Molecular image-based techniques are widely used in medicine to detect specific diseases. The analysis of the eye plays an important role in order to detect specific diseases. Eye background analysis is used in order to detect certain forms of diabetes and others diseases. In the alternative medicine plays the diagnosis of the iris an important role. One understands by iris diagnosis (Iridology) the investigation and analysis of the colored part of the eye, the iris, to discover factors which play an important role for the prevention and treatment of illnesses, but al-so for the preservation of an optimum health. Although alternative practitioner describe substantial success with the iris diagnosis. The conventional medicine is not convinced of the diagnosis method.
A big drawback of the method is the subjective interpretation of what is seen in the iris image. An automatic system would pave the way for much wider use of the iris diagnosis for the diagnosis of illnesses and for the purpose of individual health protection. With this paper we de-scribe our work towards an automatic iris diagnosis system. We describe the image acquisition and the problems with it. Different ways of image acquisition and image preprocessing are explained. We describe the image analysis method for the detection of the iris. This method is based on our novel case-based object recognition and case mining method. Results for the recognition of the iris are given. We describe how to detect the pupil and not wanted lamp spots. We explain how to recognize orange blue spots in the iris and match them against the topological map of the iris. Finally, we give an outlook for further work.