According to Lei Fengwang: The 2016 Artificial Intelligence Hunan Forum and the unveiling ceremony of the Zixing Artificial Intelligence Institute, many top experts from home and abroad gave us a report at the conference and accepted the report from Lei Feng (search for "Lei Feng Net" public number) ( Search "Lei Feng network" public number concerned) exclusive interview. Professor Sun Zhenan, research direction for biometric identification, pattern recognition, computer vision. He has published more than 50 papers in international academic journals such as IEEE Transactions on PAMI and CVPR and other international academic conferences. He has applied for 13 invention patents and obtained 6 software copyrights. The “Iris Image Acquisition and Recognition Technologyâ€, one of the major accomplishments, won the National Technological Invention Second Prize and the Eleventh China Patent Award Excellence Award, and served as the International Association for Pattern Recognition (IAPR). Secretary General of Feature Identification Technical Committee, Chairman of Asian Biometrics Forum Evaluation and Standards Committee, IEEE Member, International Conference ICB2009 Publicity Chair, ICB2007, SPIE-DS2008, BSYM2007, IWCF2008, ICWAPR2008, ICWAPR 2009 and other international conference program committee members. The scientific research aims at the major needs of the country and actively promote the practicalization and industrialization of scientific research results. The self-developed iris recognition technology was popularized and applied in the fields of Guo'an, coal mines, banks, airports, etc., breaking the monopoly of overseas iris recognition products, and successfully entering the international market through technology authorization, and achieved a breakthrough in domestic export of iris recognition technology.
1. How high is the accuracy of iris recognition on the camera?Now that iris recognition is very low for the camera, long-distance iris recognition requires only tens of millions of pixels, while close-range iris recognition can be recognized with only a few million pixels. Today's mainstream imaging devices can meet this requirement.
2. Are you currently affected by items such as glasses and contact lenses? How to overcome this effect?This will certainly have an effect, especially if the glasses lens is dirty, causing some spots on the iris texture. In addition, if there is a specular reflection, there will be light spots on the iris. For this issue, we can pre-division these causes by pre-division. The second is to overcome this effect with a more robust recognition algorithm, such as deep learning. Training with a lot of data, because the occlusion is only a small part, at present our recognition algorithm can tolerate 50% occlusion in the iris ring area, so even if there is occlusion or other will not affect the recognition result. In addition, if you wear a transparent contact lens will not basically affect the iris recognition, but if you wear a beautiful pupil, printed with some printed texture, we will detect whether the user wears the United States pupil through live detection, We will ask the user to take off the United States pupil for testing.
3. Where are the main technical challenges in iris recognition? Where is the difficulty of large-scale application?
The challenge is mainly how to facilitate the collection of the iris, because the distance the iris can image is very narrow, and the effective imaging range is about a few centimeters to ten centimeters. In actual use, if the user's posture is slightly wrong, it may not be able to acquire a clear iris. The second challenge is the challenge of robustness and multiple consistent comparisons. For example, if you use a mobile phone to collect the iris, or you use a fixed camera to collect the iris, the data comes from different devices, but you need to compare it. This is the multiplexing. Consistent comparisons. The third is the issue of security. For example, using a printed iris, the template of the iris may be stolen or falsified.
Iris recognition is most suitable for large-scale human use, and now 1 billion people in India are using the iris recognition system. Therefore, large-scale use of iris recognition is feasible. However, when the scale becomes larger, there will still be a problem of calculation speed and separability in one-to-many. However, iris recognition has the highest accuracy with respect to face recognition.
4. How to organically apply this recognition technology in improving automation technology?Automation technology means, for example, in the process of automatic customs clearance, the face or fingerprint is very easy to fake, and the iris is more difficult to fake. It will be safer to unlock the phone or the iris at the iris. For example, there is iris recognition on the Samsung note7, which is bound to Samsung payment. Current payment methods such as Alipay payment will be more secure if it can bind iris recognition. There are still many bottlenecks in the realization of the payment, such as Zhang Santao's account transfer with Li Si. If you can use iris recognition will be more secure, do not have to use the previous send SMS verification code. For example, iris recognition can also be used on a car, so that it does not carry a car key. High-speed automation, especially when personnel security clearance is required, a long queue will be assigned when crossing borders. If iris recognition is used, a machine can automatically identify the identity and match the photos on your e-passport. You can clear customs and speed up customs clearance.
5. How to evaluate the iris recognition technology used on Samsung note7? Does its safety performance meet the standard of iris recognition?We have already purchased a Samsung note7 for testing and it feels good. Represents the best level of iris recognition technology available today. However, in terms of security, we have not yet conducted a complete test. I think iris recognition is definitely more secure than face and fingerprint recognition, but whether it can overcome glass eyeballs or wear contact lenses.
6. Since iris recognition is more secure, why is the use of face recognition applications in finance more widespread than iris recognition?I think there are several problems. First, the cost of iris recognition is higher than the cost of fingerprints and face recognition. The second is that the user’s perception is not enough. If the Samsung note7 is released, the effect will feel good. There are also iPhone 8 or 9 can use iris recognition, it will stimulate everyone's understanding and use of the iris. The third point is that as iris collection becomes more convenient, everyone will gradually start using iris recognition. Now we are also working with banks to use iris recognition for withdrawals. In addition, there are also issues related to national systems and legislation. The National People’s Congress of our country has already legislated fingerprints into ID cards, but I believe that the next generation of ID cards will probably put the iris in the ID card.
in accordance of the Voltage range of the inverter, the PV boards can be installed in series to meet the input voltage of inveter, and several such series to be installed in parallel, to input into the combiner. The current goes through lightenling protection and breaker equipment, then ouput for connection with inverter. We strongly suggest the users to follow all the necessery safety regulation, take all the safefy masures mentioned in the manual, to lower the possibility of danger, make sure of the personal safety.
Jinan Xinyuhua Energy Technology Co.,Ltd , https://www.xyhenergy.com