Related Experiment Videos
Band-target entropy minimization (BTEM) applied to hyperspectral Raman image data
Effendi Widjaja1, Nicole Crane, Tso-Ching Chen
1Department of Chemistry, University of Michigan, Ann Arbor, Michigan 48109, USA.
Applied Spectroscopy
|December 9, 2003
Summary
Band-target entropy minimization (BTEM) extracts more component spectra from hyperspectral Raman images. This method improves signal-to-noise ratio and recovers spectra typically missed by other techniques.
Area of Science:
- Spectroscopy
- Chemometrics
- Biomedical Imaging
Background:
- Hyperspectral Raman imaging generates complex data.
- Extracting pure component spectra is crucial for analysis.
- Existing methods may miss subtle spectral features.
Purpose of the Study:
- To introduce and validate Band-target entropy minimization (BTEM) for spectral extraction.
- To enhance the number of recoverable component spectra.
- To improve the signal-to-noise ratio of extracted spectra.
Main Methods:
- Singular value decomposition (SVD) to obtain eigenvectors.
- Information entropy minimization applied to targeted spectral bands.
- Inclusion of noise-dominated eigenvectors for enhanced recovery.
Main Results:
- BTEM successfully extracted component spectra from synthesized and biological data.
- More component spectra were obtained compared to conventional methods.
- Recovered spectra exhibited improved signal-to-noise ratios.
Conclusions:
- BTEM is an effective technique for hyperspectral Raman spectral extraction.
- The method leverages underutilized spectral information.
- BTEM offers advantages in spectral resolution and data analysis.