Indian facial recognition startups tackle retail and KYC markets
ParallelDots has won the $50,000 Tech4Future challenge from among 300 startups for its artificial intelligence platform with biometric facial recognition for consumer brands and retailers, Inc42 reports.
The Tech4Future challenge is a collaboration between SoftBank and Invest India under the Japan India Startup Hub, and was held to identify the most promising startups in AI, machine learning, facial recognition, and cybersecurity. The competition was launched in November. In addition to the cash prize, ParallelDots will be offered a two to three-month incubation opportunity in Japan.
The company provides modules for image, video, text, and speech analysis, with features including object detection and tracking, facial recognition, emotion detection, sentiment analysis, and keyword identification. The machine learning-based technology can be used for analyzing retail shelf images, videos for retail environments, and open-ended text interactions, according to the report.
ParallelDots has previously raised $2 million, and has customers in the U.S., Europe, East Asia, and India.
KYC opportunity knocks
With know your customer (KYC) rules for banks and fintechs in India still evolving, and rapid growth in digital services in the country, Veri5Digital is offering an instant KYC service with biometric selfie photo or video liveness checks.
The company is an initiative of Khosla Labs, formerly known as Aadhaar Bridge, which is a Bengaluru-based incubator and innovation center. Veri5Digital extracts data from most government-issued ID cards, and checks documents for classification and forgery. It also provides database checks, and offers eSignature options, Aadhaar-based KYC, and a recurring payment platform called eNach.
Khosla Labs CEO Saru Tumuluri told Analytics India that the company provides its platform to Aadhaar-licensed entities like banks, insurers, and telcos, increasing their onboarding speeds while ensuring compliance with Aadhaar regulations.
“Our video ID KYC solution is completely driven using AI,” Tumuluri says. “We do a bunch of things such as document classification, ID card forgery checks, liveness check, and a live selfie capture, and face biometrics match to do authentication.”
Tumuluri also says the company’s facial recognition algorithm is one of the most accurate in the world.
“Our models are extensively trained on Indian data including PAN cards, Voter ID Cards, passports, Aadhaar, and other datasets. The best face algorithms excel at accuracy (across different lighting/angles of capture), blurriness, diversity, and occlusions.”
Some regulators in India currently allow video KYC processes, while others are running tests on the technology in regulatory sandboxes. Veri5Digital counts a number of fintech companies among its clients, and was founded to drive greater financial inclusivity. In the future, the company plans to support multilingual applications, as well as the use of other ID credentials, such as ration cards.