Misregistration Artifacts Ueda D and Katayama Y et al Published Online March 30 2021 httpsdoiorg101148radiol2021203692 In a retrospective study of 17 934 image pairs collected from 40 patients a deep learning DL model generated cerebral angiograms with fewer misregis ID: 914418
Download Presentation The PPT/PDF document "Deep Learning–based Angiogram Generati..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
Deep Learning–based Angiogram Generation Model for Cerebral Angiography without
Misregistration
Artifacts
Ueda D and Katayama Y et al. Published Online:
March 30, 2021https://doi.org/10.1148/radiol.2021203692
In a retrospective study of
17 934 image pairs collected from 40 patients, a deep learning (DL) model generated cerebral angiograms with fewer misregistration artifacts than conventional digital subtraction angiography (DSA). The DL-generated angiograms showed high similarity to DSA images with mean peak SNR of 40.2 dB and mean structural similarity index measure of 0.97. On visual evaluation, all specialists assessed the DL-generated angiograms as more clinically useful than the misregistrated DSA images.
In a 67-year-old woman, a cerebral angiogram shows an ICA aneurysm. The misregistration artifacts of the skull (arrows) in the conventional DSA images are almost invisible in the DL-generated angiograms.