Breast Density Classification Using Convolutional Neural Networks and Analysis of the CLAHE Technique

Breast cancer is the most common type of cancer and one of the leading causes of death worldwide. Early detection is essential to increase the chances of successful treatment, but current conventional methods still have limitations. Artificial intelligence has shown promise in aiding diagnosis. The project therefore seeks to use Convolutional Neural Networks to classify breast density, which is an important factor in the diagnosis of breast cancer. These networks learn automatically from raw data, classifying more accurately. The convolution operation and pooling layers are used to process the inputs and extract complex features. Fully connected layers classify and identify features. The dataset used was a combined dataset between INbreast and RSNA, pre-processed and using the CLAHE technique to increase the contrast of the mammograms. The end result was an accuracy of 89.29% for ResNet-50 and ShuffleNet.