An Adaptive Gabor Filtering Method and Its Application in Edge Detection

Low contrast and fuzzy edge are usually exist in the tissues such as liver, stomach, spleen, muscle in abdominal CT image, which makes the segmentation of these organs to be one of the most difficult tasks in medical image CAD system. Based on the fact that the imaginary part of Gabor function is an effective and robust edge detector, an adaptive Gabor filtering method is proposed in this paper, which uses its imaginary part as a detector to extract the edge in abdominal CT image. In the implementation, we first calculate the direction of each pixel in the image, then evaluate the center frequency according to the statistical characteristics of gray values of local pixels, and at last use the imaginary part of Gabor function to detect the edge of abdominal CT image. Experimental results show that the method we proposed is effective and robust.