A NOVEL APPROACH TO BREAST TUMOR CLASSIFICATION IN ULTRASOUND IMAGES BASED ON THE IMPROVED INCEPTION-V3 MODEL

Authors

Research and Development Department, Med Fanavarn Plus Co., Karaj, Iran

10.22034/icrj.2022.173654

Abstract

One of the most significant reasons for women’s decease is breast cancer. So, swift diagnosis and proper remedy have conspicuous roles in breast cancer control and its rate reduction among women.  Reliable and precise detection is highly critical to specialists. As mammography and ultrasound (US) imaging are two prevalent screening modalities for breast anomalies detection, the mass of physicians use them. However, the procedure of mammography is time-consuming and painful, contrary to ultrasound imaging. But it’s more authentic. In this study, by intending to assist doctors, we have utilized the improved InceptionV3 structure to classify ultrasound breast lesions into two classes: malignant and benign.