According to foreign media reports, if you often play cards, then subtle changes in your face may reveal important information to other players. At present, computer researchers at the University of Oulu in Finland have designed a program algorithm that can accurately capture the subtle changes in human facial expressions - "micro-expressions", so the robots assemble the program algorithm, which will have a keen observation ability, even surpassing humans.
â–² Finnish scientists' latest program can identify subtle expression changes hidden by people, especially the micro-expressions of poker players can leak information to other players.
The program uses "robot vision" to observe subtle changes in human facial expressions, which were previously magnified to reveal hidden true expressions. In the test, the program recognizes these expression changes 10% more accurately than human testers. The programming team pointed out that this is the first time to build a real version of the micro-expression database.
Usually micro-expression recognition is found by observing photos. Its accuracy is low, and it is difficult to find subtle expression changes in human faces. To solve this problem, the researchers asked 20 people to participate in the experiment, let them watch a video, triggering different emotional reactions.
They used a high-speed camera that took 100 frames per second to record the tester's facial expressions. The researchers used a penalty mechanism to encourage the tester to try to hide his expression. If the tester's facial expression leaked too much information, then he A long, boring questionnaire about emotions will be conducted. In this way, the researchers took 164 micro-expressions and repeated analysis and identification to classify these expression definitions as: happy, surprised, and depressed. Based on these micro-expression databases, this program is able to identify and analyze people's micro-expression features.
The researchers used each person's neutral expression as a baseline and used this baseline to compare other photos. This way the robot can recognize and analyze some of the subtle expression changes of people. In order to better understand the true meaning of micro-expressions, the researchers used a technique called "motion zoom", which can magnify subtle expression changes, making micro-expressions more specific and easier to recognize.
In the final experiment, the researchers placed the identification instrument in front of 15 people and let them watch the same video. The results showed that the instrument with this program had a micro-expression recognition capability that was 10% more accurate than humans.
Bulkhead Lamp,Led Bulkhead Lamp,Round Bulkhead Light,Bulkhead Ceiling Light
Changxing Fanya Lighting Co.,Ltd , https://www.fyledlights.com