Universal Image Segmentation Technique for Cancer Detection in Medical Images

Abstract

Image segmentation is key area of research and focus due to the complexity of the problem statements. Based on the application and source of images the parameters of processing the image as well as the techniques employed to get efficient segmentation varies greatly. The works in this field thus far have been on a particular modality and utilized a sinlg e or combination of techniques for improving segmentation for those modalities.This paper focuses on a universal method of identification of cancerous regions in medical images captured from different image sources such as Computerized Tomography (CT) Scan, Magnetic Resonance Imaging (MRI), X-Ray and Ultrasound. The parameters utilized in this model have been optimized so as to avoid any and all human intercention in the identification of cancer in medical images. The proposed method employs a double mask technique in order to autoidentify region of interest without any manual intervention.