26 February 2001 Pattern recognition and chemometrics for spectral classification
Author Affiliations +
Proceedings Volume 4205, Advanced Environmental and Chemical Sensing Technology; (2001) https://doi.org/10.1117/12.417466
Event: Environmental and Industrial Sensing, 2000, Boston, MA, United States
The broader definition of chemometrics includes methods such as pattern recognition (PR) and signal/image processing for noninvasive analysis and interpretation of data. These methods are among the most powerful tools currently available for noninvasively examining spectroscopic and other chemical data. Using spectral data, these systems have found a variety of applications employing analytical techniques for gas chromatography, fluorescence IR or NMR spectroscopy, etc. An advantage of PR approaches is that they make no a priori assuniption regarding the stmcture of the spectra. However, a majority of these systems rely on hunianjudgment for parameter selection and classification of spectra. Generally a spectral pattern recognition (SPR) problem is considered as a group of several subproblems. We considered a SPR problem as a group of five subproblems: spectra acquisition, feature extraction, feature selection, spectra organization, and spectra classification. One of the basic issues in PR approaches is to determine and measure the discriminatory features useful for successful classification. A spectral pattern classification system, combining spectral feature extraction and selection, and decision-theoretic approaches, is developed. It is shown how such a system can be used for analysis of large data analysis, warehousing, and interpretation.
© (2001) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Khalid J. Siddiqui and DeLyle Eastwood "Pattern recognition and chemometrics for spectral classification", Proc. SPIE 4205, Advanced Environmental and Chemical Sensing Technology, (26 February 2001); doi: 10.1117/12.417466; https://doi.org/10.1117/12.417466

Back to Top