Facebook's 'sadness' recognition ability

Facebook reassures that the new face recognition project can detect faces with accuracy of up to 97.25%, close to the ability of ordinary people.

Steps to analyze and collect information based on prototypes

The emergence of new technology means that users almost no longer have to manually upload names to downloaded images, but Facebook will do it for them.

According to Slate.com , Deep Face technology will create three-dimensional maps of facial features and create colorless models to narrow down the specific features of each case.

Facebook's API team uses 4.4 million faces tagged with 4,030 people using this social network to help the system learn and improve face recognition.

According to Facebook, face recognition consists of four phases: detection, queuing, representation and classification.

Facebook is using 3D modeling techniques to return to the alignment and representation steps, then identify faces using the 9-layer nervous system.

It's not clear when Facebook will officially launch the technology.