted by doublethreshold connecting. Experimental results show that the proposed algorithm performs well in terms of the ability to detect the edge and noise. The final edge image is precisely localized with single pixel.

Key words: edge detection, maximum objective function, nonmaxima suppression