Abstract
Acquisition of large data sets from human tissues by infrared (IR) microscopy is now routine. However, processing such large data sets, which may contain more than 10 000 spectra, provides an enormous challenge. Overcoming this challenge and developing nonsubjective methods for the analysis of IR microscopic results remain the major hurdle to developing clinically useful applications. A three-step pattern recognition strategy based upon linear discriminant analysis has been developed for use as a search engine for tissue characterization. The three-step strategy includes a genetic algorithm-guided data reduction step, a classification step based upon linear discriminant analysis, and a final step in which the discriminant coefficients are converted into a visually appealing, nonsubjective representation of the distribution of each class throughout the tissue section. The application of this search engine in the characterization of tumor-bearing skin is demonstrated.
PDF Article
More Like This
Disease pattern recognition in infrared spectra of human sera with diabetes
mellitus as an example
Wolfgang Petrich, Brion Dolenko, Johanna Früh, Manfred Ganz, Helmut Greger, Stephan Jacob, Franz Keller, Alexander E. Nikulin, Matthias Otto, Ortrud Quarder, Ray L. Somorjai, Arnulf Staib, Gerhard Werner, and Hans Wielinger
Appl. Opt. 39(19) 3372-3379 (2000)
Automated multimodal spectral histopathology for quantitative diagnosis of residual tumour during basal cell carcinoma surgery
Radu Boitor, Kenny Kong, Dustin Shipp, Sandeep Varma, Alexey Koloydenko, Kusum Kulkarni, Somaia Elsheikh, Tom Bakker Schut, Peter Caspers, Gerwin Puppels, Martin van der Wolf, Elena Sokolova, T. E. C. Nijsten, Brogan Salence, Hywel Williams, and Ioan Notingher
Biomed. Opt. Express 8(12) 5749-5766 (2017)
Cited By
You do not have subscription access to this journal. Cited by links are available to subscribers only. You may subscribe either as an Optica member, or as an authorized user of your institution.
Contact your librarian or system administrator
or
Login to access Optica Member Subscription