Deep learning for tumour classification in homogeneous breast tissue in medical microwave imaging

Deep learning has become the state-of-the-art in the area of biomedical imaging, leading to a large boost in performance that approaches human levels. Medical microwave imaging is an emerging technology that has great potential especially in the area of breast cancer diagnosis. Moreover, the obtained backscatter signals have also been shown to be a good basis for differentiating malignant and benign tumour type. We further analyse these results by applying deep learning methods to a dataset of Finite Difference Time Domain (FDTD) numerical simulations of tumour models embedded in homogeneous breast adipose tissue. Specifically we use Deep and Convolutional Neural Networks and obtain an accuracy of 93.44% which outperforms conventional machine learning previously used on the analysed dataset.