Segmentation of ultrasonic images using Support Vector Machines

Support Vector Machines (SVMs) are a general algorithm based on guaranteed risk bounds of statistical learning theory. They have found numerous applications, such as in classification of brain PET images, optical character recognition, object detection, face verification, text categorization and so on. In this paper we propose the use of SVMs to segment lesions in ultrasound images and we assess thoroughly their lesion detection ability. We demonstrate that trained SVMs with a radial basis function kernel segment satisfactorily (unseen) ultrasound B-mode images as well as clinical ultrasonic images. (C) 2002 Elsevier Science B.V. All rights reserved.