The Study of Mammary Disease CAD System

An X-ray image CAD (computer aided diagnosis) system aiming at oriental female´s dense mammary gland is proposed. The system acquires assistant diagnosis result through the synthetic analysis of both direct and indirect symptoms of the mammary disease. In the detection and decision of the direct symptom, the system adds features of micro-calcification shape description and mass margin roughness rate to achieve preferable result. Besides, the self-study faction in the detection and decision model based on the machine learning arithmetic is well actualized by building up the large digital mammary disease photo & text database. The system is capable to provide clinic assistant diagnosis for the mammary doctor as a second X-ray image reader in order to improve the diagnostic accuracy and convenience and supply reliable data source for research purpose.