Ultrasound Breast Cancer Classification

Breast cancer is the most common cancer in women. To assist in the early detection and diagnosis of breast cancer to improve patient survival rate, this algorithm uses a deep learning technique to simultaneously predict the BI-RADS categories and malignant classifications based on breast ultrasound images.

A convolutional neural network algorithm was developed to predict and classify breast cancer using ultrasound images. Each image study will be assigned labels of the corresponding BI-RADS categories from 0-6 ,and malignant classification (malignant tumor, benign lump, or normal tissue).

  • Simultaneously predicts the BI-RADS categories and malignancy levels of breast cancer
  • Highlights suspected image regions

Invasive Cancer Detection Utilizing Compressed Convolutional Neural Network and Transfer Learning

Kong B, Sun S, Wang X, Song Q, Zhang S.

International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI), 2018.