A Bayesian approach for edge detection in medical ultrasound images

Successful applications of digital image processing techniques to medical ultrasound images have been limited in part because of the lack of an useful imaging model for clinical ultrasound B-scans. In this work, the authors derive a discrete linear imaging model appropriate for clinical ultrasound B-scans. Based on the newly derived model, the authors developed a Bayesian restoration approach that is currently designed for the generation of correct edges of medical ultrasound images. Their results demonstrate that successful edge detection can indeed be achieved by the proposed method. The proposed Bayesian approach is very flexible and has the potential of being extended to perform speckle reduction, edge detection, and region segmentation at the same time.