Adaptive Non-Local Means Filtering Based on Local Noise Level for CT Denoising

Technology #2011-178

Radiation dose from CT scanning is an increasing health concern, but lowering radiation dose alone generally produces a noisier image and may degrade diagnostic performance. Non-local means (NLM) denoising has become widely accepted as a very effective denoising tool in the image processing community, and has yielded very promising results in lung and abdominal CT. However, NLM requires an estimate of the noise level in the data, and in CT the noise level varies within and across slices. We have developed a technique for efficiently estimating the local noise level in CT, and have modified the NLM algorithm to adapt to local variations in noise levels. The resulting algorithm provides more effective denoising of CT data throughout a volume, and may allow significant lowering of radiation dose. The map of local noise level can also be used in many other filters to improve the noise reduction.