The past two years, with the rapid development of technology, face recognition has gradually become an important field of biometric technology applications of the new era, Forgot your password? All right, we can "brush face"! Today, Xiao Bian will take you to find out the latest face recognition technology to see where to step up the development of this technology.
The traditional face recognition technology is mainly based on the visible image recognition, people are more familiar with this means of identification. However, the disadvantage of this approach is actually very clear, very large light restrictive, and can not meet the actual needs. Problem solving three-dimensional light image recognition, face recognition, and thermal imaging. But the two technologies are far from mature, identify unsatisfactory results.
Quickly developed a solution that is based on face recognition technology multi-source initiative to near-infrared images. It can overcome the effect of changes in the light, have achieved excellent recognition performance, overall system performance in terms of accuracy, stability and speed over a three-dimensional face recognition. The technology has developed rapidly in the last couple of years, gradually the face recognition technology to practical use.
There are three main programs on the market, they are structured light 3D program (StructuredLight), ToF3D program (TimeOfFlight, ToF techniques) and binocular stereo imaging scheme (StereoSystem).
3D structured light (StructuredLight)
The basic principle of structured light 3D technology is, by near infrared laser, having certain structural characteristics of the light projected onto the object being photographed, then collected by a dedicated infrared camera. Such structures have a certain light, because of subject areas of the object at different depths, and different images acquired phase information, by the arithmetic unit and this configuration change information in terms of depth, in order to obtain three-dimensional structures. It is simply to obtain a three-dimensional structure of the object being photographed by optical means, and then to get more in-depth information applications.
This program can be expected in the future structure of the optical advantage or promising, can expand a wider space, such as 5G-related communications fields, AR as well as with other 3D modeling can play a critical role it unique. But 3D structured light has its limitations, as compared to a conventional camera, the light from the 3D structure of the work to be shorter, the longest distance is generally about 1 m, the data given are 60cm far OPPO accuracy can be controlled at ± 1mm range, of course, the distance to face unlock is more than enough.
Using structured light 3D face recognition technology as mobile phones have what?
iphonex/xr/xs/xsmax
Last year iphoneX listed can be described as a global cause everyone's attention, because Apple represents the future of mobile phones look like iphoneX, naturally brings together Apple's top technology. IphoneX say that we remember him most worthy of black science and technology, of course, is Apple's friends faceId face recognition function, Apple's faceId not only give unlock the screen, and its level of security already reached the level payments, also released this year the new machine in a series of natural face recognition has become standard. Apple's invention of this technology in the forefront Andrews camp, but also to the Android mobile phone manufacturers a clear direction of technological innovation.
oppoFindx
One year later, the efforts of the Android mobile phone manufacturers, Android phones have begun to use the same Apple faceid similar technology, such as the O-Face3D oppofindx structured light technology, oppo official said the speed unlocking process takes only 0.5s time, FindX the 3D camera can emit 15,000 points of light, the fault rate of up to a million, far more than ordinary 2D and face the existing fingerprint recognition, has now also support Alipay and micro letter paid.
mate20pro
In addition to oppoFindXdeO-Face3D There is also a mobile phone manufacturers Andrews also developed a structured light 3D face recognition technology and has been put into use in their own flagship product that Huawei and Huawei's own flagship mate20pro. Huawei mate20 3D structured light technique carry their own research, high-precision face recognition technology and payment security technology, while supporting Alipay payment and micro-letters, through the lattice projection, not only to improve the security of secure identification, and also supports the fastest 0.6 s unlock speed. While 3D structured light technique can be used with high-definition front camera to achieve 3D modeling, the reality of the object into your phone doll, very high playability.
ToF（TimeOfFlight）
ToF is one embodiment of 3D depth camera, a structured light with the door Young. ToF ranging single point and multi-point two of them with multiple phone usually ranging. Similar single multi ranging ranging pulses and principles, but the received light to the CCD device, i.e. with a photodiode array of charge retention, in response to light having an integral characteristic. The basic principle is a certain viewing angle laser light source emits laser light, wherein dt is the length (from t1 to t2), each pixel of the CCD using two synchronized trigger switch Sl (t1 to t2) laser, S2 (t2 to t2 + dt) is controlled each pixel element of the charge retention time of collecting the reflected light intensity, the response obtained C1, C2. Distance of an object distance of each pixel = 0.5 L c dt * C2 / (C1 + C2), where c is the speed of light (which can remove the influence of formula reflecting differences in the properties of the reflector ranging). Is simply, light emitted through a process of, after the encounter objects will be reflected back and forth capture time, because the speed of light and the wavelength-modulated light is known, it is possible to quickly and accurately calculate the distance to the object.
TOF technology advantage is that the overall smaller components, do not leave big bangs, recognition distance farther, you can develop more applications; calculating the depth of field in the screen shot no post processing, popular talk time to avoid delays and save costs associated with use of an aftertreatment system to bring powerful; not easily affected by external disturbance light, compact, fast response and high accuracy identifying multiple advantages. The disadvantage is the high cost of hardware, is mainly reflected in the price of the assembly, the photon detector array, a laser and detected by the sensor, they require very high-end device. Technically the high cost of hardware, software algorithms coupled with cost even higher, and the current capacity to provide reliable hardware vendors rare, and all this will push the cost. But with the outbreak of the growing demand for ToF technology, popularization and development of ToF technology just around the corner, so decisively cut vivo TOF technology is really very favorable courage, even the strongest Kuo Ming-Ji analysts have predicted Apple's iPhone in 2019 there may be significant ToF innovations including the use of technology, we can see advanced TOF technology.
Binocular stereoscopic imaging scheme (StereoSystem)
Binocular stereo imaging (StereoSystem) using dual cameras photographing the object, then the object distance is calculated by a triangular principle. Huawei has been listed nova3 use is the binocular 3D face recognition program, and is a standard IFAA (Internet banking authentication Alliance) to provide, and this is the main reason for the payment level it reaches. Two RGB nova3 mounted (color camera) via an analog camera binocular vision system to identify a user equipment.
The simplest principle of binocular program should be a relatively early recognition program, which costs a minimum. All rely on the depth information acquired camera images acquired by software algorithms to derive, depending on the accuracy required capture resolution. And because of this embodiment relies on image analysis algorithm to obtain depth information, calculate the maximum load, the highest algorithm complexity, difficult to achieve, recognition slower. In addition, it will also be affected by light, especially under low light conditions.
to sum up
From the market response iPhoneX, millet and OPPOFindX3D exploration version 8 transparent view of the structure of the light of this program is the best, because the consumer experience for all to see, ranging scheme eyes are bound to be reduced to the expedient end models measure, but a fight of TOF technology is worth the wait, but just because Blokus, mirror month, only to wait until the launch of the real machine can judge, of course, also depends on the final test of the market.