MRI has a broad range of applications due to its flexible acquisition capabilities. This demands profound knowledge and careful parameter adjustment to identify stable sets which guarantee a high image quality for various and, especially in examinations of patients, unpredictable conditions. This complex nature and long examination times make it susceptible to artifacts which can markedly reduce the diagnostic image quality. An early detection and possible correction of these artifacts is desired. In this work we propose a convolutional neural network to automatically detect, localize and quantify motion artifacts. Initial results in the head and abdomen demonstrate the method’s potential.
This abstract and the presentation materials are available to members only; a login is required.