A. IriTech’s Superior Iris Recognition Algorithm

Since iris muscles control the size of the pupil, iris pattern shown in a captured iris image deforms significantly during pupil dilation/constriction. Daugman-based approaches assume the iris deformation is linear when mapping the iris to the polar rectangular format. However, iris deformation is not linear and in fact differs from person to person. To cope with the non-linear iris deformation, IriTech uses its patented “Variable Multi-sector Analysis” method, where the whole iris is divided into several sectors and local deformation of each sector is analyzed separately to obtain the correct sector combination for matching. Such local deformation analysis makes IriTech’s matching algorithm not only resilient to various iris deformation laws but also tolerant to a-few-pixel error in image segmentation (i.e., pupil and iris border detection). The segmentation error tolerance is important since the pupil and iris boundaries are not perfectly circular or elliptical as assumed by most iris recognition algorithms.