Recently, the National Security and Alarm System Product Quality Inspection and Supervision Center (Shanghai) and the Security and Alarm System Product Quality Supervision and Inspection Test Center of the Ministry of Public Security, which are affiliated with the Third Research Institute of the Ministry of Public Security, announced the inspection results of the Minivision Technology face control integrated machine. The host and software of the Minivision technology facial control all-in-one machine have successfully passed 11 tests, meeting the relevant standards of the "Environmental Adaptability Requirements and Experimental Methods for Security and Alarm Equipment" issued by the General Administration of Quality Supervision, Inspection and Quarantine of the People's Republic of China and the National Standardization Administration of China.
The Third Public Security Bureau strictly controls
The National Security and Alarm System Product Quality Inspection and Supervision Center of the Third Institute of Public Security and the Security and Alarm System Product Quality Supervision, Inspection and Testing Center of the Ministry of Public Security are national statutory testing institutions, responsible for various quality supervision and random inspection tasks entrusted by the state and the Ministry of Public Security. Led and participated in the revision of national, industry, and local standards for security and information security products, as well as the testing and verification of relevant standards, and formulated very strict testing standards for "Environmental Adaptability Requirements and Experimental Methods for Security and Alarm Equipment". In the detection process of the Minivision technology facial control integrated machine, the three institutes started with 11 detection items, including device support for image resolution check, device support for frontal facial angle check, pass rate check, modeling speed test, facial retrieval rate test, device software application function check, high and low temperature test, constant humidity and heat test, vibration test, etc., to verify whether the performance of the tested product meets the technical requirements, To ensure the professionalism, reliability, and safety of the tested product in delivery and application.
Inspection and Detection Report of Facial Control Integrated Machine
'Following the Map' in 100000 Portraits
The intelligent security monitoring in the factory area, the "facial recognition identity verification" by the Human Resources and Social Security Bureau, and the "face search" used by train stations with large passenger traffic are lurking everywhere, providing convenience and security for modern society. The ZS-F6000N face control all-in-one machine that passed the inspection this time uses the 1: N face recognition technology suitable for the above scenarios, combined with the silent living body technology launched by the Minivision industry, greatly improving the user experience.
1: N face recognition refers to finding and matching the face data of the current user in a massive portrait database, while silent living entities emphasize the user's non cooperation. There are three main steps involved: firstly, when a live person passes by the Minivision technology face control all-in-one machine, the all-in-one machine adopts a new recognition principle, which is different from the traditional micro expression live recognition mode. Without the need for users to cooperate with micro expressions and other actions, the face image is directly analyzed in real-time through the viewfinder to actively determine whether the operator is a live person; The second step is to compare the captured facial features of a living person in a face retrieval image set constructed from 100000 different person ID images, identify the most similar one or N images, and determine whether they are the same face; Finally, according to different scenario requirements, the all-in-one machine returns user information and matching degree. The entire recognition process is directly obtained through the camera and completed in a non-contact and non cooperative manner, which is convenient and fast.
High speed response and precise identification
The detection data shows that the Minivision technology facial control all-in-one machine retrieves 100000 facial images, with a recognition error rate set at 0.01% and a pass rate of ≥ 98%. The error recognition rate here refers to the proportion of people mistakenly identifying different individuals as the same person, which greatly affects the safety performance of the all-in-one machine. The pass rate refers to the proportion of live bodies correctly recognized by the all-in-one machine, which determines the practical performance of the all-in-one machine. This set of numbers means that when the number of face databases is 100000, the recognition accuracy of the Minivision Technology face control all-in-one machine is ≥ 98%. This high recognition accuracy has made Xiaoshi Technology's face control all-in-one machine rank first among similar image recognition products.
Although the recognition accuracy of the Minivision technology face control all-in-one machine is high, the requirement for face resolution is not high, with a minimum pixel size of only 80 * 80 (pixel). On the premise of a 100000 face atlas, Minivision Technology's facial recognition technology can respond immediately, provide efficient feedback, and track and recognize faces in real-time. In the detection, the response time of the Minivision technology face control all-in-one machine for detecting a single target face is ≤ 1 second. When the image resolution in the portrait database is ≤ 640 * 480 pixels, face pixels are ≥ 80 * 80 pixels, and the number of faces in a single image does not exceed 10, the response speed of face retrieval is>4 images/s. Thanks to lower input pixel requirements, efficient recognition speed, and accurate recognition accuracy, the application scenarios of Minivision technology control integrated machines have become more widespread. For example, during subway security checks, rapid portrait recognition can reduce the waiting time for passengers to queue up, while in providing services for information systems such as security, monitoring, and management of enterprises or government departments, all-in-one machines can promote the implementation of automated management and improve work efficiency.
The detection data shows that the Minivision technology facial control all-in-one machine retrieves 100000 facial images, with a recognition error rate set at 0.01% and a pass rate of ≥ 98%. The error recognition rate here refers to the proportion of people mistakenly identifying different individuals as the same person, which greatly affects the safety performance of the all-in-one machine. The pass rate refers to the proportion of live bodies correctly recognized by the all-in-one machine, which determines the practical performance of the all-in-one machine. This set of numbers means that when the number of face databases is 100000, the recognition accuracy of the Minivision Technology face control all-in-one machine is ≥ 98%. This high recognition accuracy has made Minivision Technology's face control all-in-one machine one of the top tier in similar image recognition products.
Although the recognition accuracy of the Minivision technology face control all-in-one machine is high, the requirement for face resolution is not high, with a minimum pixel size of only 80 * 80 (pixel). On the premise of a 100000 face atlas, Minivision Technology's facial recognition technology can respond immediately, provide efficient feedback, and track and recognize faces in real-time. In the detection, the response time of the Minivision technology face control all-in-one machine for detecting a single target face is ≤ 1 second. When the image resolution in the portrait database is ≤ 640 * 480 pixels, face pixels are ≥ 80 * 80 pixels, and the number of faces in a single image does not exceed 10, the response speed of face retrieval is>4 images/s. Thanks to lower input pixel requirements, efficient recognition speed, and accurate recognition accuracy, the application scenarios of Minivision technology control integrated machines have become more widespread. For example, during subway security checks, rapid portrait recognition can reduce the waiting time for passengers to queue up, while in providing services for information systems such as security, monitoring, and management of enterprises or government departments, all-in-one machines can promote the implementation of automated management and improve work efficiency.
