Generalized series dynamic imaging with deformable references

Many imaging applications require the acquisition of a time series of images. Conventional full-scan methods acquire each of these images independently, resulting in a tradeoff between spatial resolution and temporal resolution. To address this problem, a generalized series-based reduced-encoding method has been proposed, which collects one or a few high-resolution references and a sequence of reduced data sets. This paper extends this imaging scheme to allow deformable references to be used to compensate for possible geometric variations between the references and the reduced data sets. The deformable references are constructed using a thin-plate spline model and a mutual information-based registration algorithm. Simulation results demonstrate that the proposed method can handle large geometric changes such as those encountered in cardiac imaging.