Facing the facts

Posted Wednesday, September 23, 2020


Facing

We ask - how effective can facial recognition technology be with the new enforcement of face masks, how might this technology fare in the future of the property industry, and how useful can it be in this pandemic?

Imagine a world where your face is your door key.

You may already be in it.

Facial recognition technology has been around for some time, used in border checks and police body cameras, before moving to temperature detection for Covid-19. Contactless tech is increasing as the property industry looks for more hygienic solutions to common issues such as door entries, bin stores, etc, but how might it be used in our everyday lives, and when it comes to facial recognition, will it be effective when wearing a mask?

LettingAgent Today say the UK might be next to implement facial recognition technology in the property sector. Currently, some NYC rental apartment complexes are installing such recognition systems to enable tenants to let themselves in, replacing keys and electronic fobs. Reports show that tenants may have mixed emotions about this, and the potential to implement this in the UK brings with it some controversy.

As the country tightens its grip once again on public restrictions, we may have to rethink and embrace our use of biometric technologies. But before we make a call on whether this is a good idea, let’s look at the research:

A recent study from the National Institute of Standards and Technology (NIST) has revealed that errors in algorithms occur when using images of masked individuals, rather than unmasked. Testing images of differently shaped and coloured masks, tests were run on 89 algorithms for the Face Recognition Vendor Test (FRVT). They found that:

  • The accuracy of algorithms declined with masked faces
  • Masked faces caused more frequent “failure to enroll or template” errors
  • Greater nose coverage lowers accuracy
  • Occlusion with masks does not affect the matching of photos between two people
  • Rounded masks caused fewer errors than non-rounded
  • Colour could not be completely tested

Whilst only testing for one-to-one recognition (images of the same individual), the study found overall that individuals wearing a facial mask could spike errors in facial recognition. Sean Moore, CEO of True Face says, “if the [facial recognition] companies aren’t looking at this, aren’t taking it seriously, I don’t foresee them being around much longer”.

CNN Business state that the future of facial recognition technology rests upon the area around the eyes, and companies are now attempting to more efficiently utilize this area of the face. The race is on to adapt to the trials presented by coronavirus.

Tech5, a Swiss recognition company, has faced challenges prior to Covid-19 relating to religious face coverings. However, the technology still performed worse with a mask. They say that their algorithms were designed to identify someone wearing sunglasses, or with facial hair – not masks.

The landlord of Atlantic Plaza Towers (NYT) was recently informed that their key fob system was “being scrapped for facial recognition technology as a means of access”. As the algorithms have higher error rates in cases of nonwhite and female faces, many say they are racist, ineffective and, insulting to civil liberties. But developers say it may reduce attacks on individuals using their building late at night, helping to prevent terrorism and minimising criminal offenses.

At Kings Cross Central, developer ‘Argent’ faced criticism for implementing the technology. They are currently under investigation by the UK Information Commission, as the lines between privately owned and public spaces have been blurred.

But is it safe?

Whilst property owners are making efforts to gather ever more data on the people interacting with their buildings, there must be a “balance between national security and individual liberty”. Whilst this technology can make buildings safer, it risks the public feeling constantly watched.

One landlord says “imagine you could put on a watch and be notified when a terrorist walks onto your estate” – this thought-provoking statement may be behind the push for this technology’s implementation, but remember to keep people informed and measure its use. If public awareness/contentedness falls below a certain level, it is an indicator that something is not right.

Stay tuned for further information on this cutting-edge topic, whilst we await its adaptation to COVID-19 demands.

 

References

Kirk, D., Huff, J., Galibois, M. and Barbuto, V., 2020. [online] Available at: <https://www.reedsmith.com/en/perspectives/2020/04/facial-recognition-and-biometrics-in-a-covid19-real-estate-world> [Accessed 23 September 2020].

NIST. 2020. Face Recognition Vendor Test (FRVT) Ongoing. [online] Available at: <https://www.nist.gov/programs-projects/face-recognition-vendor-test-frvt-ongoing> [Accessed 23 September 2020].

Norwood, G., 2020. Facial Recognition Used By Property Management Firm In US - The UK Next?. [online] Letting Agent Today. Available at: <https://www.lettingagenttoday.co.uk/breaking-news/2019/5/facial-recognition-used-by-property-management-firm-in-us--the-uk-next?source=related_articles> [Accessed 23 September 2020].

Phillips, M., 2020. Facial Recognition Tech In Real Estate Pits Safety Against Liberty. [online] Bisnow. Available at: <https://www.bisnow.com/london/news/technology/facial-recognition-tech-in-real-estate-pits-safety-against-liberty-100773> [Accessed 23 September 2020].

Rachel Metz, C., 2020. Face Masks Are Screwing Up Facial Recognition Software. [online] CNN. Available at: <https://edition.cnn.com/2020/07/28/tech/face-masks-facial-recognition/index.html> [Accessed 23 September 2020].

Vincent, J., 2020. Face Masks Are Breaking Facial Recognition Algorithms, Says New Government Study. [online] The Verge. Available at: <https://www.theverge.com/2020/7/28/21344751/facial-recognition-face-masks-accuracy-nist-study> [Accessed 23 September 2020].


Back to news headlines