Main Article Content
Breast cancer is the leading cause of common cancer among women. Automated screening approaches may save time and minimize errorsin detecting and categorizing breast cancer subtypes, a crucial clinical activity. Breast cancer is diagnosed using the biopsy technique, which involves examining tissue samples under a microscope. Senior pathologists should analyze breast cell morphologies in histopathology images to determine this type of cancer. The world's population of pathologists is insufficient, and human error in diagnosing procedures is possible. Analysis of histopathology images using deep learning algorithms can aid pathologists in identifying cancer subtypes and making a better treatment plan. As a result, this research presents BRECNET (Breast Cancer Network), a dedicated architecture that employs a parallel convolution filter for screening breast cancer from histopathology images. In addition, to avoid overfitting and create high levels of reliability, a variety of augmentation procedures were implemented to improve the number of histopathological images. Based on histopathological image analysis, the proposed system was assessed and shown to have an accuracy of 87.25%and a kappa score of 85.40% in classifying eight subtypes of cancer with different magnification levels. The results show that the BRECNET model is much more successful and efficient, making them more acceptable for breast cancer screening.