Startup receives funding from Indian government for real-time biometric and thermal scanning solution
A funding body with India’s government has announced a loan for biometric facial recognition provider Cocoslabs Innovations to develop a low-cost system for identifying people with fever amongst a crowd in real-time, along with a range of other technologies for combating the spread of COVID-19 in the country.
The Department of Science and Technology’s (DST’s) Technology Development Board (TDB) announced funding for six startups in the country
Cocoslabs is expected to commercialize the solution, which provides an alert to the mobile phones and laptops of area authorities when an individual in a crowd with a fever is identified.
In addition to backing the thermal scanners with video analytics from Cocoslabs and another company, Advance Mechanical Services Private, TDB announced support for a portable X-ray machine, 3D printing anti-viral agents onto masks, and rapid diagnostic kits.
“The time of COVID-19 has shown us many paths to commercialization of indigenous technologies and products by seamlessly connecting the push of knowledge from academia and research labs to the pull of market with a clear purpose, relevance, need, ease of support, transparency, dedication, cooperation and accountability. The speed and scale of TDB support are some of the compelling examples of our new normal that would continue to serve in the Mission Swavalamban– Self-sufficient India,” states DST Secretary Professor Ashutosh Sharma in the government announcement.
Cocoslabs Founder Prathvi Palekar told MediaNama the company outlined the sectors and use cases it makes products for in a pitch to the TDB. Co-founder Akshata Kari told the publication that Cocoslabs believes the solution it pitched is worth a million dollars, and that TDB will cover roughly half that amount.
Exact terms of the loan are confidential, but it is reported to be a low-interest loan, with interest starting to accumulate after a year, and a five-year term.
The system can be rolled out in seven to ten days, though lead times could grow as Cocoslabs switches from ordering cameras as needed to bulk orders. Cameras implemented by Cocoslabs are developed in Germany and Sweden, in addition to a lower-cost model developed in Taiwan. Running Cocoslabs’ software, the body temperature of up to 30 people can be recorded at once. Temperature is read from the corner of the eyes, unless the subject is wearing glasses, in which case the forehead is scanned.
Cocoslabs’ Pixuate video analytics platform, which was trained on 3 million images, underpins the software, MediaNama reports. The solution also analyzes subject age and gender, and detects masks. Export logs with the information collected by the solution can be downloaded as PDFs, which the company recommends securing with a password.
The dataset was compiled through the company’s early partnerships, and did not involve breaching anyone’s privacy, Palekar claims. He said facial data has been collected from open source platforms on Google, but not social media, as those images tend to be compressed beyond optimal usefulness for training AI.
The company is planning to submit its facial recognition algorithms to NIST. Future development plans including recognition of people wearing masks, and for its software to identify people’s ethnicity, which it hopes to achieve within two to three months.
According to Kari, Cocoslabs is currently in talks with the Karnataka government to deploy its solutions in public institutions.