Spectral peak verification and recognition using a multi-layered neural network.
Wythoff-B; Levine-SP; Tomellini-SA
School of Public Health, University of Michigan, Ann Arbor, Michigan :1-36
The feasibility of exploiting neural network technology to recognize peak shaped signals in analytical data was evaluated. A peak detection system based on a class of neural networks known as multilayered perceptrons was created. While the system described was developed to interpret infrared spectral data, peak detection has implications in every chemical application where the recognition of peak shaped signals in analytical data is important. Chemical applications could include virtually all spectroscopic and chromatographic methods, as well as flow injection analysis and the scanning electrochemical methods. The idea that the trained human and artificial neural network can adequately perform the signal detection task under similar conditions was examined. The incorporation of a noise reference was found to aid both the human and the network signal detection process.
NIOSH-Grant; Grants-other; Analytical-methods; Chemical-analysis; Analytical-chemistry
Environmental & Indust Health School of Public Health II 1420 Washington Heights Ann Arbor, MI 48109-2029
Final Grant Report
NTIS Accession No.
Other Occupational Concerns; Grants-other
School of Public Health, University of Michigan, Ann Arbor, Michigan
University of Michigan at Ann Arbor, Ann Arbor, Michigan