Machine Learning Techniques for Autonomous Lesion Detection in Microwave Breast Imaging Clinical Data

Microwave breast imaging is becoming a promising new approach in the early detection of breast cancer. In this study, we have used advanced machine learning techniques to identify breast lesions from clinical data obtained via MammoWave scanner. To improve the performance of our model, we have explored refinements such as adjusting the number of principal components (PCs) in principal component analysis (PCA) and transitioning to generative adversarial network (GAN)-generated raw data. Leveraging this data, we have used PCA feature extraction, and utilized it to train a support vector machine (SVM). Our analysis has yielded promising results, with a ternary classification approach achieving an accuracy of 0.80.