Executive Summary : | Cancer has become a common feared word in recent years, and research on mass detection and classification systems has been conducted for the last two decades. However, due to the large number of false positives identification, most of these systems cannot be used in commercial clinical practices. Digital Radiography (DR) mammograms offer better contrast and clarity, but image interpretation processing time is comparatively high with conventional image processing algorithms. The primary objective of this research project is to develop a novel and efficient two parameter and three parameter based logistic type model with efficient algorithms-based mass detection and classification system. The proposed system is divided into five phases: pre-processing of DR images, segmentation to get region of interest (ROI), feature identification and extraction using statistical modeling-based decision-making system, and classification of features and decision making using Two-Parameter & Three-Parameter Logistic Type Distribution With U-NET+ Models technology. The Two-Parameter & Three-Parameter Logistic Type Distribution With U-NET+ Models mechanism classifies abnormalities into benign or malignant based on various parameters of intensity, texture, size, and shape in the environment of increased pixel entropy of DR mammogram images. Most conventional classifiers use limited parameters, producing greater false positive cases. The novelty of this project is the application of Two-Parameter & Three-Parameter Logistic Type Distribution With U-NET+ Models to classify mass, reducing false positive cases. The proposed system framework has been developed in consultation with medical practitioners and radiologists and can be applied as a second opinion in diagnostic radiology. In conclusion, the proposed system can be implemented as part of biomedical telemetry systems to improve healthcare and reliability. |