ALGORITHM FOR DETECTION OF USER ABNORMAL BEHAVIOR BASED ON PUPIL RECOGNITION

摘要

The present invention obtains higher reliability by obtaining information on an initial collection group through machine learning for the diagnosis of abnormal behavior, obtaining measurement information on the pupil position of an actual driver based on an AD RANK technique, and comparing and analyzing the information obtained based on the RANK technique with the information obtained through the machine learning to generate discrimination criteria. In addition, an algorithm for acquiring the information of an actual driver based on an AD RANK technique according to an embodiment of the present invention may generate a region that is a criterion based only on the information of an actual driver collected in real time and detect abnormal behavior.