In Novosibirsk, they presented the first version of the intellectual assistant for visually impaired and blind people. The device is developed on the basis of large language models and technologies of artificial intelligence (AI). His work was demonstrated during the visit of the deputy chairman of the Government of the Russian Federation Dmitry Grigorenko to the Novosibirsk State University.
The project of the project was compiled by the specialists of the Center for AI NSU and the MNTK branch “Eye Microsurgery”. Their goal is to create a compact wearable device that can help a person navigate in space with hearing.
The principle of operation of the system is simple: the video camera fixes the environment, the image is transmitted to a portable computer that converts it into a text description. Then the description is voiced through a voice assistant and transferred to the user to the headphones.
The created model is already able to recognize the inscriptions, indicate obstacles and give verbal descriptions of objects and premises. In the future, it is planned to add the opportunity to ask questions about objects in the field of view and get detailed answers.
Development is based on a multimodal language model, which allows you to achieve high accuracy of the description. According to scientists, this can become a more affordable alternative to complex and expensive medical implants.
The prototype of the device is promised to prepare by the end of 2025. It will consist of a 16-core single-pay computer, video cameras and headphones. The total weight of the system will not exceed one kilogram.
News -in -law materials cannot be equated to the doctor’s prescription. Before making a decision, consult a specialist.
