Using synthetic data generation to train a cardiac motion tag tracking neural network

• Published in: Medical image analysis Vol. 74; p. 102223
• Main Authors: Loecher, Michael, Perotti, Luigi E., Ennis, Daniel B.
• Format: Journal Article
• Language: English
• Published: Netherlands Elsevier B.V 01.12.2021; Elsevier BV
• Subjects: Artificial neural networks; Cardiac MRI; Computer Simulation; Convolutional neural network; Error analysis; Heart; Heart - diagnostic imaging; Humans; In vivo methods and tests; Machine learning; Magnetic Resonance Imaging; Motion; Neural networks; Neural Networks, Computer; Phantoms, Imaging; Synthetic data; Tag tracking; Tracking; Cardiac MRI; Tag tracking; Convolutional neural network; Machine learning; Synthetic data
• ISSN: 1361-8415; 1361-8423
• DOI: 10.1016/j.media.2021.102223

•A synthetic data generator is demonstrated that creates deforming MR images with known motion paths.•The known motion paths are used to train a tracking network for grid tagged cardiac MR images.•The tracking network can track in vivo images with high accuracy after being trained with only synthetic training. [Display omitted] A CNN based method for cardiac MRI tag tracking was developed and validated. A synthetic data simulator was created to generate large amounts of training data using natural images, a Bloch equation simulation, a broad range of tissue properties, and programmed ground-truth motion. The method was validated using both an analytical deforming cardiac phantom and in vivo data with manually tracked reference motion paths. In the analytical phantom, error was investigated relative to SNR, and accurate results were seen for SNR>10 (displacement error <0.3 mm). Excellent agreement was seen in vivo for tag locations (mean displacement difference = −0.02 pixels, 95% CI [−0.73, 0.69]) and calculated cardiac circumferential strain (mean difference = 0.006, 95% CI [−0.012, 0.024]). Automated tag tracking with a CNN trained on synthetic data is both accurate and precise.