Early Breast Cancer Detection through Optimized Ultrasound Elastography

Ultrasound elastography is a biomedical technique used to assess tissue stiffness and musculoskeletal disorder and for distinguishing benign from malignant conditions in this case of breast tissue. This study aims to enhance the diagnostic accuracy by integrating unsupervised machine learning with hybrid models. A customized Convolutional Neural Network (CNN) is employed to analyze elastography data, enabling automated pattern recognition and classification. By leveraging the model's ability to learn from labelled data, the system reduces reliance on manual annotations, improving efficiency and precision. We are trying to integrate a customized CNN0 from the ref paper with one of the preexisting architectures which when implemented alone has apparently given 91% accuracy. The hybrid approach combines the strengths of elastography and machine learning to optimize diagnostic performance, offering a potential breakthrough in non-invasive cancer diagnosis and personalized healthcare.