Ultrasonic breast tumor automatic segmentation method based on attention-enhanced U-shaped network

Abstract

The invention discloses an ultrasonic breast tumor automatic segmentation method based on an attention-enhanced U-shaped network. The method comprises the following steps: (A) constructing an attention-enhanced U-shaped network model for ultrasonic breast tumor automatic segmentation; step (B), establishing a mixed attention loss function of the attention enhancement U-shaped network model; and step (C), according to the attention-enhanced U-shaped network model and the mixed attention loss function, training the attention-enhanced U-shaped network model through a coarse and fine combination strategy to achieve segmentation of the breast ultrasound image lesion area. The method can be used for extracting the focus area of the mammary gland ultrasonic image, can effectively improve the accuracy of mammary gland tumor segmentation, is used for assisting a doctor in quickly and accurately positioning the focus area, reducing the workload of the doctor and relieving the defects of insufficient clinical experience and the like of young doctors, and has very important research value and application prospect for modern medicine.

Description

Technical field[0001]The present invention relates to the technical field of ultrasonic breast tumor detection, in particular to a method for automatically segmenting ultrasonic breast tumors based on an attention-enhanced U-shaped network.Background technique[0002]At present, breast cancer is a disease that seriously endangers women's health after skin cancer. With the development of modern medicine, early diagnosis and treatment can greatly improve the survival rate of breast cancer patients. Currently, the diagnosis of breast tumors can be divided into invasive and non-invasive diagnosis. Invasive diagnosis mainly refers to biopsy, but this will cause physical damage to breast tissue and cause pain to the patient; non-invasive diagnosis refers to the use of X-ray, MRI (magnetic resonance imaging) or ultrasound imaging to examine breast lesions Among various inspection methods, ultrasound imaging has become the first choice for early diagnosis of breast tumors due to the advantage...

AI Technical Summary

Problems solved by technology

However, due to the lack of training data, the low contrast of ultrasound images, the mutual invasion between suspicious lesions and surrounding tissues, and the diversity of tumor shapes, the research on ultrasound breast tumor region segmentation based on deep neural networks is still in its infancy. , the segmentation effect still needs to be further improved

Images