Key considerations for adopting facial recognition technology

Some items must be taken into account when choosing the hardware, frameworks, services or tools with which your software will be developed. It is important to clarify that this is a personal proposal, you can find different opinions which must be evaluated according to the scenario, client, development time, client investment costs, capabilities, hardware etc. I leave the post open for comments on it in case you have any send it to me, or write to tilcaivana@gmail.com your option will always be welcome.

In order for the system to function, it’s necessary to implement three steps.

The first thing to consider when it comes to face detection is the hardware.

In the experience that I have acquired working with cameras, it is key to consider the quality of the image that will be used for analysis. The quality not only depends on the device they use (RGB camera, camera with depth, camera with infrared or night visibility, camera of mobile device such as cell phone or tablet, etc.), but also depends on the environment in which the system is used (affected by weather, affected by natural or artificial lighting, affected by external signaling, etc). Some of the tips for the choice of hardware for Face Recognition (important these clarifications may vary for object detection always depend on the purpose of the software).

• Web cameras have face detection problems with low lighting vs a higher quality camera (more than 1080px resolution).
• The depth cameras enable us to see the shape of the face, it is not 100% perfect but eliminates security problems (which we will evaluate later).
• Infrared or night cameras do not lose face detection, as it ignores external lighting. Another benefit of such a camera is that it cannot see screen or device projections.
• Test the cameras with natural light and artificial light.

Will talk about this in the following post.

Stay Tuned!.