Why did I start the facial recognition startup Facedapter?
How does your past connect with what you do today?
n 1999, Praveen Gajjala, the Founder of Facedapter started his first job as a salesman for a sales agency that sold “Airtel Magic” prepaid and post-paid sim cards back in India. In 1999, I sold the sim cards to almost everyone. In 1999, there were only two or three ways to verify people using the driving license, pan card (Income tax ID), or Election card/Voter ID. Soon, Reliance Telecommunications launched a postpaid subscription and a mobile device for 1 INR (1/70th of US dollar), the company vision was to connect everyone at their fingertips. At the time of launch, tens of millions of subscriptions and mobiles were sold and at this point, we have witnessed millions of identity frauds, many fraudsters used synthetic and fake IDs and forged utility bills for address proof. At this time, Reliance launched a verification campaign with tens of thousands of Verification officers and I was one of them, who had to visit the customer’s house to manually verify the subscriber’s ID and the copy I had in my file (dossier in French).
When our founder was 18yrs old he was applied for a credit card impersonating his father, he succeeded in receiving a credit card with a limit of $ 2000. If a teenager can be impersonate using the Identity of someone else, imagine what an experienced Identity thief can do. Let’s leave this to your imagination.
Between 2005–2009, Praveen witnessed numerous identity and credit card frauds while working at an Insurance company and Retail store. In recent years, Praveen landed in a startup developing 3D facial recognition and his role as business development director was to develop new markets and partnerships with FinServ and integrators across EMEA and Americas. For more than 3 years, Praveen interacted with key stakeholders from various industries, and in 2018, he was amazed that identifying people is still a huge problem despite there are tens of companies that offer facial recognition technologies, for instance, black females cannot be identified or in some cases, they are not detected. He became restless and his quest for solving big problems finally started in March 2019.
What is the Problem?
Today, we have many types of facial images such as 3D/Near-infrared/Thermal captured using various types of cameras or sensing technologies. But, most governments and businesses have an existing database based on a passport or an ID card that contains a 2D image of a person. Matching any of the above image modalities to a 2D image in an existing one is simply not possible if there is no pre-enrolled database that contains 3D/NIR/Thermal images. A lot of companies across the industries talk about building digital trust, but how do you enable trust for any person from any ethnicity and all genders. Trust is often two-way and not just one-way so that businesses and their customers can experience it in the same manner.
- Enrolling a database of the latest facial images
- Bias in recognizing colored people and gender
- Doesn’t work in all weather conditions
- Re-identification is a challenge as people get older
What is the Solution?
A single API that can recognize and match 3D, Near-infrared, Thermal images to a 2D image on passport, ID, or an existing database. For example, our software API will enable the airport`s border control and security to not only recognize and match all types of images but also unifies the identity system making and fully automate access control such as the entrance of the building, check-in, printing a boarding pass, baggage drop, VIP lounge entry, embarkment, transfers, etc. In general, immigration officers manually perform the identity verification using a passport or an ID (2d image) and real-time of the passenger (Mug shot / 2D image).
Our customers believe that our technology will help them bridge their innovation gaps. We have received LOIs from three customers (USA, Switzerland, and Germany). A special thanks to them, without our customer’s belief and validation we wouldn’t be here today. The special mention goes to all the Innosuisse mentors and coaches who have supported us in building a strong CTI project that won the trust of jury members and the grant from the Innovation Agency of Switzerland that helped our founder to kick start revolutionary AI project like Facedapter with Idiap. Our first, PoC with a customer will commence in the mid of 2021. We got the attention of organizations and companies such as the World Economic Forum, GlobaliD, and SITA.
Impact and Diversity
Due to the Pandemic outbreak, many organization around the globe digitalized their services and applications. In 2021, reports show that biometric frauds tripled both online and offline causing tens of billions of dollars in losses. People are increasingly connecting through Zoom, sharing, fintech, and mobility apps. On the other hand, the travel industry is affected by the pandemic in recent times which makes it difficult for aviation organizations to bring back the industry to normal, we believe that the travel industry will drastically change over the next coming years. IATA expects 7.2 billion passengers to travel in 2035, a near doubling of the 3.8 billion air travelers in 2016.
Whether it is remote employees or governments or travel passengers, our goal is “To enable machines with a better sense of vision through AI to rebuild trust and make the world a safer place for everyone.” At Facedapter, we are looking forward to contributing to the recovery of the aviation industry.
Our team at Facedapter is diverse and exceptional people from 6 nationalities (Switzerland, France, Estonia, India, Ghana, and Egypt). Our founder supports remote-first thinking that can save a lot of time for the company and employees which gives more flexibility to work from wherever they prefer.
On a separate note, I definitely, recommend the readers to watch Coded bias on Netflix to make sense of this article. Coded Bias was directed and produced by Shalini Kantayya and cast Joy Buolamwini and Joy is the founder of The Algorithmic Justice League (http://ajlunited.org/).