Facebook revolutionized the world of social networking. The founder of the social networking site launched the site in his residence while studying at Harvard University, and simple things grew exponentially. They were recently news about the acquisition of WhatsApp messaging services and speculated on their direction. They recently invented software that claims to have better facial expressions than human faces. According to research, a person can assign two photos of unknown faces to 97.53%. This has nothing to do with the lighting situation or whether the person in the picture is facing the camera.
We wrote an article about dangerous social networking sites very early, but Facebook and its competitors have been actively developing software called deep learning. It is an artificial intelligence software that uses complex facial matching methods. The software uses neural networks to help them slowly identify patterns in large data pools. The face recognition rate of the new software has reached 97.25%, which is a significant improvement compared with the previous artificial intelligence software. According to Facebook’s artificial intelligence researchers, it is approaching human performance. Compared with the previous software, the margin of error has been reduced by more than a quarter.
The software uses a 3D model to rotate human faces, so they can face the camera even if they are looking. The researchers called facial examination facial recognition. Face verification can detect whether two images show the same face, and facial recognition software will place the face in the face. The two technologies can be used in combination, which will improve Facebook’s ability to identify and name two faces. Combining these techniques may help suggest which photos to put on, but this technique is also dangerous.
Facebook calls it the DeepFace project, which is at least in the early stages of research. Facebook recently published a paper on the project at the IEEE technical conference in June. The purpose of this publication is to obtain feedback from other members of the technical research community. The software pioneers were Yaniv Taigman, Ming Yang and another colleague Marc’Aurelio Ranzato. Professor Lior Wolf of an Israeli university also participated in the development of the project. Taigman has already expressed a lot of views on the project in public. We at perfectjammer.com hope this project will develop in a positive direction.
The DeepFace project is divided into two stages. In the first stage, the software actually changed the perspective. The correction makes the picture face forward in the picture. The software uses a three-dimensional model of a typical average front surface. In the next stage, neural networks come into play. It calculates the digital description of the face. At this stage, it will compare the numerical data collected from the two images. If there is sufficient similarity in the data, the software will determine that the two images have the same face. If you do not trust this modern technology, we recommend that you choose a portable or fixed gps blocker provided by our store.