The device, designed as an AI-powered wearable system, analyzes environmental data to help guide users who are blind or visually impaired.
Chinese Researchers Develop AI-Powered Wearable Eye to Assist the Visually Impaired
Chinese scientists have developed an AI-powered wearable system designed to function as an artificial eye, helping blind and visually impaired individuals navigate their surroundings with greater confidence and independence.
According to a study published in Nature Machine Intelligence, the system combines a camera, an AI processor, and bone-conduction headphones to analyze real-time environmental data. It then guides the user through simple audio cues and vibration alerts, offering intuitive feedback for safe movement.
The project is a collaborative effort between several institutions, including Shanghai Jiao Tong University, the Shanghai Artificial Intelligence Laboratory, East China Normal University, the Hong Kong University of Science and Technology, and the National Clinical Research Center for Neurological Diseases at Fudan University.
An AI-Powered Alternative to the Human Eye
“This System Can Partially Replace the Human Eye,” Says Lead Researcher
Dr. Guo Lili, professor at Shanghai Jiao Tong University and the project’s lead researcher, explained:
“This system can partially replace the human eye.”
She emphasized that the research team aimed to design a solution that’s simple, lightweight, and practical—something users can wear throughout the day without discomfort or fatigue.
Unlike autonomous navigation systems that often provide long and complex instructions, this wearable delivers only essential information to help the user navigate. According to Dr. Lili, it’s designed to communicate just what’s needed—information the brain can quickly and easily process.
In initial indoor trials involving 20 participants with low vision, most users were able to comfortably operate the system after just 20 minutes of practice. Setting a destination is simple: a voice command is all it takes for the system to identify a safe path and provide brief, accurate guidance along the way.
Currently, the device can recognize 21 different objects, including beds, chairs, sinks, TVs, and various types of food. The research team plans to expand its capabilities for outdoor use in future phases. Upcoming improvements include better object recognition, adapting to environmental conditions, and integrating with GPS for outdoor navigation.
Leave a Reply