WATCH: Drishti Technologies' computer vision algorithm makes humans efficient
Every now and then, there will be a technology attempting to change the world and a startup rallying behind it.
Just like the Kanban system, which was developed by a Toyota engineer. Kanban is aimed at enhancing the delivery of parts in car manufacturing with an algorithm-powered computer vision.
It is a card system using which the manufacturing line would know exactly what is being consumed. While the system solved for just in time, it did not solve for quality. Automobile enthusiasts would know that the automobile industry had thousands, if not millions, of recalls.
In fact, every year a car company recalls 10,000 vehicles on an average, which is a loss of billions of dollars thanks to legal hassles and consumer rights. But now, Bengaluru- and Palo Alto-based Drishti is solving this problem using computer vision to track quality of production and just in time movement of parts.
In a conversation with YourStory, the founders of Drishti reveal more:
Drishti’s software is not an ERP system; it is a visual network run by a machine learning (ML) algorithm that can monitor the assembly line and production quality. If a person or team is assembling a transmission or battery or any part of the car, the computer vision software does not allow the conveyor belt to move until the part is fitted to perfection. It is only when the software is satisfied will information be passed upstream and downstream in the production line for the inventory management system to know what part needs to be sent next.
This startup is the brainchild of Ashish Gupta, Prasad Akella, and Krishnendu Chaudhury who have four decades of experience in building algorithms and working in global companies. The trio got together in early 2017. While Ashish is a well-known investor from Helion, Prasad had worked in SAP, and Krishnendu had worked in Flipkart, Google, and Adobe.
“This idea began in 2016 when I quit my job and wanted to be an entrepreneur. I ran this idea past Ashish, who had been a long-time friend, and he was all in. But, we had to find a man who could build this computer vision product,” says Prasad, Co-founder of Drishti Technologies.
He adds that the startup is going to annotate industrial data by visual means.
Prasad and Krishnendu
The man from Bengaluru
In 2016, Flipkart was making cutting-edge technology by creating an ecosystem of consumers and small businesses. Krishnendu was part of that team, but his yearning was to build an image processing and computer vision platform. As luck would have it, he ran in to Prasad at a conference where the two hit it off.
By October 2017, the product was being built. They invested a good amount of money (they don’t want to disclose the numbers) and began building the product out of Bengaluru. However they received $10 million from Emergence Capital, in the USA, last year.
At present, the product is commercial with over five companies globally. “This platform allows for humans and machines to work together, taking IoT to the next level,” says Prasad.
According to a new study published by Polaris Market, the global industrial internet of things (IIoT) market is anticipated to reach $771.72 billion by 2026. Industries worldwide are looking for new business models to construct a connected enterprise to merge their operational and information departments. This transformation is expected to enhance overall productivity, operational efficiency, and visibility as well as decrease the complexities of diverse procedures in the industry.
Companies like Bosch, Siemens and Continental are all focussed on industrial IoT. “One may think that robots are replacing humans in industry. But 70 percent of the world is still a combination of men and machines, we intend to make this partnership stronger,” says Krishnendu.
“Manufacturing is going to change with real-time algorithms using computer vision. This is the future and this form of data is not annotated. Imagine being able to track why a vehicle failed! We can point to the exact time of assembly of the part and also show what happened in the assembly line. Imagine the costs saved for the manufacturer and the brand,” adds Prasad.
The company is going to win over as many corporates as possible in the next 12 months and beyond. The team does not want to reveal its business model but believes that because of data-based outcomes, the company can charge a licence fee and also charge for the outcome.
This power trio want to change the industrial world, and that revolution begins in Bengaluru.