Related Experiment Videos
"Derived spectra" software tools for detecting spatial and spectral features in spectrum images
Dale E Newbury1, David S Bright
1National Institute of Standards and Technology, Gaithersburg, Maryland 20899-8370, USA. dale.newbury@nist.gov
Scanning
|February 17, 2005
Summary
New software tools called "derived spectra" help analyze complex spectrum images. These tools efficiently identify common and rare features within large datasets, aiding scientific discovery.
Area of Science:
- Spectroscopy
- Data Analysis
- Materials Science
Background:
- Spectrum imaging generates large, complex datasets ('datacubes') that are challenging to analyze.
- Efficiently extracting information from these datacubes is crucial for scientific discovery, especially when examining unknown samples.
Purpose of the Study:
- To introduce a class of software tools called "derived spectra" for analyzing spectrum imaging datacubes.
- To demonstrate how these tools can aid analysts in recognizing both common and rare features within the data.
Main Methods:
- Developed "derived spectra" tools that create spectrum-like displays from datacube pixel intensities.
- Evaluated specific derived spectra tools: SUM, MAXIMUM PIXEL, RUNNING SUM, and RUNNING MAXIMUM.
Main Results:
- The SUM-derived spectrum effectively identifies common features in the datacube.
- MAXIMUM PIXEL and RUNNING MAXIMUM-derived spectra excel at locating rare or unanticipated features, even those present at a single pixel.
Conclusions:
- Derived spectra tools offer an efficient method for analyzing complex spectrum imaging data.
- These tools enhance the ability to recognize diverse features, facilitating the study of unknowns.