Discriminating blue ballpoint pens inks in questioned documents by Raman imaging and mean-field approach independent component analysis (MF-ICA)
| Autor: | Carlos A. Teixeira, Ronei J. Poppi |
|---|---|
| Rok vydání: | 2019 |
| Předmět: |
Correlation coefficient
InformationSystems_INFORMATIONINTERFACESANDPRESENTATION(e.g. HCI) business.industry Computer science Sample (material) 010401 analytical chemistry ComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION Raman imaging Pattern recognition 02 engineering and technology Cashier's check 021001 nanoscience & nanotechnology 01 natural sciences Independent component analysis 0104 chemical sciences Analytical Chemistry Chemometrics ComputingMethodologies_PATTERNRECOGNITION Artificial intelligence 0210 nano-technology business Spectroscopy Prior information |
| Zdroj: | Microchemical Journal. 144:411-418 |
| ISSN: | 0026-265X |
| DOI: | 10.1016/j.microc.2018.10.002 |
| Popis: | In the present work, Raman spectroscopy and chemometrics tools were explored as an analytical method to discriminate blue pen strokes on common office paper and bank check paper. The sample set was constituted of blue ballpoint pen inks available on local market and promotional pens as well. The samples were prepared to simulate an adulterated document using different pens. The proposal methodology was based on Raman imaging spectroscopy and Independent Components Analysis algorithm, which were able to recover the original spectrum of each pen, as well as images containing special information about each of the pens used. All recovered spectra could be match with references spectra at least with 0.8572 of correlation coefficient, while recovered images left no doubt about whether a forgery occurred. At the end of analysis, all the examined cases could be solved based on the discrimination between the pens inks, without any prior information. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Full Text from ScienceDirect