Proposed intelligent learning algorithm to improve the diagnosis and classification of breast cancer masses

Abstract In today's world, intelligent medical diagnostic systems, with their remarkable advances, have greatly contributed to the medical world and have accelerated and increased the accuracy of many diagnoses, including the system for diagnosing and classifying breast cancer masses. Unfortunately, breast cancer is one of the most dangerous diseases and has led to many deaths among women. Early detection can, firstly, increase treatment options and, secondly, increase life expectancy in women with earlier treatment. Early detection methods include mammography, MRI and ultrasound. Early detection methods for breast cancer, include mammography, MRI, and ultrasound. In all diagnostic methods, image processing and artificial intelligence methods can be significantly effective in the process of noise detection and reduction, tumor area detection and classification. In this paper, first we reduce the noise in mammographic images using Directional filters, then extract the tumor area by combining the support vector machine and the Evolutionary Algorithms, and finally, using deep learning, To classify the type of mass. The proposed method has better performance compared to previous methods and offers 99.1% accuracy, 98.4% sensitivity and specificity with 100%. in diagnosis and classification.