作者: Sunpreet S Arora , Mayank Vatsa , Richa Singh , Anil Jain , None
关键词:
摘要: Iris recognition has been used mainly to recognize cooperative subjects in controlled environments. With the continuing improvements iris matching performance and reduction cost of scanners, technology will witness broader applications may be confronted with newer challenges. In this research, we have investigated one such challenge, namely images captured before after alcohol consumption. Due consumption, pupil dilates/constricts which causes deformation pattern, possibly affecting performance. The experiments performed on “IIITD Under Alcohol Influence” database show that pre post consumption images, overlap between genuine impostor match score distributions increases by approximately 20%. These results a relatively small suggest about five under influence able evade identification recognition.