A Noise Reduction Method Based on a Statistical Test of High Dimensional Pixel Vectors for Dynamic and Volumetric Images

A noise reduction method applicable for dynamic and three-dimensional imaging modalities was evaluated. We applied our method to images obtained by CT, MRI, SPECT, angiography, and numerical phantom data, and measured the reduction of image noise. The standard deviation of the noise was reduced to 12–13% in a numerical dynamic phantom, 35–46% in dynamic CT, 19%-23% in a dynamic scintigram, 6–8% in dynamic SPECT, 13–15% in volumetric MRI of PVA resolution phantom, 12–17% in volumetric MRI of a PVA head phantom, and 50–53% in angiography of a moving phantom. The method reduced noise effectively, while spatial and temporal resolution remained satisfactory.