Deep motion tracking from multiview angiographic image sequences for synchronization of cardiac phases

Abstract

In the diagnosis and interventional treatment of coronary artery disease, the 3D time reconstruction of the coronary artery on the basis of X-ray angiographic image sequences can provide dynamic structural information. The synchronization of cardiac phases in the sequences 15 is essential for minimizing the influence of cardiorespiratory motion and realizing precise 3D time reconstruction. Key points are initially extracted from the first image of a sequence. Matching grid points between consecutive images in the sequence are extracted by a multi-layer matching strategy. Then deep motion tracking of key points is achieved by local deformation based on the neighboring grid points of key points. The local deformation is optimized by the Random 20Sample Consensus algorithm. Then, a simple harmonic motion model is utilized to distinguish cardiac motion from other motion sources (e.g. respiratory, patient movement, etc.). Next, the signal which is composed of cardiac motions is filtered by a band-pass filter to reconstruct the cardiac phases. Finally, the synchronization of cardiac phases from different imaging angles is realized by piece-wise linear transformation. The proposed method was evaluated using clinical X-ray angio-25 graphic image sequences from 13 patients. matching points can be accurately computed by the deep motion tracking method. The mean peak temporal distance between the reconstructed cardiac phases and the electrocardiograph signal is s. The correlation between the cardiac phases of the same patient is over . Compared with three other state-of-the-art methods, the proposed method accurately reconstructs and synchronizes the cardiac phases from different sequences of the 30 same patient. The proposed deep motion tracking method is robust and highly effective in synchronizing cardiac phases of angiographic image sequences captured from different imaging angles.

Motion-flow-guided recurrent network for respiratory signal estimation of x-ray angiographic image sequences

“…However, the widely used single-plane and rotational angiography imaging systems can only collect one angiogram at a specified angle at a time. Thus, before the reconstruction process, the cardiac and respiratory states of vasculatures on multi-angle angiograms must be analyzed for motion synchronization and compensation (Yang et al 2014, Cong et al 2015, Klugmann et al 2018, Song et al 2019a. In vascular interventional therapy, target vasculatures cannot be seen all the time during surgery because of the limited use of contrast agents.…”

End-to-End Deep Learning Model for Cardiac Cycle Synchronization from Multi-View Angiographic Sequences

“…The authors of [4] proposed a statistical method for cardiorespiratory gating on fluoroscopic images, which was also later used in [5], but that method fails when the contrast agent is injected unevenly because of image intensity fluctuations. The work in [6] made use of multi-layer matching for motion tracking to determine the cardiac phase by measuring the cardiac motion velocity of local patches of coronary arteries. While this method was shown to work relatively well to synchronize angiographic sequences, the method completely relies on the accurate tracking of the vessel segments, which is not always possible.…”