A framework for breast cancer classification using Multi-DCNNs

• Published in: Computers in biology and medicine Vol. 131; p. 104245
• Main Authors: Ragab, Dina A., Attallah, Omneya, Sharkas, Maha, Ren, Jinchang, Marshall, Stephen
• Format: Journal Article
• Language: English
• Published: United States Elsevier Ltd 01-04-2021; Elsevier Limited
• Subjects: Accuracy; Architecture; Artificial neural networks; Breast cancer; Classification; Classifiers; Computational efficiency; Computer applications; Computing costs; Datasets; Deep convolutional neural networks; Deep learning; Digital imaging; Experiments; Feature extraction; Image analysis; Image classification; Image processing; Image retrieval; Kernel functions; Learning algorithms; Machine learning; Mammography; Medical imaging; Medical research; Methods; Neural networks; Performance evaluation; Principal component analysis; Principal components analysis; Support vector machines; Deep convolutional neural networks; Support vector machines; Machine learning; Principal component analysis
• ISSN: 0010-4825; 1879-0534
• DOI: 10.1016/j.compbiomed.2021.104245

Deep learning (DL) is the fastest-growing field of machine learning (ML). Deep convolutional neural networks (DCNN) are currently the main tool used for image analysis and classification purposes. There are several DCNN architectures among them AlexNet, GoogleNet, and residual networks (ResNet).
This paper presents a new computer-aided diagnosis (CAD) system based on feature extraction and classification using DL techniques to help radiologists to classify breast cancer lesions in mammograms. This is performed by four different experiments to determine the optimum approach. The first one consists of end-to-end pre-trained fine-tuned DCNN networks. In the second one, the deep features of the DCNNs are extracted and fed to a support vector machine (SVM) classifier with different kernel functions. The third experiment performs deep features fusion to demonstrate that combining deep features will enhance the accuracy of the SVM classifiers. Finally, in the fourth experiment, principal component analysis (PCA) is introduced to reduce the large feature vector produced in feature fusion and to decrease the computational cost. The experiments are performed on two datasets (1) the curated breast imaging subset of the digital database for screening mammography (CBIS-DDSM) and (2) the mammographic image analysis society digital mammogram database (MIAS).
The accuracy achieved using deep features fusion for both datasets proved to be the highest compared to the state-of-the-art CAD systems. Conversely, when applying the PCA on the feature fusion sets, the accuracy did not improve; however, the computational cost decreased as the execution time decreased.
•Proposing a multi-DCNN based CAD system for breast cancer classification.•Deep features fusion had an influence on the classification accuracy.•Searching for the best combination of deep features.•Selecting the number of principal components with a sequential forward method.•The computer-aided diagnosis system proposed proved to be reliable.