Real-World Mapping Model for Visually Impaired People

The A-EYE model is an advanced assistive technology designed to empower visually impaired users by enabling them to perform daily tasks independently, reducing their reliance on sighted individuals. This AI-powered system provides real-time voice guidance, ensuring seamless interaction with its features. The object detection feature uses YOLOv4 on the COCO dataset to accurately identify everyday objects, enhancing users' spatial awareness. The face recognition feature utilizes Google ML Kit for detection and MobileFaceNet for recognition, allowing users to identify familiar faces effortlessly. For financial independence, the system incorporates currency recognition, employing a VGG19 model trained on a Pakistani currency dataset to distinguish different denominations accurately. Additionally, text recognition is powered by Firebase ML Kit, enabling users to read printed material with ease. The emergency shake feature ensures safety by automatically sharing the user's location with selected emergency contacts when the phone is shaken. The model's accuracy will be enhanced by refining training datasets and implementing advanced AI techniques, while the currency recognition model will be expanded to support a global currency dataset for greater versatility and usability worldwide.