Brief Introduction

Breast cancer is a common malignancy in women worldwide. For Chinese women, ultrasonography is of great value in screening for breast cancer and is the most common screening method as well. However, there are problems such as fuzzy diagnosis and operator dependence in ultrasonography, which is far from the basic concept of precision medicine.

In this paper, we attempt to design a novel system that precisely detect the nodule position in ultrasound image and accurately classify it into benign or malignant. By combining features from there different modalities of ultrasound images, the performance of our system should be better than state of the art. Furthermore, for those malignant nodules in diagnosis, we attempt to design a novel type of neural network in order to predict their pathological index, which helps to determine the prognostic evaluation.

Though, this project has just started, we are convinced that our team can achieve outstanding results since we are now moving in the right direction.