MRI Parallel Imaging Hybrid Sampling SNR

The patent segments the K-space sampling into two distinct parts: low-frequency fully-sampled data and high-frequency uniformly under-sampled data. This segmentation allows the critical low-frequency information (which determines image contrast and SNR) to be fully captured while accepting under-sampling in the high-frequency regions (which contribute less to overall image quality). The segmentation resolves the contradiction by strategically allocating sampling resources.

The patent applies different sampling strategies to different regions of the K-space based on their local importance. The low-frequency central region receives full sampling to preserve image quality and SNR, while the high-frequency peripheral regions accept under-sampling to increase imaging speed. This local differentiation of sampling quality resolves the contradiction between speed and image quality.

By segmenting the sampling strategy, the patent ensures that the low-frequency data (which contains reliable structural information) is fully sampled, maintaining image reliability. The high-frequency data is under-sampled but reconstructed using the segmented approach, minimizing artifact introduction while maintaining speed.

The patent converts the potential harm of under-sampling artifacts into a benefit by strategically applying under-sampling only where it matters less (high-frequency regions) while maintaining full sampling where it matters most (low-frequency regions). The hybrid approach transforms what would be a harmful uniform under-sampling into a beneficial selective sampling strategy.