Breast Cancer Detection on Mammographic Image using Convolutional Neural Network

Breast cancer had led the prevalence cancer cases in Indonesia and it had become the fourth leading cause of death worldwide in 2020. The risk can be minimized by regularly conducting early screenings. This study implements the Deep Convolutional Neural Network (CNN) method to build an artificial intelligence model for early detection of breast cancer on mammography images. Using 2400 training data in total and 120 test data for testing which each contains 50% normal and cancer images. The proposed CNN Architecture has 3 layers which implement average and max pooling layer. The training reach 100% of maximum training accuracy and 91.67% of maximum validation accuracy. The selected model is used to classify new unseen dana and it produces 60.83% accuracy which achieved by using optimizer adam, 0.001 learning rate, and 10 epochs. The result was obtained after several parameter training using epoch, learning rate, layers, optimizer, and adjustment to the dataset.