Abstrato

A HIGH SENSITIVE APPROACH FOR GENDER PREDICTION BY USING PUPIL DILATON

Ms. Priyanka Thapak , Prof. Ajit Kumar Shrivastava

Pupil dilation is rarely analyzed in usability studies although it can be measured by most video-based eye-tracking systems and yields highly relevant workload information. Algorithms developed by the researchers for recognizing gender by their Pupil dilation patterns have now been tested in many field and laboratory, producing no false matches in several million comparison tests enabling real-time decisions about personal identity with extremely high confidence. The variety of factors that can influence pupil dilation and the distortion of pupil-size data by eye movements yields the size of the pupil as seen by the eye-tracker camera depends on the person's gaze angle. The high confidence levels are important because they allow very large databases to be searched exhaustively without making false matches despite so many chances. In the present study, we developed and implemented a neural-network based calibration interface for eye-tracking systems, which is capable of almost completely eliminating the geometry-based distortion of pupil-size data for any human subject. It also helps, for providing more security to the information.