The apparatus (100) has one or two sensors (111, 111a, 111b) for capturing a subject’s eyes in one image (305) or two images (302, 303), and processor(s) (113) in a housing (100a, 100b, 100c) with the sensor(s) and/or in a computer system (124). The processor(s) (113) determine a head tilt angle between a virtual line extending between the eyes of the subject in accordance with predefined features ((xpL’,ypL’), (xpR’, ypR’)) in the image(s) and a dimension of pixels in the image(s) characterizing zero head tilt, segments left and right iris images, and rotates the segmented images to substantially remove head tilt when present (905, 906, 1002-1006). The processor(s) (113) generate template(s) representative of one or both the left and right irises of the subject (907, 908, 915, 918, 919, 1007), which are either added (1008) or compared to identification data of templates of subjects (909, 910, 916). The processor(s) (113) may also determine interpupillary distance (IPD) or IPD-to-iris ratio (606) using the image(s), and as such may be used for comparative searching of the identification data having IPD or IPD-to-iris ratio of subjects.