Early detection of breast cancer in histopathology images employing convolutional neural network (CNN)

• Published in: AIP conference proceedings Vol. 2453; no. 1
• Main Authors: Hilaliyah, Putri Khalifa, Irfan, M., Lestandy, Merinda
• Format: Journal Article Conference Proceeding
• Language: English
• Published: Melville American Institute of Physics 25.07.2022
• Subjects: Acoustics; Artificial neural networks; Breast cancer; Classification; Deep learning; Heredity; Histopathology; Machine learning; Medical imaging; Neural networks; Object recognition
• ISSN: 0094-243X; 1551-7616
• DOI: 10.1063/5.0094608

Breast cancer is currently acknowledged as a serious disease that affects both men and women. Breast cancer could develop as a person ages, either as a result of an unhealthy lifestyle or due to heredity-related factor. As a result, early detection towards the presence of cancer is required for early prevention. Upon considering its enormous skills in modeling complex data such as images and sounds, deep learning has served as one of the most debatable subjects in the area of Machine Learning. The Convolutional Neural Network involves the Deep Learning approach that offers the best results in image recognition (CNN). Breast cancer diagnosis in this study was conducted by utilizing the Convolutional Neural Network (CNN) approach, which has proven to be accurate in various investigations, especially regarding the picture data. This study employed histopathology images data from Kaggle in the form of breast histopathology images. The total number of bytes is 549 further utilized to import data, examine data, preprocessing, define helper functions for the classification task, and assess classification models as the five stages of the research method. Implementing a modified CNN model, researchers were able to detect breast cancer in histopathology images with an accuracy of 80%.