Motion-guided anisotropic filtering of ultrasound sequences

In spite of its many advantages, the use of ultrasound imaging remains limited by the quality of the images. Several filtering techniques have been proposed to overcome this limitation. In the case of 4D studies such as those available in echocardiography, filtering can be performed both in the spatial and in the temporal domain, making possible a further reduction of speckle. We present an algorithm that combines spatial and temporal filtering in a single framework. The algorithm is based on the coherence enhancing anisotropic filter, modified to make possible the introduction of motion information in echocardiographic sequences. Motion is estimated using a similarity measure specifically designed for the characteristics of ultrasound acquisition. Results show a clear improvement when compared to spatial-only filters, or to filters which do not differentiate between temporal and spatial dimensions