According to foreign media reports, during the new crown epidemic, many people wear masks when they go out, which makes it difficult for people to distinguish the expression on other people’s faces. But a new system is reported to be able to do this, using cameras mounted on their headphones. The device, called C-face, was developed by a team led by Professor Zhang Cheng of Cornell University. It combines two RGB cameras connected to microcomputers and is mounted under the subjects’ ears on third-party headphones. < / P > < p > the system can determine the current positions of 42 key facial feature points by analyzing images of changes in the contour of a person’s cheek (the back of which is captured by a camera). The data is then used to determine the current shape of their mouth, eyes and eyebrows – these combined shapes make up their overall expression, which is represented by one of eight corresponding expressions displayed on a computer screen. < / P > < p > in the test of nine volunteers, the new technology was used in conjunction with state-of-the-art settings that track the location of facial landmarks. The latter is realized by the front camera, which can capture the whole face image. Compared with this system, the error range of C-face is less than 0.8 mm. In addition, the expression it shows can also reflect the real expression of people, with an accuracy rate of more than 88%. With the further development of the system, this number should increase. Global Tech