In early 2017, huiding technology demonstrated the case of off-screen fingerprint identification through two modified mobile phones, Samsung Galaxy S7 edge and vivo xplay6. This fingerprint identification technology without opening a hole on the screen has attracted the attention of countless players.From "high and low" at the beginning of 2018 to the end of 2018, oppo will enter the K1 supporting screen fingerprint into the popular price of 1599 yuan. This "black technology" is destined to become the "standard" in the field of medium and high-end smart phones in 2019.So what are the differences between the early screen fingerprint and the latest screen fingerprint scheme? Why are new products armed with this technology getting cheaper and cheaper, while the accuracy of fingerprint identification has been improved?
Development history of screen fingerprint technologyFrom the technical level, screen fingerprint identification can be derived into two categories: "optical" (also known as "photoelectric") and "ultrasonic".Unfortunately, the ultrasonic screen fingerprint scheme has just been trial installed by Samsung Galaxy S10 series (the feedback from netizens' actual experience is mostly slow and low recognition rate). The screen fingerprint from 2018 to early 2019 is still the world of optical screen fingerprint scheme.
Theoretically, the optical screen fingerprint technology can be infinitely close to the traditional capacitive fingerprint identification technology in terms of recognition rate and unlocking speed. Taking the latest off screen optical fingerprint module (CMOS scheme) of huiding technology as an example, it is applicable to both OLED soft screen and OLED hard screen (ultrasonic scheme can not be used for hard screen), and has created a separate design, so that the optical fingerprint module no longer needs to be attached to the OLED screen (ultrasonic scheme needs to be attached to the screen), which greatly reduces the commercial difficulty, It effectively reduces the hidden cost and risk cost of production.
Therefore, optics has a greater cost advantage than ultrasound. Screen separation can effectively reduce the production cost, and it adopts a fully mature industrial chain.Optical screen fingerprint technology needs light emitting and sensing devices. Theoretically, TFT (LCD) and OLED screens can be used as light sources. However, the TFT screen is passively illuminated, and the LED backlight at the bottom needs to pass through the TFT panel to display. If you want to realize the screen fingerprint identification technology in the TFT screen, you need to perform a "major operation" on the whole module, and the cost is very high. In contrast, OLED can accurately control each sub-pixel because of its ultra-thin and self luminous characteristics.The general principle of screen fingerprint is that when the user slightly presses the screen with his finger, the RGB pixel of the OLED screen will emit light and illuminate the fingerprint texture through the small hole of the OLED screen to the screen surface. Because different lines of fingers lead to different reflected light, the light reaching the fingerprint reflects through the screen and then reaches the fingerprint sensor. Finally, the fingerprint sensor can form a fingerprint image according to the reflected light.
Take vivo as an example. As early as January 2018, vivo X20 plus, the world's first screen fingerprint identification mobile phone, was mass produced. Then X21, nex, x23, nex dual screen version and x27 were armed with the second, third, fourth, fifth and sixth generation photoelectric screen fingerprint technology respectively.In fact, the iterative update of optical screen fingerprint identification technology is not so fast. In essence, it has just experienced two generations of innovation.First generation: off screen fingerprint sensor scheme
The first generation of screen fingerprint is also known as the "collimator scheme". For early mobile phones that supported screen fingerprint identification, such as vivo X20 plus, X21, Xiaomi 8 exploration version, Huawei mate RS off screen fingerprint version and Meizu 16th, they all embedded fingerprint identification sensors under the OLED screen to collect light imaging from the small holes of the OLED screen through the micro lens, so as to realize the identification of fingerprint information.Considering that the pixel arrangement of OLED may produce moire marks, which may affect the recognition, most of the fingerprint sensors of this kind of mobile phones use the method of tilting a certain angle to eliminate the moire marks.In order to get a clear fingerprint image, the "collimator scheme" also needs to add a "collimating layer" between the sensor and the OLED screen to filter out the refracted and scattered light and ensure that the light reaching the photosensitive element is collimated light. In addition, based on the collimation layer scheme, a "small hole imaging scheme" is derived, which can further reduce the module thickness and cost, and make the fingerprint image reaching the light sensor clearer.
