Computer-aided diagnosis of breast cancer from mammogram images using deep learning algorithms

• Published in: Journal of Electrical Systems and Information Technology Vol. 11; no. 1; pp. 38 - 16
• Main Authors: Dada, Emmanuel Gbenga, Oyewola, David Opeoluwa, Misra, Sanjay
• Format: Journal Article
• Language: English
• Published: Berlin/Heidelberg Springer Berlin Heidelberg 01.12.2024; Springer Nature B.V; SpringerOpen
• Subjects: Accuracy; Algorithms; Artificial Intelligence; Artificial neural networks; Breast cancer; Classification; Computer-aided diagnosis (CAD); Datasets; Deep learning; Diagnosis; Electrical Engineering; Electrical Machines and Networks; Engineering; Image classification; Image processing; Information Systems and Communication Service; Lymphatic system; Machine learning; Mammography; Mechanical Engineering; Medical imaging; Power Electronics; Simulation; Tumors; Deep learning; Computer-aided diagnosis (CAD); Breast cancer; Medical imaging; Classification
• ISSN: 2314-7172; 2314-7172
• DOI: 10.1186/s43067-024-00164-y

Even though accurate detection of dangerous malignancies from mammogram images is mostly dependent on radiologists' experience, specialists occasionally differ in their assessments. Computer-aided diagnosis provides a better solution for image diagnosis that can help experts make more reliable decisions. In medical applications for diagnosing cancerous growths from mammogram images, computerized and accurate classification of breast cancer mammogram images is critical. The deep learning approach has been widely applied in medical image processing and has had considerable success in biological image classification. The Convolutional Neural Network (CNN), Inception, and EfficientNet are proposed in this paper. The proposed models attain better performance compared to the conventional CNN. The models are used to automatically classify breast cancer mammogram images from Kaggle into benign and malignant. Simulation results demonstrated that EfficientNet, with an accuracy between 97.13 and 99.27%, and overall accuracy of 98.29%, perform better than the other models in this paper.