Due to the rapid increase in use of CT imaging and the recently-heightened awareness of radiation-induced cancer, improving the diagnostic quality of low dose CT has become increasingly important. The signal-to-noise level of low dose CT scans is significantly lower than for normal dose scans, and thus may cause a decrease in diagnostic accuracy. Mayo Clinic researchers have developed an improved non-local means filtering method for improving quality of low dose CT images. Existing image sets and neighboring slices are utilized to inform the algorithm and reduce noise. Additionally, the algorithm utilizes a smaller search window size which can increase filter performance while at the same time allowing faster computation due to fewer calculations.
Filtering method to improve the quality of low-dose CT images, as well as MR, ultrasound, and magnetic resonance elastography images.
Stage of Development
This software system has been implemented in the Department of Radiology at Mayo Clinic.
Kelm Z, Blezek DJ, Bartholmai BJ, Erickson BJ. “Optimizing non-local means for denoising low dose CT” IEEE ISBI: From Nano to Macro. 2009. pp. 662–665.