BreastCNN: A Novel Layer-based Convolutional Neural Network for Breast Cancer Diagnosis in DMR-Thermogram Images

Breast cancer is one of the most prominent sources death in females. Every year many women suffer breast cancer, and, end, occurs. The early detection may cause to reduce rate and save women’s lives. medical care cost prevention are costly become a priority diagnose at its stages. Initially, mammography technique was leading detect stage cancer. However, it cannot deal with tumor size less than 2 mm. To overcome this challenge, by considering DMR-thermogram images, novel layer-based Convolutional Neural Network (BreastCNN) for classification proposed. BreastCNN method works five different layers uses types filters. learning structures change after every convolution layer. proposed tested on Database Mastology Research (DMR) having 745 healthy 261 sick images. performance calculated as statistical values known sensitivity, specificity, precision, accuracy, F1-score. shows better accuracy 99.7% related already presented methods.