A novel algorithm for modifying echocardiographic particle image velocimetry flow field

Echocardiographic particle image velocimetry (E-PIV) technology is commonly used to trace the contrast bubbles intravenous injected in order to describe the blood flow field in heart chambers. However, affected by speckle noise, the E-PIV output vector fields in echocardiographic images contain a large number of wrong velocity vectors. To correct the wrong vectors, we proposed a new iterative algorithm that smooth and modify the E-PIV output vector fields. The method is based on E-PIV, hydrodynamics partial differential equation and anisotropic diffusion. It detects local blood flow trendency, smooths the flow fields and corrects the wrong vectors without blurring important characteristic of the blood flow. Experiments on artificially generated images and vivo echocardiographic images show that the proposed method can noticeable improve the accuracy of the flow fields.