Temporal methods for ultrasound speckle reduction

With the exception of direct averaging, temporal filtering has not found application in medical ultrasonic imaging. Indeed while spatial and frequency compounding have been the subject of research effort, time compounding has largely been ignored despite its intuitive advantages. Schistad and Taxt (see Nuclear Science Symposium and Medical Imaging Conf., p.2210, 1991) have used an adaptive linear filter, median filter and quantum mechanical field model for one-dimensional temporal filtering. This paper describes a new technique for temporal image filtering which uses image statistics to control the filtering algorithm. The technique is based on extending an established TDLMS filtering technique. Results are presented in comparison with direct averaging to illustrate the advantages that this new technique can accrue