Amrita Vishwa Vidyapeetham researchers develop eye-to-speech device

NEW DELHI : Researchers at the Humanitarian Technology (HuT) Labs of Amrita Vishwa Vidyapeetham, have developed an eye-to-speech device. The eye-to-speech device is India's first device enabling individuals with speech impairments to communicate with the world using simple eye signs. The device has been named Netravaad.

Netravaad, as per an official statement, encompasses a range of essential components, including a camera, display, speaker, controller, and rechargeable battery. The camera's primary function is to detect eye signs, which are then converted into alphabets, words, or even complete sentences through a custom-built AI algorithm called Sharani. The detected words and sentences are then displayed on the device's screen.

Additionally, the statement added, the device incorporates a speaker that vocalises the text. The Netravaad device is set to undergo further development to include regional languages such as Malayalam and Hindi, allowing users to communicate in their preferred language.

Rajesh Kannan Megalingam, Director, Humanitarian Technology (HuT) Labs, Associate Professor, ECE Department, Head, Amrita Center for Advanced Robotics said: "We created a new language using eye signs, called Netravaani which allows users to utilise all 26 English alphabets to create any word and sentence. It also supports the numbers 0 to 9, with which the users can create any number,“ the head added.

Amrita Vishwa Vidyapeetham Development: Advantages of Netravaad

One of the notable advantages of Netravaad, it added, is its portability and lightweight design, allowing it to be adjusted to any position for the user. The Netravaad unit comprises a display monitor, camera, speaker, smash box, and a rechargeable 24V battery. By tracking the user's eye movements, the camera converts them into words or sentences, which are then displayed on the screen and vocalized through the speaker.

Netravaad is portable and lightweight device

Kannan said, "Netravaad supports two modes, English 1 and English 2. In the first mode, patients can select the ten most commonly used words for day-to-day activities, enabling quick communication. The English 2 mode offers word, sentence, and number sub-modes."