ONSTRUCTION OF PRECISE HUMAN BODY MODEL AND DATABASES USING PET?CT IMAGES FOLLOWING AUTOMATIC DIAGONOSIS

点击翻译按钮获取中文摘要

This research is a celebrative research project between Graduate School of Medicine of Yokohama City University and Graduate School of Environment and Information Sciences of Yokohama National University. It aims to construct a CAD (Computer Automated Diagnosis) system for the 'automatic cancer detecting' using the whole-body PET (Positron Emission Tomography) and CT (Computer Tomography) Images. To achieve this, we made many interviews with medical doctors (Radiologists) and finally established the formal algorithm to detect abnormal 3D-area in the human bodies. Then we developed the CAD Engine which interprets the algorithm (written by the special language MDPL), make diagnosis and produce the formatted data including abnormal 3D?area and the annotations. This system enables us to store the CAD results and to retrieve them by using Database Management System. Within the project, we asked several medical organizations (hospitals and clinics) to evaluate the reliability and efficiency of the CAD system. Total evaluation and Improvement of the algorithm to cover all the variations of human bodies are left for future.