An Experimental Investigation of Classifying Breast Cancer Using Different CNN Models
Hirenkumar Kukadiya, Divyakant Meva · International Journal of Computer Applications · 2025
Breast cancer continues to be one of the most common cancers that affect women worldwide.Improving patient outcomes requires an early and precise diagnosis.The performance of many Convolutional Neural Network (CNN) designs for breast cancer image categorization is compared experimentally in this publication.We tested a number of cutting-edge CNN models, such as VGG16, ResNet50, DenseNet121, EfficientNet, and MobileNet, using a number of publically accessible datasets related to mammography and breast cancer histology.According to our tests, EfficientNet-B3 demonstrated the best trade-off between computational efficiency and performance, while DenseNet121 obtained the highest overall accuracy (94.8%) and F1-score (0.937).Additionally, we suggest a brand-new ensemble method that leverages the advantages of several CNN designs, improving classification accuracy by 2.3% over the top-performing single model.Our results offer important new information for the practical application of deep learning algorithms for the diagnosis of breast cancer.