Facial Recognition - Are we ready for that?

-

While recognizing faces and emotions is simple for humans, it is quite a challenge for computer systems. In order to transfer the same skills to machines, the computing power of computers and their ability to learn had to increase.
The very principle of this operation is quite simple, the human face consists of a set of features such as the distance between the eyes, the shape and length of the nose, the contour of the lips, the shape of the ears or the chin. By imposing an appropriate grid of points on our face, the machine can create a virtual image of the face and compare it with the model that is stored in the database. In addition, to make such a system more accurate, i.e. to cope with changing emotions, different lighting, setting or aging, additional algorithms supported by artificial intelligence (neural networks) were used, which by minimizing errors increase its effectiveness.

Source: https://www.analyticsinsight.net/the-impressive-growth-of-facial-recognition-technology/

 

Cameras are ubiquitous these days, they are placed in the streets, parking lots, in shops, in our smartphones and laptops. No wonder that along with the development of this technology, its use in our everyday life grows rapidly.

Source: https://www.crn.pl/artykuly/rynek/monitoring-potrzebuje-infrastruktury

 

 

Application areas

Main advantages of Facial Recognition are:

  • high versatility,
  • speed of action (real time),
  • touchless character

Below we will take a look into two areas of human activity which can be or are already in use today:

  • Security and monitoring
  • Health and medicine

They also can be adapted to many different functions or activities and certainly, will also find a place in other market sectors as well.

 

Security and monitoring

There are other biometric measurement methods such as a retinal scan or a fingerprint, which may be more accurate in verification than Facial Recognition, but they are not fast enough and require the involvement (cooperation) of the person we want to identify. Meanwhile, the scan in the Facial Recognition technology takes literally a second, and additionally, it can be performed even when the person whose identity we verify is not fully aware of it.

The Facial Recognition technology has many positive aspects as long as we use it with right intentions. It will help us to find in the crowd a wanted person, identify a dangerous criminal or a hooligan with a stadium ban. In factories, power plants, large industrial plants or military facilities, it will allow us to control access to specific rooms only for employees with appropriate authorization. We will identify the driver and vehicle registration plates. Based on this data, we will be able to verify whether the driver has a valid driving license, whether the vehicle has a valid technical inspection and a valid liability insurance policy. It is also possible (using additional devices), to check the speed and weight of a given vehicle and thus to select cars that are overloaded or exceeded the speed limit. Such systems can significantly minimize the risks on highways and expressways, increase safety in such sensitive places as border crossings, airports, railway stations or during mass events.

We have to be aware of that not all institutions that use Facial Recognition have the right intentions. In wrong hands, this technology can become a seed of public invigilation and violate human rights. Therefore, it is worth observing what is happening in this area in other countries.

 

Health and medicine

A detailed scan of our silhouette can provide artificial intelligence with information, that will be useful to monitor our health. From our face, it is possible to detect emotions that accompanies use at a the given moment, and determine the present state of our fettle. Going even further, to identify past and predict future diseases.

Facial Recognition can help us:

  • detect depressive states
  • detect various types of genetic diseases, the symptoms of which are changes in facial features
  • track the use of medications that the patient should take
  • support the elderly, people with disabilities or speech impediments
  • support the treatment of pain and monitor its intensity

In the current corona virus (COVID-19) pandemic situation, cameras with thermal imaging are used to automatically measure body temperature. By identifying people with fever, companies and institutions are able to detect and prevent infection of other people earlier by cutting the chain of virus transmission.

Source: https://www.securitas.ie/

 

A quick look into selected available solutions

 

Amazon Rekognition

Deep learning-based visual analysis cloud service. Provides highly accurate facial analysis, face comparison, and face search capabilities. You can detect, analyze, and compare faces for a wide variety of use cases, including user verification, cataloging, people counting, and public safety.

Source: https://azure.microsoft.com/pl-pl/services/cognitive-services/face/#demo

 

 

Azure Cognitive Services

Cloud based service to detect, identify, and analyze faces in images and videos. U can use this technology to support various scenarios, for example, authenticate people for access, count people in a space for crowd control, or garner crowd insights for media campaigns. You can detect one or more human faces along with attributes such as: age, emotion, gender, pose, smile, and facial hair.

Source: https://azure.microsoft.com/pl-pl/services/cognitive-services/face/#demo

 

Source: http://www.proto.pl/aktualnosci/emocje-w-reklamie-im-wiecej-tym-lepiej

 

 

Face recognition

Using face recognition, you can for example recognize a large set of known people, and make a prediction for an unknown person in short computation time. The classifier is first trained on a set of known labeled faces and can then predict the person in an unknown image by finding most similar faces in its training set and return a most likely label. You can use this library to do real-time face recognition as well.

Traning set

Source: https://www.imdb.com/

 

Test faces

Source: https://www.imdb.com/

 

Results:

- Found Lagertha at (35, 118)
- Found Ivar at (291, 93)
- Found Rollo at (593, 98)
- Found Ragnar at (902, 118)
- Found Aslaug at (1191, 118)
- Found Biorn at (56, 552)
- Found Floki at (476, 563)
- Found ??? at (784, 546)
- Found ??? at (1166, 510)

 

Conclusion

As long as the proper use of Facial Recognition can bring many benefits, keep in mind that this technology, in addition to the light one, also has a dark side. Let us try (as far as possible) to verify for what purposes our effigy is used for and do not mindlessly feed data in the form of our effigy, into applications of unknown origin. It may turn out that in the near future, our anonymity to-date will be one of the few things for which we will really miss.

 

References

  • https://azure.microsoft.com/
  • https://aws.amazon.com/
  • https://github.com/ageitgey/face_recognition/

 

 

AMB'S TECH INSIGHTS - AN ARTICLE BY KRZYSZTOF

Krzysztof has over 20 years of IT experience, starting as a software developer, Krzysztof soon become a Software Architect.

Today as a Professional Consultant, Krzysztof supports our clients with business intelligence solutions and product development consulting. Krzysztof is very into AI topics and certified in the Microsoft Azure Cloud.

 

 

DO YOU NEED HELP WITH WEB DEVELOPMENT? CONTACT US AT

CONSULTING@AMBSOFT.DE