Phone:+86 181 9406 4618
1. Dynamic face detection and tracking recognition algorithm based on video stream;
2. Binocular (aperture :F2.0, anxious :6mm) dynamic camera with RGB supplementary light and automatic white balance width;
4. 8-inch, IPS LCD,350 LU, 800*1024 resolution;
5. CPU4 core, 1.8ghz, 2GB memory, 8GB storage;
6. Support camera capture and face activation device;
7. Power consumption: 10W MAX;
8. Equipment specification: 240*120*24 (mm);
9. Support device local storage of 10000 face library;
When the face database is 3000, the recognition accuracy of 1:n is 99.7% under the condition of error rate of 3/10000.
Fast recognition, (a) about 20ms for face tracking and detection; (b) about 200ms for face feature extraction; (c) 0.2ms for face comparison (1000 people database, mean value for multiple recognition); 0.5ms for 10,000 people database, mean value for multiple recognition);
10. Support the saving of on-site photos during face recognition or stranger detection;Support the function of stranger identification comparison;
11, support HTTP interface docking, support public network, LAN deployment;Support SDK secondary development.