1, privacy security
Since it is a face, it must involve personal privacy. With the increasing popularity of information technology, personal big data will be retained in different degrees on different platforms. At this time, how to ensure the security of personal privacy on the Internet needs face recognition technology and related laws and regulations to improve and manage, but this part is still in its infancy. It also needs the promotion of the whole society from cognition to technology.
Application scenario
The accuracy of face recognition is highly dependent on the application scenario. Someone once tested a platform-type intelligent robot, claiming to integrate the most NB face recognition technology and camera, which can also have high recognition accuracy in dim light. Once when the robot moved to the door, I walked out of the house. When I walked in front of the robot, half of my face was exposed to the sun, while the other half was in a dark state. At the same time, the angle of the robot is laser shooting, so it failed! The results of many tests are the same. When a person moves from a dark environment to a strong light environment, the camera angle is under strong light, and the recognition effect is very poor, which shows that the adaptability and accuracy of face recognition can be greatly improved.
User Experience
The user experience of face recognition is also a challenge. Sometimes I feel embarrassed when I pass the airport security check. When everything is running normally, you can't get through the identification channel. There may be several reasons: 1. Because of network transmission problems, the most accurate photos cannot be uploaded in time, which leads to the slow display of test results; 2. When the user has undergone cosmetic surgery and the facial contour and depth features have changed greatly, it is possible to detect errors; 3. The recognition range of the detection equipment is limited (too many faces will appear when it is too large, and the framing of too small faces will be incomplete). At this time, users need to bend over, pad their feet, or move to adapt to the camera, which is like someone carrying a bicycle when climbing a mountain! Because of the wide application of character recognition, how to provide the best user experience in the whole scene is still one of the challenges faced by character recognition.
Other implementation algorithms, efficiency and accuracy can be gradually solved with the application of big data and the strengthening of computing power, but it is still a local challenge!
Wan Wei appraisal answers for you.