An Improved Breast Cancer Detection Based on Convolutional Attention Mechanism

Computer Science

Breast cancer can effectually be treated through its early detection. As such, the availability of proper screening methods is important for detecting the initial symptoms of breast cancer. Various imaging techniques are used for the screening to identify this disease, the popular ones of which include: mammography, ultrasound, and thermography. This project proposes a model using ResNet50 Convolutional Neural Network to detect breast cancer from mammograms, using Convolutional Block Attention Module (CBAM) as an attention mechanism. A varying number of blocks is used to check for optimality, among which the most optimal was identified. Experimental results on MATLAB 2021a show that the proposed model attains the most optimal performance for the case of 12 CBAM blocks with a less overfitting rate. Thus, this suggests that by reducing the number of CBAM blocks, the model can attain an optimal performance without compromising the training and testing accuracies, proving the assertion made by previous research. Further study will employ a regularization technique to address overfitting

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.