Four-Dimensional Angiography Data Processing for Phase Separation

The patent applies segmentation by dividing the overlapping contrast agent concentration time profile into separate arterial and venous phases using a maximum likelihood estimation algorithm. This computational segmentation separates the mixed signal into distinct phases, allowing individual visualization of arterial and venous structures even when they overlap in time, thus resolving the contradiction between maintaining vessel filling quality and preserving phase separation information.

The patent applies preliminary action by first capturing a complete 3D vessel data set with sufficient contrast agent filling (using a longer injection time), then subsequently applying computational phase separation algorithms to the already-acquired data. This allows the vessel data set to be created with optimal quality first, and the phase separation to be achieved through post-processing, thus avoiding the need to compromise injection time while still achieving phase separation.

The patent applies preliminary action by first acquiring a complete angular range of projection images with sufficient contrast agent filling, then using computational phase separation on the reconstructed 4D data set. This allows the full angular range to be captured for optimal image quality without extending the contrast agent administration beyond what is needed for complete vessel filling, as the phase separation is achieved computationally rather than temporally.

The patent applies segmentation by implementing a maximum likelihood estimation algorithm that segments the overlapping contrast agent concentration curves into distinct arterial and venous phases. This computational segmentation is applied to the time-resolved 3D images generated by multiplicative back-projection, separating the mixed arterial and venous signal into individual phase-specific images, thus preserving both temporal resolution and phase separation.