Histopathological Image-Based Breast Cancer Detection Using Deep Learning Models

Breast cancer exists as a widespread disorder within the female population because its prompt detection enables successful care and reduces mortality statistics. The integration of deep learning into breast cancer research through improved traditional machine learning tools occurred because of extensive research that employed decision trees and KNN and support vector machines algorithm with naive Bayes. The major advancement in image analysis results from CNNs because these architectures allow users to extract complex semantic information from raw data. High-quality results from breast cancer diagnosis and classification emerge from implementing state-of-the-art ResNet, Inception V3 and VGG19 CNN architectures. Two distinct classification types located in the primary dataset serve as the base elements of the research study: Class type 0 designs the malignant category and Class type 1 signifies benign tumors. Ultrasound images supplied multiple details that enabled the models to successfully classify between benign and malignant tumors. Among three CNN architecture models ResNet Inception v3 and VGG19 provide excellent performance in detecting and classifying breast cancer from sonographic images. Modern CNN architecture models applied to sonographic data yield top diagnostic outcomes that improve opportunities for breast cancer screening and diagnosis enhancement.