Abstract
This paper presents a computer-aided diagnosis system for breast tumor by using three-dimensional ultrasonic images. The system consists of three parts. The first part is to reconstruct a volumetric ultrasonic image data using both the echographic image slices, obtained by commercial ultrasonic echography device, and the probe's position and orientation data tracked by a magnetic position sensor. In the second part, breast tumor regions are automatically extracted by using fuzzy reasoning techniques. Finally, breast tumors are quantitatively evaluated based on some parameters by using fuzzy reasoning in order to differentiate malignant tumors from benign tumors. So far, the system has been applied to over 100 cases of clinical breast tumor data, and the results show excellent agreement with those of fine needle aspiration cytology. It is concluded that this system has the potential for greatly increasing the accuracy and efficiency of breast cancer examination.
