Extended iris color features analysis and cluster headache diagnosis based on support vector classifier

Our long-term is to build an automatic diagnosis detection system of Cluster Headache, whose signs are often noticeable in terms of changes in chrominance and luminance of the iris from the induced pain side. We used a Support Vector Classifier in order to detect differences among 45° direct frames pigmentation of iris by using the error probability as a representation of color differences between both eyes pigmentation. The classification performance was compared when considering different color spaces, in such a way that the error probability provided by the classifier when comparing the iris color of 45° direct frame of both patient eyes provides with a quantitative headache diagnosis measure. Systematic tests were performed on 11 patients images database. We conclude from this work that the iris color features through statistical learning emerges as an interesting technique to study disorders affecting the sympathetic system.