Abstract
This paper describes the creation of a cutting-edge AI-powered book reading assistant intended to help people with vision impairments overcome the difficulties involved in accessing printed and digital books. Conventional methods frequently depend on static image processing or PDF readers, which are neither very flexible or useful for dynamic content like books with intricate layouts or different typefaces. Due to a lack of real-time help that can manage a variety of formats, unstructured text, and dynamic page-turning, many visually impaired people find it difficult to read physical books. To solve these problems, this research introduces a video-based text recognition system that enables text from physical books to be processed continuously and in real-time. The system utilizes cutting-edge technologies such as natural language processing (NLP), machine learning, and computer vision to provide accurate and efficient text recognition and reading.
These assistant records real-time video feeds from books and processes the content dynamically as the user interacts with it, in contrast to previous approaches that either capture a static image or rely on pre-existing digital forms like PDFs. By responding to voice commands, the assistant gives users control over their reading experience and provides them with audible feedback. With increased flexibility thanks to this method, people with visual impairments can read printed books more naturally by flipping the pages and hearing the content read out as they go. The technology is a major step in developing more accessible and user-friendly reading tools, even though it hasn't been tested with visually impaired people yet.