Diagnosis of Breast Cancer Based on Hybrid Features Extraction in Dynamic Contrast Enhanced Magnetic Resonance Imaging

Breast cancer develops in breast cells. It is the most common type of cancer in women and the second most lethal disease after lung cancer. The presence of breast masses is an important symptom for detecting breast cancer in its early stages. This study proposes a hybrid features extraction method to improve the automatic detection of breast cancer by combining three feature extraction methods: Kinetic Features, convolutional neural network (CNN) deep learning features, and the newly proposed quantum-Chebyshev polynomials model (QCHPs). The long short-term memory (LSTM) model is used as a classifier in this study to detect breast cancer automatically, which could reduce human errors in the diagnosis process. The experimental results using a large publicly available dataset achieved a detection accuracy of 99.50% for hybrid features in post-contrast 2, potentially reducing human errors in the diagnosis process.