Analysis Of Nuclear Magnetic Resonance Spectroscopy Based On Wavelet Transform And Pattern Recognition

MRS and MRI are the main techniques of application of nuclear magnetic resonance technology to clinical diagnosis of cancer. Because of better visual effects, MRI is used widely in clinical application, but MRS is still on the stage of experiment research. According to the characteristics of FID signals and MRS, the paper presented the new scheme of analysis of FID and MRS based on wavelet transform and pattern recognition technologies. Wavelet transform is used to decompose FID and MRS, and the feature of cancer is recognized and extracted from the MRS based on pattern recognition. The methods of decomposing, de-nosing and feature extraction based on wavelet transform and pattern recognition were experimented and the key techniques of selection of mother wavelet, decomposing level, and feature classification etc. were investigated.