20-year-old Madhav Lavakare from Delhi has built Transcribe Glass, inspired by Google Glass, a device that processes information from live conversations, Google, Microsoft and other platforms to convert speech to text
In 2017, when Madhav Lavakare was in Class 11, his friend, who suffered from a hearing impairment, had to quit school. When Madhav probed the matter to find out why, he learned that it was because his friend found it painfully difficult to communicate with his peers and teachers due to his condition. As the issue became insurmountable, he had to stop attending classes.
But rather than turn a deaf ear to his friend’s problem, Madhav decided to find a way to help him.
So this year, the 20-year-old has built an innovative TranscribeGlass, an assistive technology device for the deaf, those who are hard of hearing, and the elderly in need of technological aid.
An affordable solution
“My friend was deaf and hard of hearing (DHH). It affects the understanding of linguistic information and impacts performance in an educational environment,” Madhav tells The Better India.
Madhav says he was disheartened that his friend had to quit studies due to his condition. “I wanted to make learning easier for him in every possible way. He found it difficult to follow what was being taught in class and communicate with his friends. He needed a solution that could improve his linguistic comprehension,” he explains.
He asked his friend if could use hearing aids or cochlear implants. “But he said both the solutions were expensive, especially the cochlear implants, which can cost lakhs. I asked him about speech-to-text apps, but he said the caption quality was not up to the mark. It also demanded the user to look at the phone every time they wanted to read the captions. It would be inconvenient to understand what was going on in class while having to constantly refer to the app on the device,” he notes.
Madhav began to think of a solution that could translate information into text captions and put them on a heads-up display. “The concept stemmed from Google Glass, where it showed information in the user’s field of vision through augmented reality. But the former was expensive and I was determined to make the solution affordable,” he adds.
He began researching on Google, watching YouTube videos, and teaching himself aspects of programming, mechanical engineering, design, optics, electronics and other aspects for building the device. “I approached a few technical mentors in the process, including Professor P V M Rao at IIT Delhi, two persons from Google in the United States, and two working professionals from the UK and India, both with hearing loss. They gave me encouragement, guidance and essential feedback to understand the challenges faced by people with hearing impairment,” he says.
For the first few years, Madhav worked on the project during his spare time. By the end of 2018, he was running a small crowd-funding campaign on Ketto and raised about Rs 2.5 lakh.
“In 2019, I founded a startup named TinkerTech Labs and built two versions of the product. In 2020, I received a grant under the Pfizer-IIT Delhi Innovation Program, who incubated the startup. And towards the end of the year, another grant came from the US-India Science & Technology Endowment Fund (USISTEF),” he says, adding that using the grants, he hired a team of engineers to build better versions of the prototype.
After four prototypes and five years of work, Madhav conceived the device.
Explaining its functionality, he says, “The user first launches a mobile app and chooses the source to caption information that converts speech-to-text. The caption source can be automatic speech-to-text models from companies like Google and Microsoft, or live human interactions. The captions received are then transmitted via Bluetooth to a display within the hardware device.”
He further says that the hardware device is attached to one of the glasses on the spectacle frames worn by the user. The captions from the display are projected in realtime in the wearer’s field of vision using the principles and optics of augmented reality. It enables the wearer to see the person speaking at that moment and read the captions comfortably.
Madhav says that his most recent prototype has given the desired result, and is lighter and sleeker. “It weighs about 20 grams and costs less than Rs 4,000. It also has optical adjustments to reduce the strain on the eye, allowing usage for a longer period,” he adds. The innovation also won Madhav the NCPEDP-Mphasis Universal Design Awards 2021.
Sharing his challenges, Madhav says, “The optical system demanded a lot of work. It was also difficult to identify vendors to cut the lens and mirrors to the right size to achieve the desired 3D printed enclosure. Since there was no initial funding, I relied on pro-bono help from other young professionals who helped me develop the firmware and the mobile application.”
He says that the first 100 devices are in the making and will be shared with beta testers. Speaking on further improvements, Madhav says, “The device currently displays captions only in English. Plans include captions in regional languages, ability to translate in addition to transcription, and a better product design.”
“I hope to make the desired difference to the lives of many suffering from hearing ailments through the device,” he says.
Edited by Divya Sethu