 The proposed breast ultrasound image-based assisted diagnosis method based on convolutional neural networks improves the diagnostic speed and early screening rate of breast cancer by identifying tumor location in size using attention-based semantic segmentation, cropping identified nodules to construct a training dataset, and training a convolutional neural network for benign malignant breast nodule diagnosis. The system provides a valuable aid for the ultrasonic diagnosis of breast cancer and is superior to junior doctor's diagnostic performance. This article was authored by Lei Yong, Baichuan Zhang, Fei Ren, and others.