Facial Recognition
Deep learning algorithm independently
developed by KANKAN AI
-
Face Detection And Tracking
• Face detection is a key part of dynamic facial recognition system. It can accurately detect human faces in the environment with complex background and open space, to detect and track faces in real time with external interferences such as obscure or shaded side faces.
• Applicable for face tracking in video streams, avoiding repeated capture recognition and improving efficiency. -
Face Retrieval
• Dynamic: Support functions such as video-stream-based face detection and tracking, feature extraction and comparison retrieval with low performance consumption, and it could complete retrieval of a million-face database in 30ms with a single-core GPU.
• Static: Support static image face detection, feature extraction, attribute analysis and ten-million face database retrieval. -
Face and ID Card Comparison
• Face photos were collected by image acquisition equipment and compared with id photos one by one for face information, and the accuracy rate was more than 99% under one in a million false recognition rate.
• It is available for scenarios such as airport or train stations, financial fraud prevention, visitor registration. -
Face Attribute
• It could analyze basic face attributes in real time, including the characteristic attributes such as gender, age group, glasses, etc.
• Applicable for face data analysis. -
Anti-Spoofing Detection
• It could detect whether a user in front of the camera is a real person. This can effectively resist common attacks such as photos, face changes, masks, covers and screen copying, thus to helps users to identify frauds and guarantee user benefit.
• It has been widely used in application scenarios such as access control, mobile devices, and financial fraud prevention.
• Support structured-light camera, binocular infrared camera, monocular camera.