Linear and Quadratic Classifier to Detection of Skin Lesions "Epicutaneus"

Spectrograph-based spectral imaging systems provide images with a large number of continuous spectral channels per pixel, while a gray-value image typically reflects the light intensity over a part of the electromagnetic spectrum in a single band, and a color image reflects the intensity over the red, green, and blue parts of the spectrum in three bands, increasing the number of bands can greatly increase the amount of information in an image. Hyper spectral images commonly contain about 100 to 300 bands with resolution save of 1 to 10 nm. Principal Components Analysis PCA, is one of the most important methods for blind source separation, which we used to estimate the most important components. That means we reduce the dimension of the image-data and the size of the processed data, to keep the operating time short and the coast of the processing low. To separate different tissues of skin lesions we applied the linear- and quadratic discriminant classifiers LDC and QDC. in this experiment we used a training set of 4-different skin lesions (allergy test lesions) and other data group for testing with LDC and QDC. the results provide information about the LDC and QDC as suitable and promised methods for skin lesion classification classification.