If you have seen the American drama "Lietome", you will not be surprised by the setting "Identify the perpetrator through the identification of micro-expressions." In fact, since the invention of photography by humans, the academic and artificial intelligence community has officially opened the brain about "how to identify criminals through facial recognition technology": The advent of the photographic era has led some 19th-century scientists to believe that machines are absolutely capable of accurately finding criminals by recognizing facial features. Although their hypothesis was eventually labeled as “untrustworthy†and “absurdâ€, after 200 years, this once “paradox†may restore their reputation: A new generation of artificial intelligence technology may prove that their comments are well documented. Shanghai Jiaotong University: The accuracy of the neural network to identify prisoners is as high as 89.5% In China, researchers from Shanghai Jiaotong University, Wu Xiaolin and Zhang Xirang, return to the public's field of vision, the controversial traditional issue of “recognizing criminals through facial recognitionâ€: They built a neural network that was said to find out the prisoner simply by verifying facial features. In their paper, in order to prove the feasibility of this theory, the researchers mixed a large number of suspects and non-offender photos, and then used a large number of machine vision algorithms to test these people's facial expressions, the goal is to discover a set of Reliable identification of the prisoner's neural network system. In the process, scientists need to infuse 1856 photos of Chinese people without facial hair (including eyebrows, mouth hair and beards, etc.) between the ages of 18 and 56 to the neural network. Half of these photos (about 730 have been identified as criminals) have a criminal record. Cm Nozzle,Panasonic Cm Nozzle,Smt Cm Nozzle,Panasonic Smt Cm Nozzle Shenzhen Keith Electronic Equipment Co., Ltd. , https://www.aismtks.com