Abstract
Automatically detecting tumors and extracting lesion boundaries in ultrasound images is difficult due to the variance in shape and the interference from speckle noise. In this study, we present a novel algorithm to automatically find lesion contours in ultrasonic breast images. With inherent property of ultrasound image embedding speckle noise, the image can be regarded as degraded paint character (DPC) image containing closure noise, which is well known in perceptual organization of psychology. An effective scheme of removing closure noise using iterative disk expansion method has been addressed in our previous works and it has been successfully applied to many applications. The contour extraction of ultrasonic breast lesions can be equivalent to the removal of speckle noise. In this study, simulations and clinical images were performed to evaluate the performance of the proposed algorithm. The results show that the mean normalized true positive area overlap between simulated contour and contour obtained by the proposed algorithm is 82%. There is a strong correlation between experienced physicians manual and automated contour extraction in clinical breast images. Furthermore, due to running time for contour extraction in a single breast lesion image was 0.06 second, the algorithm is feasible to be implemented in a real-time ultrasound imaging system. © 2005 IEEE.