Source:http://linkedlifedata.com/resource/pubmed/id/15338736
Switch to
| Predicate | Object |
|---|---|
| rdf:type | |
| lifeskim:mentions | |
| pubmed:issue |
8
|
| pubmed:dateCreated |
2004-9-1
|
| pubmed:abstractText |
A technique has been developed for combining a series of low signal-to-noise ratio (SNR) real-time magnetic resonance (MR) images to produce composite images with high SNR and minimal artifact in the presence of motion. The main challenge is identifying a set of real-time images with sufficiently small systematic differences to avoid introducing significant artifact into the composite image. To accomplish this task, one must: 1) identify images identical within the limits of noise; 2) detect systematic errors within such images with sufficient sensitivity. These steps are achieved by evaluating the correlation coefficient (CC) between regions in prospective images and a template containing the anatomy of interest. Images identical within noise are selected by comparing the measured CC values to the theoretical distribution expected due to noise. Sensitivity for systematic error depends on the SNR of the CC (=SNR(CCmax)), which in turn depends on the noise, and the template size and structure. By varying the template size, SNR(CCmax) may be altered. Experiments on phantoms and coronary artery images demonstrate that the SNR(CCmax) necessary to avoid introducing significant artifact varies with the target composite SNR. The future potential of this technique is demonstrated on high-resolution (approximately 0.9 mm), reduced field-of-view real-time coronary images.
|
| pubmed:language |
eng
|
| pubmed:journal | |
| pubmed:citationSubset |
IM
|
| pubmed:status |
MEDLINE
|
| pubmed:month |
Aug
|
| pubmed:issn |
0278-0062
|
| pubmed:author | |
| pubmed:issnType |
Print
|
| pubmed:volume |
23
|
| pubmed:owner |
NLM
|
| pubmed:authorsComplete |
Y
|
| pubmed:pagination |
1034-45
|
| pubmed:dateRevised |
2006-11-15
|
| pubmed:meshHeading |
pubmed-meshheading:15338736-Algorithms,
pubmed-meshheading:15338736-Arteries,
pubmed-meshheading:15338736-Artifacts,
pubmed-meshheading:15338736-Coronary Vessels,
pubmed-meshheading:15338736-Feedback,
pubmed-meshheading:15338736-Humans,
pubmed-meshheading:15338736-Image Enhancement,
pubmed-meshheading:15338736-Image Interpretation, Computer-Assisted,
pubmed-meshheading:15338736-Information Storage and Retrieval,
pubmed-meshheading:15338736-Magnetic Resonance Imaging,
pubmed-meshheading:15338736-Motion,
pubmed-meshheading:15338736-Numerical Analysis, Computer-Assisted,
pubmed-meshheading:15338736-Online Systems,
pubmed-meshheading:15338736-Pattern Recognition, Automated,
pubmed-meshheading:15338736-Phantoms, Imaging,
pubmed-meshheading:15338736-Reproducibility of Results,
pubmed-meshheading:15338736-Sensitivity and Specificity,
pubmed-meshheading:15338736-Signal Processing, Computer-Assisted,
pubmed-meshheading:15338736-Stochastic Processes,
pubmed-meshheading:15338736-Subtraction Technique
|
| pubmed:year |
2004
|
| pubmed:articleTitle |
Adaptive averaging for improved SNR in real-time coronary artery MRI.
|
| pubmed:affiliation |
Department of Medical Biophysics, Sunnybrook Health Science Centre, University of Toronto, Toronto, ON M5G 2N2, Canada. marshall.sussman@utoronto.ca
|
| pubmed:publicationType |
Journal Article,
Clinical Trial,
Comparative Study,
Research Support, Non-U.S. Gov't,
Validation Studies
|