Decision support system for breast cancer screening

Abstract

Currently, decision support systems in radiation mammology focus on the detection and classification of neoplasms, despite the fact that the real work of a radiologist does not imply a diagnosing. Computer vision systems use a black box model and do not explain the results of work, which is unacceptable in medicine. The proposed hybrid approach, combining the work of computer vision subsystems and a problem solver based on graphodynamic sc-memory, can make a structured research report based on the detected anomalies and production output. This approach will reduce the time of the doctor-radiologist. Also, for the first time, the decision support system of an X-ray technician is presented to improve the quality of breast styling. The system will improve the quality of the X-ray image and, accordingly, improve the evaluation of the study.