FACIAL SOFT BIOMETRICS - Library of Ph.D. Theses | EURASIP

89and office lights, daylight, flashlight and a fluorescent table light. Clear shifts in the color spacecan be observed. For conciseness and to avoid overlap we portray only 5 subjects. It is clear thata robust eye color categorization technique must consider and cope with illumination variations.A possible solution is to estimate the illumination condition. A self suggesting illumination estimationmethod is proposed by [DYC06], where the color of the sclera serves as estimator of theROI.7.4.2 Glasses influenceThe presence of glasses is another interfering factor, primarily examined in the context of facerecognition and naturally of importance in the current eye color categorization study.Figure 7.7: Eye colors behavior with and without glasses (constant illumination).The subjects were asked for this test to wear 2 different pairs of glasses. It is interesting tocompare this graph with the illumination variation one (Figure 7.6) in order to comprehend theimmense drift eye glasses cause in eye color. It is evident that a stable eye color categorizationsystem should include a priori glasses detection. To detect the presence of glasses in an efficientand robust manner, we can perform histogram normalization, followed by Canny edge detectionon the area between the eyes. A further line detection indicates the presence of the frame part ofthe glasses. This algorithm was deduced from [JBAB00]. For eye color classification in the case ofpresence and absence of glasses the eye color classifier should be able to estimate and compensatethe color shift of the estimated values.7.4.3 Consideration of left and right eye colorWe here show that the strong illumination influence has not only impact on images capturedunder different illumination conditions, but also on the color perception of left and right eye.Although none of our subjects has the seldom condition of heterochromia (different iris colors ofleft and right eye), a drift between the colors of left and right eye can be observed. The illuminationfor this graph was constant daylight falling sidely on the face of the subjects, to achieve in order amaximum illumination difference between left and right eye.