Deep Learning in Grading Prediction of Breast Cancer Ultrasound Images

Abstract

This study proposed a breast cancer ultrasound image grading prediction method based on multi-layer perceptron (MLP), aiming to improve the accuracy and efficiency of breast cancer automated diagnosis. Early diagnosis of breast cancer is crucial to the prognosis of patients, and ultrasound imaging, as a non-invasive examination method, is widely used in breast cancer detection. In order to adapt to the characteristics of breast cancer of different grades, this study extracted features from ultrasound images and converted the texture, edge and shape information of the image into feature vectors suitable for MLP model processing. Experimental results show that the MLP model performs well in evaluation indicators such as accuracy, recall, precision and F1 score, showing its potential in breast cancer grading prediction tasks. Comparative experiments with other deep learning models further verified the efficiency and robustness of MLP. This study provides reliable theoretical and experimental support for breast cancer ultrasound image grading prediction based on MLP, and provides a new technical solution for early diagnosis and grading management of breast cancer.