For Full-Text PDF, please login, if you are a member of IEICE,|
or go to Pay Per View on menu list, if you are a nonmember of IEICE.
Qualitative Decomposition and Recognition of Infrared Spectra
Qi ZHAO Toyoaki NISHIDA
IEICE TRANSACTIONS on Information and Systems
Publication Date: 1996/06/25
Print ISSN: 0916-8532
Type of Manuscript: PAPER
Category: Artificial Intelligence and Cognitive Science
artificial intelligence, qualitative reasoning, inaccuracy handling, infrared spectrum recognition,
Full Text: PDF>>
The objective of this paper is to provide an effective approach to infrared spectrum recognition. Traditionally, recognizing infrared spectra is a quantitative analysis problem. However, only using quantitative analysis has met two difficulties in practice: (1) quantitative analysis generally very complex, and in some cases it may even become intractable; and (2) when spectral data are inaccurate, it is hard to give concrete solutions. Our approach performs qualitative reasoning before complex quantitative analysis starts so that the above difficulties can be efficiently overcome. We present a novel model for qualitatively decomposing and analyzing infrared spectra. A list of candidates can be obtained based on the solutions of the model, then quantitative analysis will only be applied to the limited candidates. We also present a novel model for handling inaccuracy of spectral data. The model can capture qualitative features of infrared spectra, and can consider qualitative correlations among spectral data as evidence when spectral data are inaccurate. We have tested the approach against about 300 real infrared spectra. This paper also introduces the implementation of the approach.