Tapway brings facial recognition solution to brick-and-mortar businesses
IN THE brick-and-mortar world, retail chains and shopping malls have long relied on counting the number of occupied parking lots to estimate the number of visitors to their establishment. However, this highly manual method tends to be time-consuming and inaccurate, to say the least.
One Malaysian startup, Tapway Sdn Bhd, is looking to turn this old method on its head with the use of modern technology like the Internet of Things (IoT) sensors paired with Artificial Intelligence (AI) technology. Recently the company launched a new facial recognition solution called FaceTrack that is said to detect faces and process images to provide insights.
Started in 2014, Tapway was founded by chief executive officer Lim Chee How and chief operating officer Justin Shye Kang Loh.
Lim explained that Tapway’s solutions were made to improve customer experience, streamline operations and enhance security during a recent interview with Digital News Asia at the company’s office in Petaling Jaya.
“Our core is in people tracking and we do this for our clients, most of whom are in the retail sector. We are able to do people tracking using special purpose-built cameras that can accurately track the number of people entering and exiting a premises. We are also able to tell how long they spend in a particular area and track their location, providing a heat map,” he said.
There is even a feature that combines with wearable Bluetooth devices that help exclude staff from being counted, so as to provide higher accuracy.
He cited an example where FaceTrack could be used to improve a customer’s retail experience by recognising loyalty card members and alerting the staff to give special attention to the customer.
Conversely, a “blacklist” of faces that could include suspected shoplifters can be created for retailers and malls so as to alert personnel, via SMS, email or app notification, that the individual is on the premises.
Tapway deployed its FaceTrack solution at the Pavilion Kuala Lumpur last month and it is currently being tested. A total of over 200 sensors have been installed around Pavilion and Pavilion Elite, gathering the flow of people coming into the mall and providing valuable insight to the mall’s management.
The company plans to conduct more pilot tests in the coming months. It added that FaceTrack is currently only available in Malaysia.
But is it expensive to install such a high tech solution? Not all according to Lim as Tapway aims to make facial recognition more mainstream and affordable. “Previously the main barrier to traditional facial recognition technology was the expensive cameras and servers but it is now much more economical,” he said.
Tapway described the solution as primarily running off on an “AI box” that sits within the client’s premises. Essentially, a computer equipped with a Graphics Processor Unit (GPU) that is powerful enough to run their software and link the multiple cameras spread throughout the site.
Tapway has also partnered with Amazon Web Services (AWS) to use Amazon’s Rekognition software together with Tapway’s own AI software to analyse images that have been detected. The process is said to determine the age, gender and emotions on faces that are identified.
All faces detected go through an identification process at AWS to receive their own face ID that is encrypted so as to protect the privacy of people, Lim explained.
The company even has plans to run the service off a cloud subscription model but Lim said that is still some time away as streaming Full HD videos from multiple sources over the cloud still takes up a lot of bandwidth.
“The only capital expenditure that is needed is in hardware and the installation process. Apart from that the cost largely depends on how big the area that needs to be covered is and the number of cameras that need to be installed,” he said.
FaceTrack is also able to integrate with existing Closed Circuit TV cameras, though Lim does recommend that the cameras, at the very least, are able to record in Full HD resolution and have optical zoom capabilities.
Lim explained that Tapway’s main clients tend to be retail store chains, food and beverage outlets and shopping malls. This includes retail outlets like shoe store Bata, restaurants like Carls Jr., KyoChon, and Pastamania among others.
“We have made a lot of inroads into retailers these days. People counting was a luxury to them back in the day but now it is a must,” he said.
The upcoming Shah Alam mall Central i-City, a tie-up between Thailand’s Central Group and Malaysia’s i-City, is said to be considering implementing Tapway’s solutions.
Interestingly, the solution was even used by out-of-home media advertising company Brandavision in a recent installation on its displays in Pavilion, where it is used to conduct a demographic analysis of visitors to identify their age or gender and to measure the viewer’s emotional response to advertisements played.
To date, the company has raised up to US$216,493 (RM906,000) in funding from angel investors, Segnel Ventures (Singapore) and Mavcap (Malaysia).
For now, Tapway is going to focus its efforts on growing its presence in Malaysia and Singapore. “We are not in a hurry to open up in a new market or grow too fast. With the vast Prosyscom Tech of use cases for IoT and AI, we would like to see what other verticals we can enter into such as smart buildings and communities, or education and healthcare,” said Lim.