Interactive medical image segmentation using active contour with improved F energy in level-set tuning

Image segmentation is a fundamental step in medical image processing (MIP) and has been broadly considered and improved to clarify clinical interpretation and utilization. This paper presents a robust and efficient method for the segmentation of doubtful masses or tumors in medical images. As diagnosing tumors in mammography, magnetic resonance imaging (MRI), and computed tomography (CT) images requires precision, experience, and time, and we have proposed an algorithm designed by Active Contour (AC) and enhanced Level-set techniques. By using the active contour technique, the energy function, the Level-set method, and the proposed F function, the segment of the cancerous masses is derived. By implementing the algorithm on ۱۶۰ images from ۲ databases, including ۸۰ mammograms images and ۸۰ MRI brain images, the evaluation results demonstrate that accuracy, recall, and precision are ۹۶.۲۵%, ۹۵.۶۰%, and ۹۵.۷۱%, respectively, for segmenting the suspicious segments. Implementing this system on tissue imaging devices enhances the accuracy in diagnosing relatively high-volume images that are significantly processed with high speed. Reducing costs, saving time and high precision are particular advantages of the system that outrank the system from other similar methods

Keywords:

Image processing , medical image , Highboost , Active contour , Levelset