Menu

Media makeup students hack facial recognition software

24 October 2019

City of Oxford College students worked alongside makeup artists from Charlotte Tilbury last night in an attempt to manipulate face detection technology.

As part of the Oxford IF Science and Ideas Festival, students created two different theatrical makeup looks.

Two students sat as models for the interactive Making Faces event held at John Lewis’s Place to Eat in the Westgate shopping centre.

Practise makes perfect

Amy Johnson, a media makeup artistry student at City of Oxford College said: “We’ve been practicing for a few weeks now and I am really pleased that we managed to trick the facial recognition software with our makeup looks.

“Being involved in events like this are so important because it shows people the kind of work we do and how different is it to traditional makeup.

“When I started my course, I knew the basics of makeup but have since learned about special effects, prosthetics and face and body paint. These looks we created are not your everyday looks, that’s for sure.”

Working alongside the University of Oxford

The City of Oxford College students successfully harnessed the power of makeup to trick technology with dramatic makeup techniques which focused on the eyes.

Niki Trigoni, a professor of computing science at University of Oxford, joined the event for a discussion on multimodal technology, explaining how it can be used to recognise the whole person.

Niki Trigoni, said: “I really enjoyed tonight's event. It was fun to see makeup artists in action changing facial features of their models and exposing the vulnerabilities of face recognition systems. It highlights the need for these systems to continually improve over time and learn feature representations that are robust to variations in people's appearance due to pose, light, make up, and so on.

“We discussed how leveraging data from diverse sensors can be a practical and effective way of boosting person identification accuracy in challenging real-world conditions.

“Sensor fusion and cross modality training is one of the active areas of research at the Oxford Cyber Physical Systems group. It is not only applicable to improving face recognition systems; it can impact a wide range of context inference problems, from indoor localisation to scene understanding.”

On completion of her presentation, Niki hosted a question and answer session alongside Dane Comerford, the festival’s director.

Niki added: “It was great to see a diverse audience interested in this topic from junior and senior school pupils, to students and professionals. A great meeting where art both informs and challenges science in a way that engages an audience of all ages.”

The challenges the students faced

The models from City of Oxford College and Charlotte Tilbury used the iPhone X Face ID app upon themselves to conclude the evening.  

This application uses light and sensors, taking several images of the users face when set up in order to recognise it when using the phone.

Dane Comerford said: “Advances in machine learning, often described as artificial intelligence, are becoming an increasing feature of our daily lives. In parallel, social media has captured and even encouraged extreme makeovers, promoting a new type of celebrity, the social media influencer, who dedicates an enormous amount of effort and skill to creating adventurous looks using makeup.

“The event brought these two topics together to explore whether we could disguise a face using makeup techniques that a smart phone had learned to recognise.”

If you would like to know more about media makeup artistry programmes at Activate Learning , please visit activatelearning.ac.uk/mediamakeup or contact on 0800 612 6008.