Extracting acoustic shadowing from ultrasound image using local difference

Abstract

We propose a method for extracting acoustic shadowing from ultrasound image. If we locate the acoustic shadowing exactly then we can also extract the hyperechoic substances with a high probability. The proposed method reconstitutes an original ultrasound image to simplify calculations and uses the locally horizontal difference to extract shadow candidates. The shadow candidates are classified into start and finish point, and shadow regions are extracted by using them. The experiment results show the proposed method extracts 23 of 27 acoustic shadows from ultrasonogram efficiently and it can be used in wide applications.