Face detection and recognition: 2 tables of camera installation parameters
Training of neural networks, expansion of the capabilities of smart analytics, and ways of transmitting and storing information are promising directions of evolution of modern technologies. When installing a video surveillance system, do you need detection and facial recognition, where and how to properly place the video camera? We are sorting this out in the article.
Recommended locations for installation
Cameras with face detection should be installed centrally above entrances and exits, at passageways, in areas with turnstiles, in corridors. The camera should be installed so that it “looks” at the faces of passers-by.
In such places, people either slow down, which allows the camera to “see” the face or stay in the camera’s field of view long enough for detection.
Face detection requires high-quality light and locating the face in the frame for about 3 seconds.
Table of parameters
Let’s outline the main parameters that should be taken into account when installing a surveillance camera:
| Factor | Recommendation | Admissible error |
| Height, m. | 2,5 | +/- 0,5 |
| Distance, m. | 6 | from 4 to 20 |
| Tilt angle, deg. | not more than 15 | from 10 |
In addition to the factors mentioned above, you need high-quality counter lighting of the faces on the video. Possible darkening of faces due to natural light from glazed areas of the room should also be considered.
Table of the recommended height
Next, let’s consider recommendations for the height of the location of the video camera, taking into account the possible distances between the camera and the object of detection.
| Distance, m | Height, m |
| from 4 to 20 | 2,00 |
| from 5 to 20 | 2,25 |
| from 6 to 20 | 2,50 |
| from 7 to 20 | 2,75 |
| from 8 to 20 | 3,00 |
An alternative option is a camera with a varifocal lens. It will allow you to zoom in on the image by changing the focus distance of its lens. Read all about choosing a lens for an outdoor camera in an expert thematic article.
The high resolution of the video equipment will also help to recognize a face at a great distance.
What is necessary to know before buying software?
Before you make a choice and buy a video surveillance system, you need to scrupulously analyze the functionality offered, the possibility of long-term operation, reviews of the manufacturer, etc.
Let’s identify several important nuances that should be considered before buying video surveillance software with facial recognition function.
- The speed of face recognition rate of a person from the white/black list in online mode. What is the proportion of false positives?
- Operation of the surveillance camera in complex and non-standard conditions. It is reasonable to test-drive several video cameras.
- Integration of the face detection function in an already installed video surveillance system.
- Possibility to use the facial recognition function when monitoring several large premises and controlling the system in a single location.
- Evaluation of functional work with recordings (deletion, method, and file storage period).
- Recognition of faces of different sex, age, race, and skin color.
- The cost. How much does the software, additional equipment, maintenance, upgrades, data storage, and other additional bonuses cost?
Faceter Face Recognition
Faceter’s online facial recognition feature offers a host of benefits and opportunities for its users. Much of it has to do with keeping analytics for the business owner: the number of unique and repeat visitors, their classification by gender and age. Black and white lists are also maintained with notification of the appearance of a person from the established list.
The innovative features of the Faceter cloud surveillance service allow you to mount a surveillance system using any two smartphones, even older ones. The entire video surveillance system can be quite budget-friendly and even free of charge. Access to a wide range of functionality is retained in full.
Conclusion
Smart video surveillance is becoming more accessible to a wide range of users. It is already possible to do without the purchase of additional equipment. Facial recognition is currently one of the most popular features, although it has appeared relatively recently.
Intelligent algorithms are intelligently applied in Faceter cloud video monitoring. Their effectiveness has been confirmed by the national research University of Washington.
