Facial recognition for remote ID verification wins Africa Prize

Charlette N’Guessan is the first woman to win the prestigious prize since it was established in 2014, and the first winner from Ghana. The prize was awarded for BACE API, which uses facial recognition to verify identities remotely.

N’Guessan and her team decided to go ahead with the project in 2018, after they discovered that Ghanaian banks have a significant problem with identity fraud and cybercrime. Their research estimated that approximately $400m is spent every year by Ghanaian financial institutions to identify their customers.

“Online identity fraud is a huge problem, especially in Africa, because the growth of online businesses and the enthusiastic embracing of new technologies on the continent have not been matched by an equivalent commitment to cybersecurity,” N’Guessan told E&T. “As software engineers and data scientists, we wanted to bring our contribution by building a solution that can help people to feel safe while using online services across Africa. Since day one, the goal has always been the same: fight against online identity fraud.”

BACE API is aimed at financial institutions and other industries which rely on identify verification when offering services. In partnership with a data controller which deals with government-issued ID, it has access to Ghanaian passports and other documentation which it uses during its verification process.

The software uses a phone or computer’s built-in camera and requires no specialised hardware. BACE API uses live images or short videos taken on cameras to detect whether the image is of a real person or a photo of an existing image.

It is already being used by two financial institutions to verify customers’ identities, and being tested on an event platform to confirm registration of attendees. During the coronavirus pandemic, it has emerged as a viable alternative to in-person identify verification processes used by most businesses, allowing them to authenticate new customers without ever meeting face-to-face.

In contract to commercial AI systems, it has been developed specifically to identify Africans. Studies have demonstrated that commercial facial recognition software tends to perform comparatively poorly when analysing the faces of women and people with dark skin tones, in large part due to poor representation in training datasets. D’Guessan explained that this has caused “misunderstandings and fears” about the technology in Africa, so few local businesses have invested in the technology. In deciding to develop their own facial recognition, they took a risk that it would not be accepted.

“BACE is an effective solution built by a Pan-African software team born and raised in Africa,” she said. “Our solution has been developed and trained in the local context through different use cases to ensure reliable, smooth, efficient protection of data and to build trust between local businesses and people.”

At a virtual awards ceremony, four finalists for the Africa Prize delivered presentations, and then judges and a live audience voted for the most promising engineering innovation.

N’Guessan earned the first prize of £25,000. The three runners-up, who each received £10,000, are: Farmz2U, a digital platform providing tailored agricultural data to farmers; PapsAI, a low-cost microscope which speeds up cervical cancer screening diagnosis; and Remot, a system which manages off-grid power grids by monitoring the condition of solar arrays.

The fifteen shortlisted Africa Prize entrepreneurs received eight months of training and mentoring, during which they developed their business plans and were connected to individuals and networks in the UK and across Africa who could help accelerate their business. N’Guessan and her team used this as an opportunity to focus more on business development, as well as improving the accuracy of the model and reducing the verification time.

“Being part of the Africa Prize has given us such confidence,” said N’Guessan. “We focus on Africa because we want to make sure BACE API is used by our people, and works for them. We are so grateful to the Academy, and cannot wait to take our innovation to new heights.”

Rebecca Enonchong, Africa Prize judge, commented: “We are very proud to have Charlette N’Guessan and her team win this award. It is essential to have technologies like facial recognition based on African communities, and we are confident their innovative technology will have far-reaching benefits for the continent.”

James Duddridge, UK Minister for Africa, added: “Congratulations to all the participants in this year’s Africa Prize. The UK is a hub of engineering innovation, and home to a wealth of entrepreneurial talent and experience. By partnering this talent with the most promising African innovators, we can create local solutions to global challenges, transforming lives and improving economies.”