Facebook Is Considering Facial Recognition for Its Upcoming Smart Glasses

Apr 28, 2021
4 min read

Introduction

Facebook is currently checking the legal implications for using facial recognition technology in a pair of smart glasses. The company is planning to launch the product by the end of this year. Facebook’s AR and VR Hardware Chief, Andrew Bosworth, told employees that the company is evaluating the legal framework to integrate facial recognition technology into such devices.

Bosworth tweeted, “we’ve been open about our efforts to build augmented reality glasses and are still in the early stages.”

The development of smart glasses is yet to begin. As it is illegal to search people based on their facial features, Facebook is currently assessing and planning the use of facial recognition technology in the pair of glasses.

In response to a question asking whether users would be able to mark their faces as unsearchable, Bosworth responded, “we don’t know where to balance those things”. He added that there would be a public discussion about the benefits of facial recognition in upcoming smart glasses and the risks involved.

What is Facial Recognition

Facial recognition technology is defined as identifying and verifying a person's identity by capturing, analyzing, and comparing facial details.  Facial recognition technology can detect and locate human faces in videos and images. The face of a person is captured and transformed into digital information based on facial features. An algorithm is used to convert analog information (facial features) to digital data.  The digital data helps in comparing two faces.

Facial recognition is one of the most natural processes among other biometric measurements. The recognition and authentication of a person is done by looking at the facial details such as the bridge of the nose, spacing of the eyes, the contour of the ears, lips, chin, etc. whereas in other biometric methods, it is done by scanning iris, recognizing voice or fingerprints.

Evolution in Facial Recognition Technology

Facial recognition technology is gaining huge recognition since 2013. As per NIST report:

· The facial recognition algorithms developed during 2018 offer great accuracy.

· The best facial recognition algorithm of 2020 has shown only 0.08% error rate.

Top Spot Projects in Facial Recognition and upcoming markets

Top software companies regularly launch new products based on  technologies like image recognition, face recognition, and artificial intelligence. Some of them are:

Academia

In 2014, the GaussianFace algorithm was developed that gained an excellent rating. It was the first algorithm developed for face recognition. It scored 98.52% in facial recognition as compared to 97.53% by humans.

Google

FaceNet by Google gained a new record of 99.63% accuracy. This technology of face recognition is used in Google Photos for automatically sorting and tagging. It was also incorporated in OpenFace, an unofficial open-source version.

Microsoft

In 2018, Microsoft shared an improved biased facial recognition technology.

Amazon

In 2018, Amazon promoted Rekognition, a cloud-based face recognition service. It was capable of recognizing around 100 people in a single image. But it went wrong when it falsely recognized 28 US Congress members as criminals.

Facebook

In 2014, DeepFace by Facebook had an accuracy rate of 97.25% as compared to 97.53% by humans. It was developed for matching two pictures of a person.

Facebook deployed DeepFace in 2015. It has the largest repository of photos uploaded by users. DeepFace, based on facial recognition technology, was used for photo tagging, identifying, and suggesting people in pictures, videos, etc. Facebook hasn’t marketed DeepFace and other similar products so far. Facebook has confined facial recognition technology to its own social networking usages.

The other top facial recognition vendors are Accenture, BioID, Aware, Certibio, Fulcrum Biometrics, Fujitsu, Thales, Idemia, HYPR, Leidos, M2SYS, Nuance, NEC, Smilepass, and Phonexia.

Even though facial recognition technology is facing scrutiny and hated debates for security reasons by police departments, federal agencies, etc., Facebook is considering this technology for its new product. The upcoming Smart Glasses can recognize someone at a meeting or party even if the person does not recall their name or face. It will be a challenge for Facebook as some cities like Boston and San Francisco have already banned facial recognition technology from government use.

According to Maxine Williams, Facebook Chief Diversity officer, there might be a need to develop facial recognition principles to take precedence in the absence of technology governing laws. She said, “Just because you can build something doesn’t mean you will”. Facebook would need to check the potential for discrimination and harm.

Bosworth remarked that he had discussed facial recognition technology with Facebook CEO and reviewed the audience's privacy concerns for upcoming smart glasses. As per the Big Technology Newsletter report, the Facebook vice-president had mentioned in an internal memo about the necessity of product differentiation on a privacy basis.

Bosworth believed that Facebook should become the undisputed leader in providing privacy-aware software. He is aware of privacy concerns related to facial recognition technology but he is not in the support of current legislation such as the Biometric Information Privacy Act (BIPA) of Illinois.

Legal policies don’t allow face recognition. As it was seen last year when Facebook had to pay 650$ million to Illinois citizens for using the photo-tagging feature. It was a violation of BIPA. Bosworth said, " That’s ok. We can do that as a society and this product will survive and thrive without it. I do think there are some lost opportunities, though.” BIPA was passed back in 2008 and it prevents private companies from collecting and storing data without the consent of the person.  Bosworth questioned the feasibility of age old legislation for today’s technologies.

Conclusion

Facebook is getting ready to release its smart glasses product capable of facial recognition in partnership with Luxottica (Ray-Ban makers) sooner or later in 2021. Although, Bosworth said, “If people don’t want this technology, we don’t have to supply it” in an interview. The product is expected to provide an augmented reality experience and challenge the products by Snapchat and Amazon.

It is estimated that by 2024, there will be a boom in the global facial recognition market. It would generate $7 billion of revenue over 2019-2024. Asia-Pacific region is experiencing the fastest growth in facial recognition technology and India is leading in this field.