PARTIAL BLINDNESS ASSISTANCE USING IMAGE CAPTION GENERATOR THROUGH USING TRANSFORMER AND TEXT-TO-SPEECH

Abstract

This project discusses the global issue of vision impairments and the WHO Global Report on Vision, emphasizing the significant number of avoidable impairments. It highlights the significant number of avoidable impairments and emphasizes the importance of taking action. The project presents data from the Rapid Assessment of Avoidable Blindness (RAAB) conducted in Indonesia from 2014 to 2016, indicating the prevalence of Moderate to Severe Visual Impairment (MRVI) in the population. It delves into the emotional impact of visual impairment, its effects on individuals, families, and society, and the broader consequences on independence, healthcare, employment, and quality of life. The project also acknowledges the potential benefits of AI/ML technology, specifically the Image Caption Generator combined with Text-to-Speech, in assisting individuals with impaired vision. It explains the chosen methodology, Rapid Application Development (RAD), and describes its four-step process of defining project requirements, user design, rapid construction, and cutover. The final part highlights the successful development of an Android application capable of generating captions from images using a transformer algorithm and providing text-to-speech functionality in English and/or Indonesian.