Automated separation of low quality and artifact spectra by pattern recognition in the processing of MR spectral images
Menze B, Hamprecht F, Kelm B
University of Heidelberg
We propose a methodology, which automates the decision of a human operator about the quality and interpretability of an MR spectrum. It separates both spectra of low signal-to-noise ratio and spectra corrupted by artifacts. The method is applicable in conjunction with a following quantification or a pattern recognition. The primary application will be in the automated processing of high resolution MR spectral images.