Pattern Recognition Algorithms For Tissue Diagnosis By Near-infrared FT-Raman Spectroscopy

The newly developed technique of near-infrared laser excited Fourier transform (-)-Raman spectroscopy is capable of accurately detecting diseased and healthy tissues, because of its fingerprinting and in-vivo capabilities. This opens up the possibility that this near-IR FT-Raman technique may be developed into a powerful "smart" fiberoptic sensor for biomedical applications. The development of such an automated FT-Raman diagnostic system requires the obtained spectral data to be processed. By focusing on the FT-Raman data obtained for urinary calculi and normal kidney tissue, we demonstrate in this study the utility of pattern recognition algorithms in Raman diagnosis and classification of diseased tissues.