The facial recognition technology first saw roots in the 1960s.

Fascinated by it, Varun Chatterji has spent more than eight years learning and grasping various image recognition techniques and while he was at it, he attained a specialisation in Deep Learning-based Computer Vision.

He went on to conceptualise Seventh Sense, an AI tech company with a focus on creating power-efficient and low-cost computer vision systems.

In a response to the battle against the pandemic, they have implemented a derivative of their technology in AI-driven facial recognition fever detection solutions placed in Singapore’s shopping malls and healthcare facilities.

This facial recognition and thermal camera processing technology allows these public places to do away with the need for manual screenings.

Pioneering Mobile Security

The name Varun Chatterji may not ring a bell immediately, but this man has co-founded various startups — of which, at least one, would trigger your memory.

Back in 2003, he met one of the co-founders of his first startup Darius Cheung (he’s the founder of property portal 99.co) during the NUS Overseas Colleges Programme (NoC) in Silicon Valley.

Varun graduated with a degree in Computer Science in 2005 and set up tenCube with Darius and two other co-founders, Rishi and Indradeep, shortly after.

(From left to right) Varun Chatterji, Darius Cheung and other co-founders of tenCube, Indradeep Biswas and Rishi Israni / Image Credit: Knowledge Enterprise

The startup pioneered mobile security and personal information management (contacts, text, emails, phone calls etc) for the then-emerging smartphone market.

“We built “mobile software” before the term “app” became fashionable,” said Varun.

It had counted Singapore Police Force, Defence and Science Technology Agency (DSTA), Centre for Strategic Information Technology (CSIT) among its clients.

Image Credit: NUS

He put Singapore onto the world map when tenCube was acquired by McAfee in 2010. That was also when he joined the McAfee team, but left in 2012 when he became interested in Computer Vision instead.

Hailing from India, the 39-year-old was quick to tell us that he “almost didn’t come to Singapore”. He had initially enrolled in an undergraduate course for Physics at St. Stephen’s College in New Delhi but he “really wanted to study Computer Science”.

On seeing how five of his best friends gained admission into local universities here, he applied to the National University of Singapore and got in. He only joined the NoC programme in his third year.

“So, you could say that if I had not been a part of [NUS] and NoC, my life might be quite different,” he laughed.

The Startup Itch

After tenCube, Varun started exploring Computer Vision to make smart toys for kids and niche hardware such as smart mobile tripods.

These smart gadgets could record a variety of standard shots for videography (smooth pan, smooth zoom etc) to make the life of amateur film makers easier.

Those ideas unfortunately didn’t take off so he went on to start Sent.ly, which was focused on providing customer support over text messaging.

It was during the Sent.ly days when he met Sam, one of his co-founders of Seventh Sense.

Sam was introduced by Ashish (another co-founder at Seventh Sense), whom Varun knew from his childhood days in Kanpur. In later years, Sam became a board member of Sent.ly.

With the emergence of messaging platforms such as WhatsApp and WeChat, the relevance of SMS declined rapidly and so did Sent.ly.

Prior to winding it up in 2017, they tried to pivot to introducing chatbot and chatbot automation for customer support.

The pivot towards chatbots got me interested in Machine Learning and AI. After Sent.ly, I took an online course on Big Data and Social Analytics taught by Alex Pentland from MIT.

– Varun Chatterji, co-founder and Chief Software Architect of Seventh Sense

The course was an interesting mix of economics and machine learning, and inspired him to look up his other works — that’s how he came across Alex Pentland’s works on facial recognition.

“Facial recognition was quite exciting to me, especially given my earlier interest in Computer Vision, and, in the next year, I educated myself in various image recognition techniques and also completed a Deep Learning specialisation on Coursera taught by Andrew Ng,” said Varun.

Varun Chatterji, founder and Chief Software Architect of Seventh Sense

Along with Ashish and Sam, a veteran CFO and ex-board member in Sent.ly, they founded Seventh Sense in 2017.

Ashish had made his first attempt in AI back in college and spent months working on implementing fuzzy logic and artificial neural networks in hardware. He eventually gave up, attributing to “being couple of decades too early”.

I proceeded to develop a prototype system on a small, low power, compute constrained hardware.

This caught the interest of multiple units within Singapore’s Ministry of Defence (Mindef) eco-system, such as DSTA, FSTD, and subsequently, Cap Vista, the strategic investment arm and fully-owned subsidiary of the DSTA, a statutory board under Mindef.

– Varun Chatterji, co-founder and Chief Software Architect of Seventh Sense

Finding His Calling In His Third Startup

With his third company, Seventh Sense, Varun found an intersection of his interest in security, machine learning and Computer Vision.

