• +91-9916368800
  • nfbkarnataka@gmail.com

NFB News Alert Gurugram girl's visionary move: Virtual-reality glasses for visually impaired. Dated

Amity School student Saumya Chauhan hopes her object-detection device will help the visually challenged handle daily challenges more confidently

Innovation and entrepreneurship have nothing to do with age. Saumya Chauhan, 17, is a testament to that. The Class 12 student of Amity Global School in Gurugram has developed an intelligent vision-based object detection device for the visually/partially impaired to help them get through their daily lives with some ease.

The idea dawned upon Chauhan when, as a Class 8 student working with some NGOs, she happened to witness first-hand the problems faced by those with failing eyesight. It made her think hard how she could help them. A bit of research and Chauhan soon began work on her project, giving a final shape to the prototype early last year.

To be used like a headgear, the virtual reality goggles/glasses are made from biodegradable material. Being environment-conscious, Chauhan says she didn't want her innovation to harm the planet. "In future, after these devices are used to an optimum number, I will give the option of buy-back to recycle the e-waste," she says.

The algorithm used in object recognition is YOLOv4 (YOLO is a real-time object recognition system that can recognise multiple objects in a single frame; YOLOv4 is its fourth version) and TensorFlow Lite (a deep-learning framework). Her device has an accuracy of 67.23 per cent and an object detection time of 0.03785 seconds. Initially, she used an Arduino UNO board (a microcontroller board), later shifting to a Raspberry Pi board (single-board computers).

The visually impaired need to be aware of their surroundings and recognise objects to avoid challenges and mishaps. The YOLO object identification technique recognises the item in the real world. The single-board Raspberry Pi circuit makes it easier to install the camera sensor. The board is cut to the size of a little credit card and object detection is through a web application.

Chauhan, who spent about Rs 14,000 on her project, now plans to start testing the prototype. Ask about the experience of working on the device and getting recognition (a research paper has been filed on it in the International Journal of Research and Analytical Reviews), Chauhan says the feeling is unique and amazing. "You feel you can accomplish anything and not be limited by age and resources."