Hearing aids or cochlear implants designed for deaf people or with auditory difficulties can be, in addition to expensive, uncomfortable or uncomfortable and, in some cases, not work properly or not finish solving the needs of users. In his search for a solution, Madhav Lavakare, a computer engineering student at Yale University, has developed a system that combines smart glasses, subtitles and functionalities of artificial intelligence.
Earl this year he officially launched Transcribess, The company behind its functional innovation, glasses with the capacity to subtitle and translate conversations in real time. Has worked on the project, as stated Yale Daily Newsfor seven years, when the institute still was in, and has carried out seven different versions of the prototype applying its knowledge in hardware, mechanical engineering and 3D printing.
The result of all years of research, work and tests is transcribess, which transcribe spoken conversations and project subtitles on the glass right in front of the eyes. The glasses have no cameras, microphones or speakers, only a small projector that Project the subtitles directly on the glass. However, the device requires connection with a mobile application, so that it generates the subtitles from the conversations captured by the phone’s microphone.
The application It allows to customize the subtitles, both in regards to the size of the source and the text line shown at the same time, as well as their location in the field of vision. The glasses also have Bluetooth, a battery with autonomy for 6 hours, and weigh less than 40 grams.
The price of the device is 377 dollars, plus a Monthly subscription of $ 20 to access the transcription service and the discourse recognition system. Transcribeglass is currently available for iOS system, and in English, Spanish, French, Chinese, Italian, German, Portuguese, Korean, Japanese and Vietnamese.
They are designed for deaf people and, mainly, for those who have hearing loss and have difficulty reading the lips or following a conversation in a noisy environment. In this sense, the main focus of the company is the functional simplicity, That is, to improve the system to facilitate user communication.
It is what, as stated Wired, Lavakare considers that it is what differentiates their device from others existing in the market, such as Even Realities, Xrai or, even Google or Meta, which are making more colorful intelligent glasses and with more functions, especially linked to entertainment, such as recording videos and taking photos or interaction with chatbots.
“All these smart glasses exist, but no one has yet found a truly useful case of use”, Lavakare commented to the aforementioned medium.“We believe we found it: one that has enormous value for the end user”
In this regard, the company is betting on the artificial intelligence To improve your proposal, and work in your application for the Recognition of emotional tone of the speaker from the voice. Transcribeglass is testing with the function of labeling emotions during conversation with the intention of facilitating greater context, understanding and emotional processing to facilitate communicative exchange, for example, for people with autistic spectrum.
Throughout the last five years, Lavakare has conducted field tests with more than 500 deaf people and with auditory difficulties in the United States, Europe, Australia and India. And has collaborated with various advisors and collaborators throughout the development process. He currently works with manufacturers that provide hardware, Among them Vuzix, and installs their software in the glasses.
As collected Yale Daily NewsLavakare’s intention is to continue working on the product to make it accessible To everyone, a process that he considers will still require time. He is satisfied with the impact generated and the way users have adopted glasses, and imagines a not too distant future in which they can be used in everyday situations, such as going to church, the theater or talking during a dinner with friends.
