METHOD FOR OPTIMIZED BIAS AND SIGNAL INFERENCE IN DIFFUSION-WEIGHTED MAGNETIC RESONANCE IMAGE ANALYSIS

摘要: An approach to estimate noise, Rician signal bias and true in magnitude data obtained with magnetic resonance imaging is proposed. Rather than relying on repeat measurements for estimation of noise expected at given scan parameter settings, the method uses multiple different also referred as weightings, an iterative algorithm associated bias. The fact that behavior response these weightings individual image pixel locations can well be described analytic functions employed infer separate each weighting level measured. Measurements all measured levels contribute ultimate bias-free decay function. Therefore, so processed data, weighted signals computed arbitrary considerably better signal-to-noise ratio originally corresponding weightings. Bias-free desired levels, maps function fit parameters, or a combination such parameters used rapid highly sensitive tissue characterization.