Enhancement of CT pancreatic features by a simple cascading filter

• Published in: IEEE, PISCATAWAY, NJ, (USA). 1994
• Main Authors: Tsai, Du-Yih, Kojima, Katsuyuki
• Format: Conference Proceeding
• Language: English
• Published: 01.01.1994
• ISBN: 9780780314870; 0780314875

A simple method for detection of subtle abnormalities in CT pancreatic images is proposed. The algorithm of the proposed spatial filtering consists of three modules that are applied to the original CT images in sequence. In the first step a square-of-logarithm operator is introduced to enhance the edges with low gray levels. Then a gray-level shifting to remove the blurred area on the image is applied in the second step. In the final step a logarithmic operator is used to enhance the contour of subtle details. This procedure brings out features on the image with little or no enhancement of the image noise. Two CT pancreatic images with lesion subtlety are selected for experiments. The performance of the method presented is compared to that of four other commonly used techniques. Both visual and quantitative results reveal that our method is superior. Recent work to extend the method to other organs such as the liver in abdomen is also encouraging.