DreamEyes- AI assistant

How could deep learning generate a narrative description of the surrounding environment?As of 2021, the world health organization (who) estimates that there are approximately 36 million visually impaired individuals worldwide[1]. This disability can significantly impact their ability to perform daily tasks and engage in social and professional activities. However, visually impaired individuals have the same potential as those with normal vision and can be just as productive.Therefore, it is crucial to take advantage of technology and advancements in assisting devices to create tools to support them. Recent developments in the ai industry, particularly in deep learning, have spurred researchers to explore the creation of such assisting devices.

DreamEyes is an AI based project that aims to translate real-world data into rich and understandable paragraphs that will be converted into audio format using deep learning models. This will greatly contribute to improving the interaction of visually impaired individuals with their surrounding environment. The prototype consists of a wearable smart cap equipped with internet of things (iot) components such as microprocessors that analyze real-world data and inform the user of what is happening around them. The algorithms used in this project are inspired by the learning mechanism of the human brain and are based on deep learning, a subset of machine learning where deep neural networks learn and make decisions from complex and large data sets. Dreameyes is designed to replace traditional solutions that are limited in their technology and features. This device will allow individuals with partial to full vision loss to communicate with the external world and have an enhanced experience of their surroundings.In this report, we will focus on describing the implemented methods and approaches followed during the design as well the development of this contribution. The scientific part will present a descriptive top-down approach of deep learning and the sequence models architecture whereas the technical part will cover the development and deployment stages of the program.