Evaluation of simple acceleration strategy for advanced neural diffusion models based on half q-space under-sampling
Min-xiong Zhou1, Huiting Zhang2, Yang Song3, Guang Yang3, and Xu Yan2
1Shanghai University of Medicine & Health Sciences, Shanghai, China, 2MR Scientific Marketing, Siemens Healthcare, Shanghai, China, 3Shanghai Key Laboratory of Magnetic Resonance, East China Normal Univeristy, Shanghai, China
The
half q-space under-sampling scheme did not significant sacrifice the accuracy
of quantitative maps of advanced diffusion model but could accelerate the
acquisition by 2 fold. In addition, a symmetrically data copy step is needed to
improve the estimation accuracy.
Figure 1. ODI and ICVF maps (NODDI model) of
three different schemes (Full, Half and Half-to-Full) and their corresponding
percent of square coefficient of variation (CV) for a patient.
Figure 2. RTAP and NG maps (MAP-MRI model)
of three different schemes (Full, Half and Half-to-Full) and their
corresponding percent of square coefficient of variation (CV) for a patient.