3D reconstruction of face recognition
Structured light three-dimensional reconstruction (3D imaging) is actually an active 3D measurement technology for the measurement and reproduction of 3D parameters of the target. Because this technology needs to actively project structured light onto the measured object, and determine the size parameters of the measured object through the deformation of the structured light (or flight time, etc.).
Structured light is the structure of light, and the rough classification can be divided into point structured light, line structured light and surface structured light. At present, the more mainstream schemes are random lattice structured light (speckle) and multi-line structured light. After the structured light is projected to the measured object, the surface of the measured object is highly modulated or temporally modulated to the structured light, and the modulated structured light is collected by the camera, and then the three-dimensional surface data of the measured object can be obtained after analysis and calculation by the software algorithm.
Different from ordinary 3D imaging and 3D reconstruction, face recognition is more secrecy and accuracy, because it is a higher-end structured light 3D imaging application. Concealment is generally achieved by using infrared lasers, while accuracy is achieved by a combination of optics and algorithms.
The iPhone X uses a structured light solution for human eye recognition, and it has been a great success. The principle of structured light face recognition is to project an infrared laser with certain structural characteristics onto the human face, and then collect it by a special infrared camera. The energy distribution of this structured light is carefully designed, and the camera will collect different image phase information due to the different depth areas of the object, and then the 3D reduction algorithm converts the change of this structure into depth information to obtain a model of the human face.
A good structured light 3D reconstruction solution can build a fine and complete 3D model of the user's face, more accurately identify the user's biological characteristics, and the security is far better than fingerprint recognition and traditional facial recognition solutions, almost without interference from the surrounding environment.