Abstract: Breast cancer remains a critical global health challenge where early and accurate diagnosis is essential for improving survival rates. Conventional diagnostic approaches are often limited by subjectivity and inconsistent accuracy, demanding the need for reliable computer-aided diagnostic systems. This study examined the use of the EfficientNetV2-B0 model, for breast cancer prediction using histopathological images. The images used in this study were obtained from the BreakHis database, which contains images of biopsies at varying levels of magnification that includes both benign and malignant samples. Images underwent comprehensive preprocessing, including size reduction, normalization, and application of augmentation techniques and random cropping to improve generalization. The dataset was divided into a training (70%) and testing (30%) set and all known benign cases were oversampled to eliminate class imbalance between the benign and malignant cases. EfficientNetV2-B0 was utilized to extract features and generate predictions. Additionally, ResNet50 and DenseNet121 were also used in this study for comparison purposes with EfficientNetV2-B0. Model training and evaluation were conducted in Python 3.11 using TensorFlow and Keras. Experimental results demonstrate that the EfficientNetV2-B0 model achieved performance of 99% prediction accuracy on the test set with precision of 100% and recall of 99%. In contrast, ResNet50, and DenseNet121 achieved lower accuracies of 89% and 94% respectively. The results demonstrated the robustness and diagnostic reliability of EfficientNetV2-B0 for breast cancer prediction over existing models. The EfficientNetV2-B0 model is therefore, recommended for integration into clinical workflows to assist pathologists, minimize diagnostic errors, and improve patient outcomes through timely and accurate detection
Keywords: Breast cancer prediction, histopathological images, efficientNetV2-B0, transfer learning, deep learning.
Title: A Pre-Trained EfficientNetV2-B0 Model for Breast Cancer Prediction
Author: Samuel Oluwatayo Ogunlana, Olanrewaju Raymond Adeniyi, David Adeola Akinwumi
International Journal of Computer Science and Information Technology Research
ISSN 2348-1196 (print), ISSN 2348-120X (online)
Vol. 14, Issue 2, April 2026 - June 2026
Page No: 159-168
Research Publish Journals
Website: www.researchpublish.com
Published Date: 27-May-2026
