Detection of brain tumor contours by using edge detection method over intensity versus perceived brightness gradients

Article type: Research Article

Authors: Dirican, Bahar | Oysul, Kaan | Pahsa, Alper | Beyzadeoǧu, Murat

Affiliations: Radiation Oncology Department Gulhane Military Medical Academy, Ankara, Turkey

Note: [] Corresponding author. E-mail: bahardirican@yahoo.com

Abstract: In this work, a new contour detection method is studied for detecting brain tumor regions based on their gradient magnitude information. Gradient magnitude data, an edge detection method, is generated from the brain slice image intensity or perceived brightness information. Contour map of the brain tumor is generated by using the gradient magnitude differences of the template masks (cropped from brain slice tumor image) and the sample masks (traverses the image) raw pixel and perceived brightness (luminance) date. Then these differences are averaged and normalized to produce edge profiles of the brain tumor region contours. This data is used by the remote surgical devices for removing the tumor area.

Keywords: Perceived brightness (luminance), gradient magnitude, contour profiling, edge detection

Journal: Journal of X-Ray Science and Technology, vol. 16, no. 1, pp. 1-8, 2008