AI-Powered Visual Memory Assistant for Blind and Visually Impaired Users
MEMORA2.0 is an assistive AI system designed to help blind and visually impaired users locate specific people and everyday objects using computer vision and visual memory.
Instead of only describing a scene, MEMORA focuses on finding the specific person or object the user is looking for.
Python · OpenCV · YOLO-World · DINOv2 · ByteTrack · PyTorch · Computer Vision
MEMORA2.0 uses Amazon S3 as the cloud-storage layer for its visual-memory architecture.
The real-time computer-vision pipeline runs locally for responsive processing, while AWS provides the cloud-storage component for the project's memory architecture.
Our long-term goal is to create a memory-based visual assistant that can learn familiar people and important personal objects, recognize them later, and guide a blind user toward them.