A Pipeline for the Generation of Synthetic Cardiac Color Doppler

Abstract

Color Doppler imaging is the modality of choice for simultaneous visualization of myocardium and intracavitary flow over a wide scan area. This visualization modality is subject to several sources of error, the main ones being aliasing and clutter. Mitigation of these artifacts is a major concern for better analysis of intracardiac flow. One option to address these issues is through simulations. In this paper, we present a numerical framework for generating clinical-like color Doppler imaging. Synthetic blood vector fields were obtained from a patientspecific computational fluid dynamics CFD model. Realistic texture and clutter artifacts were simulated from real clinical ultrasound cineloops. We simulated several scenarios highlighting the effects of i) flow acceleration, ii) wall clutter, and iii) transmit wavefronts, on Doppler velocities. As a comparison, an "ideal" color Doppler was also simulated, without these harmful effects. This synthetic dataset is made publicly available and can be used to evaluate the quality of Doppler estimation techniques. Besides, this approach can be seen as a first step towards the generation of comprehensive datasets for training neural networks to improve the quality of Doppler imaging.

Phase Unwrapping of Color Doppler Echocardiography Using Deep Learning

“…Although this remains to be verified, it is likely that a similar strategy could also work with disturbed flows, provided that we have access to Doppler data with their alias-free references. Such ground truths could be obtained by supervised correction, as in this study, and by simulations [25].…”

SIMUS: An open-source simulator for medical ultrasound imaging. Part II: Comparison with four simulators

“…In our team, SIMUS has also recently been applied in the generation of simulated moving ultrasound phantoms and left ventricles to train convolutional neural networks for motion estimation in echocardiography [ 28 , 29 ]. We also simulated clinical-type Doppler-echo cineloops with the future objective to train deep learning algorithms for intracardiac flow imaging [30] . In the same vein, Milecki et al .…”

Machine Learning in Cardiovascular Imaging: A Scoping Review of Published Literature

“…Another interesting area of research in medical imaging is image generation. Although current research efforts have utilized generative DL for objectives, such as removing artifacts from imaging data [ 32 ], raising the resolution of imaging [ 33 ], producing synthetic imaging datasets [ 34 , 35 ], and improving segmentation outcomes [ 36 ], and the capabilities of generative models are far broader. It has been demonstrated that generative models can convert two biplanar CXRs to a natural-looking chest CT scan and even incorporate synthetic tumoral lesions into normal imaging data [ 37 , 38 ].…”