Image Reconstruction in the Grappa Algorithm Formalism

Yufang Bao, Andrew Maudsley

Abstract: In this paper is presented a novel parallel image reconstruction algorithm that follows the same formalism as the existing GRAPPA reconstruction, and is termed as the extended GRAPPA algorithm, which is equivalent to the generalized SENSE (GenSENSE) algorithm in reconstructing a high-resolution sensitivity-invariant MR image. Based on this understanding, a concept of weight map is introduced for reconstruction, and the existing GRAPPA algorithm is further interpreted as an approximate of the extended GRAPPA algorithm by selectively choosing weights of significant value in the map for reconstruction