Abstract #2723

Arterial Spin Labeling Denoising with Convolutional Neural Network and Convolutional Long-Short-Term-Memory (ConvLSTM)

Qinyang Shou¹, Chaitanya Gupte², Danny JJ Wang¹, and Hosung Kim²

¹Laboratory of Functional MRI Technology (LOFT), Stevens Neuroimaging and Informatics Institute, University of Southern California, Los Angeles, CA, United States, ²Neuroimaging with Deep Learning Lab (NIDLL), Stevens Neuroimaging and Informatics Institute, University of Southern California, Los Angeles, CA, United States

The purpose of this study was to develop a deep learning algorithm for 3D spatial-temporal denoising of ASL perfusion image series. 162 datasets from the Pediatric Template of Brain Perfusion database were used for model training and testing. The results showed that the proposed method can achieve higher Peak Signal-to-Noise Ratio (PSNR) and higher Structural Similarity Index (SSIM) than averaging of the time series and using traditional Principal Component Analysis (PCA) denoising. This result was robust when reducing the input measurements to one quarter of the total measurements, which shows the potential to reduce the scan time for ASL imaging.

This abstract and the presentation materials are available to members only; a login is required.

Join Here