Skip to Main Content
Hyperspectral imaging is a novel technology for obtaining both spatial and spectral information in the area of aerospace and military industries for last 20 years. The multifusion, multispectral, hyperspectral and polarimetric imaging can provide both anatomical and physiological or metabolic information; it can play an important role in the development of molecular imaging technologies with enhanced specificity and sensitivity, capable of identifying the presence versus absence of cancer, detection of margins, stage, distribution, and type of cancer, but most importantly, techniques capable of assessing the progress of the disease and its response to treatment. Hyperspectral Image modeling provides practical alternatives for the field measurements in human tissue sample due to sample coupling problems and heavy absorption due to protein and water in UV and NIR regions respectively in the region of therapeutic window. In this paper we have proposed the Lorentz Oscillators techniques to generate the multivariate signal for a pixel signature over spectrum channel of interest. The model has been demonstrated to synthesize the image with proper signature for the pixel. The spectra having customized signatures are generated using proper values of the central frequency, Strength of Oscillators and Width of oscillators. These models can be customized for the specific vision under observation and serialized for the quality, intrusion or diagnostics using pattern recognition principle for both anatomical and physiological information in human tissue.
Date of Conference: 5-7 Jan. 2012