“Over the next year, we continued to develop our technology. Along the way, we underwent thorough due diligence processes from the multiple parties within Mindef, ST Engineering, SingTel, etc. Subsequently, Cap Vista and 500 Startups came (onboard) as investors,” he said.

Image Credit: Seventh Sense

In the firm’s initial years before funding, Seventh Sense was “completely financially supported by Ashish and Sam,” said Varun. Both had day jobs and essentially provided him with sustenance to work on his ideas.

Fast forward three years later, they have partnered with the largest telecommunications operator in Singapore to provide live deployments for retail and workplace safety during the Covid-19 pandemic.

Their retail products help in measuring and controlling overcrowding, perform temperature screening and have features that help businesses comply with safe work practices.

Unlike conventional technology, the system can also provide a real-time aggregated people-count at entry and exit points. These analytics on building occupancy are helpful to prompt retailers to promote social distancing and/or limit the number of entrants.

Image Credit: Seventh Sense

Their products range from a retail analytics and temperature screening solution to an employee-and-visitor management solution with access control linked to gantries.

Clearly, this product comes at an opportune time given Singapore’s gradual reopening of schools and retail spaces. Moreover, public places like malls are doing away with manual screenings.

The technological solution can also be positioned at crowded locations like healthcare facilities and educational institutes and is linked to the SafeEntry system.

Explaining To The Layman

Varun revealed that the biggest challenge in setting up Seventh Sense was “communicating the hard technical details”.

With the implementation of face recognition in smartphones, they often get asked that if the technology is becoming commoditised.

Most people don’t realise that a 1:1 match in the case of a phone is very different from a 1:N match.

Furthermore, people often think of the technology as matching a frontal (mugshot) picture against another frontal (mugshot) picture. The reality is quite different for ‘in the wild matching’.

– Varun Chatterji, co-founder and Chief Software Architect of Seventh Sense

Apparently, it took them more than a year to educate investors and customers on the realities of facial recognition.

Dashboard / Image Credit: Seventh Sense

“We also get asked how we can compete with graphics processing units (GPUs) with lower-powered devices. The answer is not always as simple as “we do X”. We do a variety of things, that, over the years, have helped us get where we are,” continued Varun.

Another challenge they have been experiencing is finding talent in machine learning.

“Getting to a team size of 12 has been quite a challenge,” he said.

Raising Funding For Series A Round

To date, Seventh Sense has raised a seven-figure seed round with Cap Vista and 500 Startups, along with two local angel investments.

They are currently raising a Series A round and have “commitments from all their existing investors, with a handful of new investors underway”, which according to Varun, is “interesting and challenging” in a post-Covid-19 world.

Image Credit: Seventh Sense

At this stage, they have a pipeline of several large contracts and product deployment engagements and he sees the need to expand the team to deliver on them.

“Our next goal is to build a bigger research team and tackle some problems that require hard-core research,” said Varun.

“With the expansion of our team, we will be able to realise some of our harder goals such as developing and building products around our hard-earned IPs — US patent (US16/431,818) and a few more in the pipeline.”

So, How “Deep” Is Deep Tech?

With more enterprises adopting solutions that embody AI, machine learning or blockchain, deep tech is becoming increasingly common in corporations.

To further accelerate deep tech commercialisation, the Singapore government has committed S$19 billion to the development of deep tech startups.

But does that mean that deep tech is becoming commercialised?

Far from that, said Varun.

It can take as long as four to five years for deep tech companies to get off the ground. Afterall, it’s not called “deep tech” for nothing.

Very simply, deep technology may be defined as a scientific breakthrough  that can potentially change human life for the better.

Sectors wherein deep tech has made giant strides include Biotech, Life Sciences, Cleantech and Computer Science. The newest medicines and medical devices, as well as 3D printing are all examples of deep tech.

Only through extensive testing can the full impact of each deep tech development be determined.

This is how deep tech businesses distinguish themselves from companies such as Facebook or Airbnb.

For budding entrepreneurs who are looking to stick their hands into deep tech, Varun has some advice:

Don’t jump on a clever idea. Validate the idea before thinking of making it a business. Try to find backers for your idea. Then, and only then, should you startup.

Ideas are cheap. Execution is not.

– Varun Chatterji, co-founder and Chief Software Architect of Seventh Sense

Featured Image Credit: Seventh Sense

The post This Startup Utilised Its Tech In Facial Recognition Fever Detection Solutions We See At Malls appeared first on Vulcan Post.