Numerical analysis of endoscopic autofluorescence imaging for cancerous tissue detection

This paper presents an application of numerical analysis to autofluorescence imaging for pathological tissue (mainly cancerous) tissue detection. The analysis was carried out using an application created in MathWorks Matlab. Autofluorescence images produced under ultraviolet illumination whose source was a xenon lamp or a semiconductor laser were subjected to processing. Classifying functions were designed and implemented on the basis of information about discrete red and green color values. The suitability of statistical analysis algorithms for the precise determination of areas corresponding to cancerous lesions at different stages of the disease is demonstrated. The classifying functions were also used in the real-time analysis of images obtained from conventional white-light endoscopy.