Unfortunately, the off-screen fingerprint sensor scheme is not perfect. Firstly, the module of this scheme must fit closely with the OLED screen, which is difficult to process and has the problem of yield. Secondly, the cost of the scheme is directly proportional to the effective fingerprint identification area. The larger the area used to collect fingerprints, the higher the cost.The fingerprint sensor can be seen whether the screen is off or onIn addition, in order for the sensor hidden under the screen to receive the light penetrating the screen to the greatest extent, it is necessary to improve the light transmittance of the screen. Therefore, all mobile phones adopting the "collimator scheme" have a common problem - the screen can vaguely see the outline of the fingerprint sensor in strong light, which greatly affects the view.
Second generation: off screen camera sensor schemeStarting from vivo nex, most of the new products represented by oppo R17 and K1 adopt the second-generation screen fingerprint identification technology. They replace the fingerprint identification sensor hidden in the screen with a special camera module (composed of fingerprint chip and ultra short focus CMOS)Fish eye camera, light sensor, nor flash and filter are also called "CMOS scheme", which essentially completes fingerprint acquisition and comparison through "off screen camera".
Compared with the first generation "collimator scheme", the module of "CMOS scheme" does not need to be combined with OLED screen. It can be fixed on the frame, and the process difficulty and cost are greatly reduced. In addition, the scheme does not need to consider the problem of collimating hole. It can further improve the imaging quality by improving the pixels of the camera. It is also difficult to see the fingerprint module (lens) in the screen under strong light.In early 2018, the rejection rate of the "collimator scheme" was 7% 8%, but now the "CMOS scheme" has been reduced to 2%, and the unlocking speed has been reduced from 1 second to 0.24 seconds.In terms of price, the cost of screen fingerprint chip during the "collimator scheme" period is about $8 to $9, and the cost of additional modules is about $7 to $9. The total cost of the whole scheme is as high as $15 to $18, that is, about 103 to 123 yuan. Today, the screen fingerprint chip of "CMOS scheme" has been reduced to $6, and the module cost is only $2. The cost of about 55 yuan has laid a foundation for it to become the standard configuration of mainstream price mobile phones (more than 1500 yuan).
"CMOS solution" also has a killer advantage - since its main body is the camera, camera related technologies can be introduced to further improve the imaging quality and recognition efficiency.As a representative of the first batch of mobile phones adopting the "CMOS scheme", vivo nex has added a large number of photography derived technologies such as multi frame synthesis, distortion processing, stability judgment and scratch detection, as well as algorithm optimization for fingerprint images such as feature extraction optimization, dynamic base adaptation, mask mechanism and DPI adaptation.At the x23, vivo, in conjunction with Qualcomm, conducted in-depth optimization at the SOC level, launched the "screen fingerprint DSP acceleration technology", and handed over the fingerprint image enhancement and contrast recognition originally undertaken by the CPU to hexagon DSP, which greatly improved the efficiency and significantly reduced the power consumption.
At the same time, x23 also increased the image density of the fingerprint camera under the screen to 750ppi and the aperture to f / 1.5. Through the later OTA System update, it can further optimize the algorithms such as first frame acceleration, HBM synchronization, water ripple elimination, mask preloading and removal, so that the fingerprint unlocking speed of x23 is further improved from 0.35 seconds at the beginning of release to 0.24 seconds!The vivo nex dual screen version launched last year adopted the "fifth generation photoelectric screen fingerprint", which upgraded the 2p lens to 3P lens, opened a fingerprint processing area on the SOC (CPU), and made another breakthrough in the bright screen unlocking speed combined with DSP level optimization. In short, "CMOS solution" has the characteristics of lower cost of ownership and continuous optimization and upgrading through similar camera algorithms. It has become the most popular optical screen fingerprint solution for new mobile phones.Vivo x27 screen fingerprint recognition demonstration
The newly launched vivo x27 adopts the "sixth generation screen fingerprint HD version", which is characterized by doubling the sensing area of the module sensor. Compared with the previous generation of screen fingerprint technology, the detected finger area is increased by 27%, and the acquired image semaphore and dynamic range are also increased by 30%. Moreover, the larger sensing module makes the sixth generation screen fingerprint perform well in special scenes. For example, in the case of dry fingers and low temperature environment, its recognition efficiency has also been improved by nearly 50%.