PDF version of this entire document
The proposed approach needs to be explored in the context of recent papers including those whose authors choose to work with non-rigid registration (see Isola et al. [1] from 2010) and/or simple tagging. Isoda et al. [2] have only just published an article which looks at three-dimensional phase-contrast MRI and results include an experiment that deals with the cardiac phase.
Last year's review of shape models for 3-D image segmentation [3] may be worthwhile for extracting the signal from larger images, but given the distinctly dark background captured when isolation is applied at the scanner, there ought to be simpler ways to achieve the same effect and obtain similar segmentation. A paper from Zhuang et al. [4] covers cardiac MRI for segmentation of the heart without human intervention. It does not deal with the problem of motion and real-time (or `off-line') tracking of parts of the heart (4-D), though.
Motion caused by respiration is studied in [5]. Many papers on tracking exist and half a year ago there was a speckle-tracking echocardiographic study into myocardial deformation at the University of Hong Kong [6].
Other authors who had studied myocardial strains over a decade earlier [7] more recently looked at what they described as a ``non-invasive method for assessing regional myocardial'' [8] (paper from 2003).
Cardiac tagging was done in Johns Hopkins University last year. It was tissue tagging [9] with cardiovascular magnetic resonance (CMR). More experimental work on CMR tagging (from 2008) can be found in [10] where mice hearts are studied. A paper from Wei et al. also considered CMR tagging by experimenting on mice [11]. Methods involved ``[i]n vivo myocardial function was evaluated by 3D CMR tagging in mdx mice.'' More material on CMR tagging [12] ``introduce[s] a standardized method for calculation of left ventricular torsion by CMR tagging and [determining] the accuracy of torsion analysis in regions using an analytical model.'' Delling et al. [13] looked at CMR and explained: ``We sought to assess the correlation between mitral valve characteristics and severity of mitral regurgitation (MR) in subjects with mitral valve prolapse (MVP) undergoing cardiac magnetic resonance (CMR) imaging.'' Some less relevant papers on the subject may include [14,15] on CMR tagging.
The problem of tracking parts of the heart is a complicated one and a particular group has proposed ``a novel dynamic model, based on the equation of dynamics for elastic materials and on Fourier filtering.'' [16]
To some groups, robotic ``intervention on the beating heart'' is said to be a worthwhile route of exploration as well [17]. This one study on pig hearts [18] -- like [6] -- considers speckle-tracking. There is also lot about HARP, including a paper from last year [19]. Lastly, Cardiac Imaging has this broad new survey from 2010 [20].
