News

With the wide application of face recognition system, its product categories also tend to be more diversified to fit more demand scenarios. Take the Surveillance Camera, an important component of face recognition, for example, all kinds of products not only have different prices and styles, but also have monocular and binocular distinctions. What is the application for the face recognition function?

Choose monocular or binocular? It depends on the specific needs of in vivo detection

Monocular camera is mainly RGB camera, with monocular algorithm can quickly complete the acquisition, high quality images to the back-end for identification comparison, suitable for face arming, accurate promotion, passenger flow statistics, access control management and other application scenarios. Such cameras with appropriate live detection algorithms, you can do monocular live detection. For example, ordinary surveillance cameras through the introduction of open platform free open face recognition algorithm, you can achieve a silent live body recognition, the whole process without user action with. In addition with the face recognition algorithm, it can also quickly achieve a full set of functions such as face recognition, age detection, gender detection, face recognition under large area occlusion.
The binocular camera is based on the RGB camera, adding an IR infrared lens, both near-infrared light and visible light video capture function. Based on the principle of infrared imaging, the screen class can not be imaged, so it inherently has the ability to resist the screen imaging fake face attack.
Therefore, compared to monocular cameras, binocular cameras are better in the defense against fake face attacks, and the defense against paper photos and screen photos is quite excellent.
However, due to the difference in hardware, the cost of binocular live is relatively elevated compared to monocular RGB live. In general, monocular face recognition solutions are easy to deploy and low cost, while binocular face recognition solutions are more accurate in live detection and more adaptable to factors such as light changes and complex background environment, and can support true unattended. Therefore, in the selection process, specific considerations should be made according to the actual application scenario, the live effect you want to achieve, and the final cost, balancing the needs of all parties and not making generalizations.

Camera imaging core components and key parameters

The working principle of the camera is to convert the light signal into image signal, so strong light, backlight, dark light and other complex light environment will directly affect the camera imaging quality. The quality of camera imaging will also directly affect the effect and experience of face recognition and in vivo detection. Therefore, during product selection, attention should also be paid to the core components and key parameters of camera imaging to ensure that the image quality collected at the front end meets the face recognition requirements.
Among the three core components of face recognition camera module, namely lens, image sensor and DSP image processor, the lens mainly undertakes the task of capturing the object and focusing the captured image on the image sensor, whose main parameters are field of view, aperture, CRA, aberration, glare, etc. Face recognition camera module generally uses a fixed focus lens with fast focusing speed, stable imaging quality and accurate metering.
For the image sensor responsible for converting the light signal into digital signal output, its main parameters include sensor size, effective pixels, pixel element size, etc. CMOS sensors with high integration, low power consumption, fast speed, low cost, etc. are recommended. The last is the DSP image processor that transmits the digital image signal optimized processing to the face recognition equipment, and the high-quality DSP can improve the original chip wide dynamic range by more than 30%.

How to test the basic function and reliability of the product?

To ensure that the product can be used smoothly after landing, testing is one of the essential parts. In the case of camera module, its basic performance test includes resolution test, splash test, line test, off-color test in normal shooting test environment and dark corner test, dirty test and bad spot test in whiteboard test environment.
Among them, the resolution refers to the clarity and image detail expression ability that can be achieved by the camera module. A more intuitive evaluation method is to visually identify the black and white line pairs in the chart by shooting the ISO chart test and see the limit line values that the human eye can identify the black and white line pairs. In the same environment, the image should also be checked for abnormalities such as splashes, lines, and off-color. And in the whiteboard test environment, the dark corners are usually checked to detect the brightness uniformity of the module, to detect any dirty abnormalities within the lens, and to detect any bright spots and bad spot abnormalities in the chip.
In all these links, testing is not only to propose bugs, but also an important step for development and product design improvement, so the testing work often comes with extreme professionalism and rigor for product effect.