Vision Bridge: Empowering Blind Individuals Using Machine Learning

Abstract

Visual impairment and blindness are the usual outcomes of eye disorders. The World Health Organization estimates that 250 million people worldwide suffer from some form of visual impairment and 40 million individuals are blind. In their daily lives, they encounter a lot of difficulties, particularly when they are traveling alone. To cover their basic needs, they usually depend on help from other sources. Thus, it's a challenging task to establish a technology solution to support them. To assist the blind, numerous devices have been developed or visually impaired. One such endeavor is to create on the other hand, teaches blind sufferers to recognize and categorize every conceivable everyday object a voice warning the listener of both nearby and distant objects in their environment. Two distinct algorithms are used in the development of the system: Yolo is used for object identification, while GTTS (Google Text to Speech) is used for audio warnings. Both techniques are tested using the MS-COCO Dataset, which has about 200 K pictures. Webcams are used to assess both algorithms in a range of scenarios to be able to gauge each algorithm's accuracy under all conditions.