Abstract #2700

Evaluation of Image Quality Improvement using Wavelet Denoising Based on Stein's Unbiased Risk Estimate (SURE)

Tao Zhang¹, Peng Lai², Shreyas Vasanawala³, Robert Herfkens³, Kedar Khare⁴, Luca Marinelli⁴, Kevin F. King⁵, Anja Brau²

¹Electrical Engineering, Stanford University, Stanford, CA, United States; ²Applied Science Laboratory, GE Healthcare, Menlo Park, CA, United States; ³Radiology, Stanford University, Stanford, CA, United States; ⁴GE Global Research Center, Niskayuna, NY, United States; ⁵GE Healthcare, Waukesha, WI, United States

Low SNR in MRI images can usually be found in cases with parallel imaging and acquisitions with high spatial resolution. Image denoising methods can increase SNR. But some of them will introduce artifacts such as image blurring. In this work, wavelet denosing based on Steins Unbiased Risk Estimate (SURE) was evaluated on various MRI applications. Comparisons on overall image quality and image sharpness were carried out on these applications. Based on the evaluation results from radiologists, wavelet denoising based on SURE can effectively increase SNR without introducing image blurring.