ZKTeco applies the ground-breaking 3D surface imaging system and develops its 3D Structured Facial Recognition, which proactively projects light to an object and detects light coding of the object’s surface to analyze the disparity between the unique projected patterns and the observed patterns deformed by the surface, the systems enable higher accuracy of facial recognition, better functioning in weak light source environment and capability to operate from shorter distance.
Introduction of 3D Structured Light Facial Recognition What 3D structured light facial recognition is with constant advancement of high definition and high-speed sensor, 3D surface imaging technology has achieved a new height in biometric technology. the 3D structured light surfacing imaging technology has been applied to the industry of access control.
ZKTeco fuses its facial recognition technology with 3D structured light technology, which not only acquires vertical and horizontal details but also depth, considerably reducing limitations on the perception and understanding of real objects.
3D structured light facial recognition is advance technology, it assesses the depth and surface data of the objects by projecting light models with a high-speed camera system.
• Higher precision
• Better performance in weak light environment
• Highest performing in a short to med distance
Depth is calculated with Speckle Images with special coding
• Actively projects speckle light spots to assess depth, which is less impacted by the external light source environment
• Higher preciseness with specially coded speckle images
The 3D depth is calculated with the visual disparity of the imaging by the duo-lens camera.
• It requires pairing of the resolution difference of two images, as the duo-lens camera performs passive imaging, the modifications of external light source environment have great influence.
• Low preciseness due to calculation with raw images.
• The binocular algorithm has huge computational amount and high requisite of processor. It requires high cost for high preciseness in embedded products.
Generally, the more laser spots on facial recognition, the more accurate of facial recognition detection is, most of the facial recognitions offer 68 focus spots, ZKTeco’s 3D Structured Light Technology offers 30,000 laser spots on face which are spread around eyes, mouth, nose and jawline, permitting outstanding recognition flexibility for users and enterprises, ZKTeco Structured Light Recognition are classified as Class 1 laser product with a safe design, use and execution of laser to minimize the risk of laser accidents, especially those involving eye injuries.
• Anti-Spoofing With applied structured light facial recognition, Anti-spoof function has been significantly improved and is able to 100% against fake faces, photo, and video attacks, which is promoted from the power of structured light to perform liveness detection by the 30,000-laser spot.
• Anti-Fake Photo and Video Attack Anti-spoofing function effectively avoids interlopers to access the locked areas of a building, as it is able to recognize real human faces and photos, it prevents unauthorized attempts of access by showing fake photos and videos to cheat the system.
Advantage of using 3D structured light
At present, there are 3 prevalent 3D imaging technologies, which are binocular stereo vision, ToF (Time of Flight) and structured light. Seeing the highest efficiency of light projection, ZKTeco has applied of the 3D structured light technology.
• Higher precision Taking advantage of using 3D structured light technology, it always achieves higher accuracy than TOF and Stereo Vision by discovering the depth of image and comparison with facial features.
• Better performance in weak light environment As an active technique of 3D structured light, it works well in conditions of lacking light, compared with Time of Flight method
• Highest performance in a short to med distance Structured light camera is able to capture an object up to 4 meters, which is suitable for access control that requires a high accuracy of facial recognition in short to medium range.