App Profile: FindyARI

FindyARI helps people with low vision to independently locate everyday items (like keys or TV remote) by training an on-device personalized AI model with photos of their belongings, and then guiding users with real-time voice directions and AR cues to the item’s real-time location through the device’s camera.
1. Save and recognize objects
Findy allows users to save objects by taking six photos, providing a comprehensive view for accurate recognition.
2. On-device AI training
Using Create ML, Findy trains a custom AI model entirely on the user’s device, ensuring privacy and personalization.
3. Enhanced image processing
The Vision framework improves image data by separating backgrounds, isolating objects, and generating relevant tags for better AI training.
4. Live object detection
With CoreML, Findy analyzes the live camera feed to identify saved objects in real time.
5. 3D navigation assistance
ARKit converts 2D object coordinates into 3D spatial data, and RealityKit renders an arrow pointing to the item’s location.
6. Spatial audio guidance
Findy provides voice instructions and spatial audio cues to guide users toward their objects.
7. AI-powered image enhancement
Users can leverage via Image Playground to generate high-quality preview images of saved objects.