What are the identifiable drawbacks of using iris for classification and recognition process?
In general, we may speak about: eyelids recognition, noisy samples like flash light spot inside iris region, and partially occluded eye.
Illumination dependency
Spatial displacement (I used for gaze tracking)
In some cases, it is tedious to perform an iris scanning due to the presence of reflections. It could happen in case of eyelashes, lenses, and anything in general that would cause a reflection. These aren’t very uncommon in our society, so people would face some difficult during iris scanning process.
Also, Iris scanners are relatively higher in cost compared to other biometric modalities. Moreover, A person has to be steady in front of the device to be enrolled by iris scanners. It means this device can’t be used like face recognition devices to scan anybody, regardless of their movements. Sometimes it is quite difficult to be steady enough to complete the scanning process in first attempt.
These 2 reasons may be drawbacks:
1. LOCATION: its location has made it susceptible to forgery has high-quality photograph of iris images captured at a distance of few centimeters to meters and even on the motion have been reported to be used in developing synthetic images that have been used to spoof an iris recognition system
2. High Error Rate: Iris has a high error rate of 1 in 131,000 individual
AKANDE NOAH OLUWATOBI Your first point as drawback for iris recognition system is quite informative and revealing. However, don't you think having an error rate of 1 in 131,000 subjects isn't a bad system as the reason for such error maybe attached to some factors? Can I have more clarification on this as regards the implication of have 1 in 131,000 error rate
High error rate in Iris recognition system can be reduced, as many researchers have proposed numerous approaches in that regard. Moreover, scientific and academic research have long ago proven that high error rate in Iris recognition system can be reduced through the use of multiple classifiers. Moreover the advent of Machine and Deep Learning have really contributed a lot to the eradication of high error rate in Iris Recognition System.
Halleluyah Aworinde The error rate is on the high side when compared to other biometric traits. Retina recognition has the best error rate of 1 in 10,000,000 compared to iris recognition of 1 in 131,000, fingerprint of 1 in 500+, speaker recognition of 1 in 50 and hand geometry of 1 in 500. You can read more in the attached article.
Oluwashina Akinloye Oyeniran Yeah, research is not stagnant, definitely, that would have called for more research efforts. I will appreciate if you can drop a recent article to back up the claim. Academics is all about finding new things and backing it up with claims
- Badges
- Science topic
- Similar topics
- Data Mining
- Classification