Quantification of regional cardiac function is of central importance in cardiology, but has yet to be adopted into clinical practice due to limitations of the current techniques. Here we present a method requiring minimal human intervention for tracking “cine watermark” images, in which features have been encoded into the phase image of a cardiac cine series. The method employs nonlinear least squares optimization, which allows the sum of squared wrapped phase differences between patches in successive frames to be minimized globally across all frames, while regularizing over physically-motivated metrics. Preliminary results in healthy human volunteers show robust tracking.
This abstract and the presentation materials are available to members only; a login is required.