A new approach to the smoothing of dynamic nuclear medicine data: concise communication

A weighted three-dimensional digital filter that smooths data in both space and time has been developed for use with dynamic nuclear medicine studies. This smoothing algorithm allows a large improvement in signal-to-noise ratio without unacceptable degradation of spatial and temporal resolution. The initial results of using this smoothing algorithm suggests that it is superior to a standard nine-point smoothing function used on dynamic data. This is particularly encouraging since the parameters of the digital filter have not been optimized. The quality of the processed digital images is at least equivalent to that of analog images, and the digital images may be of more diagnostic value. The new algorithm also appears useful in preparing dynamic dynamic data for other manipulations such as the creation of parametric images or the extraction of quantitative measurements.