Automated Ultrasound Doppler Angle Estimation Using the Probabilistic Hough Transform

Abstract

Modern Doppler ultrasound can estimate the velocity of blood flow in vessels quantitatively and noninvasively. Once the Doppler angle, i.e. the angle between the insonating beam and the direction of blood flow, is given, the velocity of flow can be estimated using the standard Doppler equation. However, the operator is required to manually tune the vessel slope cursor based on the direction of the vessel wall in the B-mode image to specify the Doppler angle. This manual tuning process is time consuming and error prone. In this paper, a novel automatic way of specifying the Doppler angle is proposed to improve the efficiency and accuracy of the blood flow velocity measurement. Our approach starts with the image preprocessing and segmentation of blood vessel in ultrasound common carotid artery image. The segmentation step is followed by an estimation of Doppler angle based on the probabilistic Hough Transform technique. On 15 ultrasound video sequences of the common carotid artery taken from 4 healthy subjects, the results generated by the proposed method are compared to the angles specified by an expert operator. The deviation of the angles estimated by the proposed method from the angles specified by the operator is 4.15 degrees ± 2.92 (mean ± standard deviation). The proposed method eliminates the need of manually tuning of the vessel slope cursor and the estimated angles are in highly agreement with the ones specified by the operator. It has the potential to enhance the work flow and improve the speed of the ultrasound spectral Doppler examination.