Solutions that adapt to the new needs of facial recognition.
by RealNetworks
2019 turned out to be an exciting year for many reasons, such as economic growth, the rise of certain technologies, the boom of Artificial Intelligence (AI), its benefits and possible improvements in society.
However, a storm was looming and we had no idea what was coming and what would hit the world: Covid-19 strangled and confined the entire society to their homes, with a huge impact on various levels, from economic to human to psychological.
As a company, we have worked for a long time to build improved worlds, that is, smarter and more secure thanks to AI and our strategic partners. Specifically, our technologies contribute to creating ideas and providing data so that better decisions can be made, either by security guards who protect us or to feed information and data to systems, platforms or machines, which once improved can help us in certain situations.
When developing facial recognition systems, the challenges to delivering excellence are not small. They come, for example, from the technological complexity to be precise, fast and available in lean infrastructures. Others come from the perspective, in terms of privacy; or simply in a way by third-party agents, they arrive changing the conditions and the context, and they make it move quickly; as hackers with their counterfeiting and fraud techniques or others such as new realities such as Covid-19 and the imposition of the use of masks on others.
Last year we defined our Guiding Principles and argued that AI is essentially good when used correctly, with good judgment, goodwill, justice and to solve society's problems.
With this mindset and to face Covid-19, we carefully analyze the new reality and move forward to continue building improved worlds, making them safer and smarter.
We recently launched a webinar to put the benefits of SAFR, a facial recognition and video analysis solution, in a new context, ranging from precision in facial recognition with occluded faces, which allows:
1. Healthcare professionals move through safe areas without removing protective equipment;
2. Providers of essential services continue to provide products and services without delay; and
3. Security threats do not go unnoticed and actions are taken quickly.
For that we needed to deliver value as accurately as possible. We currently offer 93.5% accuracy when a face is occluded against a clean photo in the database, or 99.87% if you have a regrouped photo of the face with a similar occlusion level.
When we talk about occlusion, we could go to binary optics (occluded/non-occluded) and highlight the benefits of its detection, for example, in certain restricted areas (hospitals, supermarkets, pharmacies and others) in accordance with the requirements of GDPR, in this specific case, which provides a view of whether people's faces are occluded or not, being 100% GDPR compliant, as there is no biometric signature assigned to the face of the person in question.
Photo 2.
In this new context, contactless is the key to securing access, confirming identity and granting access to secure areas without the use of hands. This results in more hygienic procedures, such as queuing at an airport, without the need to exchange documents or passports, for example.
Photo 3.
Or even without the need to place the fingerprint and increase security measures with 2FA processes and the use of AprilTag badges related to biometric signatures of the face.
Photo 4.
When it comes to monitoring and verifying the intensity of the subject in front of the camera, which requires a certain identity, there are several methods, both collaborative and passive. In the case of SAFR, it offers native collaborative models in the software and, in combination with Intel's RealSense 3D camera, can provide the following experience.
Photo 5.
In addition, and in the age of social distancing, it is extremely important to understand the crowds or numbers greater than the number of people allowed in a given space. SAFR can contribute positively by counting people, locating subjects, and triggering alerts as needed.
Photo 6.
When we have a clear view of who the virus is most aggressive against and who has tested positive, this technology can help monitor the age of people in a given location, comprising higher volumes of potential risk groups, or through watch lists of those registered at a hospital or medical center. monitoring them in the city.
The identification of people registered on a watch list can alert monitoring centers, as in the previous case, and at the same time, this technology can be configured to produce alerts and SMS notifications for mobile devices, get faster responses and join the realities of mobility.
Photo 7.
With all that said, we reinforce our principles by taking responsibility and adding value to help society with our technologies. In this way, our teams continue to work and build smart technologies that, together with smart processes, will help now with Covid-19 and in the future.
