Tumor Segmentation and Related Analysis on Breast Ultrasound Image

Tumor Segmentation and Related Analysis on Breast Ultrasound Image

The purpose of this paper is to deal with breast cancer on ultrasound image. Through the region of interest (ROI) on the selected image, we can obtain the tumor in ROI, and advance to process the preliminary determination of benign and malignant for breast tumors. Benign tumor on breast ultrasound image presented a relatively smooth edge and stable content, and also can be moved with elastic deformation. Malignant tumor has irregular edges, unstable content, and sound attenuation. Based on the different characteristics of benign and malignant on breast ultrasound images, we can design an image processing procedure, which combined MeVisLab software, to set a reasonable threshold for the tumor region on breast ultrasound images. Then, we can apply the technology of image segmentation of other image processing methods into the breast ultrasound. We hope that this information technology can be used in clinical experiments, and to help determine breast tumors which is benign or not.