Towards dynamic cardiac scenes interpretation based on spatial-temporal knowledge

Motion estimation of 3D coronary vessel skeletons from X-ray angiographic sequences

“…In order to quantitatively and objectively estimate the precision of the method, the result should be compared with the ground truth. However, since actual motion of coronary arteries during cardiac cycles cannot be quantitatively measured in vivo, as reported so far, a widely accepted golden standard has not been existed till now in research of tracking and estimating cardiac and coronary arterial dynamics from angiographic sequences including coronary angiography and left ventriculography [4,5,9,10,[12][13][14]20,[22][23][24][25][26][27]. Considering that our method includes two separate steps: sequential reconstruction of coronary vessel skeletons from CAG sequences and motion estimation of 3D vessel skeletons based on elastic registration, we evaluated the accuracy of the two steps respectively.…”

“…The major drawback of the conventional optical-flow-based methods [20,[22][23][24][25][26][27] mentioned in Section 1.2.2 is that errors existing in calibration parameters of the imaging system may be introduced into motion estimation results again during backprojection. Also, results of OF-based methods are sensitive to possible noises existing in the images because the OF field of the target is obtained according to partial derivatives of the intensity value of each pixel with respect to the pixel's position.…”

“…Since problems to be solved for both steps are to globally minimize a predefined function (i.e., energy function for sequential reconstruction and matching error function for motion estimation), both of the optimization problems can be solved with dynamic programming, an efficient optimization tool, to obtain global optimums with low computation cost and numerical stability. Compared with the method in [20,[22][23][24][25][26][27], not only possible errors introduced by calibration parameters of the imaging system are avoided, but also estimation procedures are simplified. And application of DP algorithm ensures low computation cost.…”

“…Finally, they reconstructed the 3D displacement vector of each skeleton point according to its two projections on both image planes. Puentes et al [22][23][24][25][26][27] used the OF-based approach presented in [19,20] to extract dynamic features of coronary vessel skeletons from biplane digital subtraction angiographic sequences. Furthermore, they developed a symbolic system to qualitatively interpret coronary arterial dynamic performances to assist clinical diagnosis of coronary artery diseases.…”

Directional analysis of cardiac left ventricular motion from PET images.

“…4D images are routinely acquired in the modern hospital using magnetic resonance imaging (MRI), X-ray computed tomography (CT), ultrasound, PET and SPECT. Manual analysis of these image sequences is a difficult task, which requires the skills of highly qualified specialists to interpret pathological motion caused by circulation problems in coronary arteries, anomalies in the cardiac muscle or disruption in electrical conduction to the ventricles (PUENTES et al, 2000;MURAMATSU;MATSUMOTO;NISHIMURA, 2005), and can lead to various problems, such as inter and intra-operator variability, high financial cost, slow speed of producing results. The proposed study therefore aims to quantify a 4D cardiac MVF from 4D cardiac images using natively 4D and automated techniques.…”