Motion blur parameter estimation based on autocorrelation for liver ultrasound image

Liver ultrasound images are an important tool of the bile duct cancer surveillance. But it has difficult to analyze because the liver ultrasound images are motion blurred. The motion blur of the liver ultrasound image is the one factor to decrease the liver ultrasound image quality. The restoration method for the motion blur image is one choice to enhance the liver ultrasound image. However, the liver ultrasound images are differential of the blur level value. Thus, the blur level value estimation for the liver ultrasound image is the way to improve the restoration performance. This paper proposed the new method of the motion blur parameter estimation for the liver ultrasound images (ELUS). The main idea of the ELUS method is to estimate the blur level value of the liver ultrasound image. The proposed methods consist of three processes, which are the autocorrelation method, the reference number method, and the motion blur estimation method. To evaluate and compare the performance of the proposed method, two measurements are applied, the accuracy estimation of the blur level value and the scan line pixel. The experimental results showed that the proposed method is able to give the higher success rate of the blur level estimation for liver ultrasound image than the other method.