Application of the principal components transform to colposcopic color images

Many elementary images have to be recognized in each one of the three colposcopic tests, to give a gynaecologist at first the opportunity to build his colposcopics tables, and in the second place to establish the basic elements to make a semi-automatic system for pathological detection. The first step is a segmentation on the basis of the color, and we have implemented the algorithm of the principal components transform suggested by Umbaugh et al. (1993). Our first results are encouraging and are still in evolution, as presented in this paper.