GEMID.
Starting gemstone identification with a photograph.
A mobile application foundation for account access, guided onboarding, and camera-based image preview.
01 / The challenge
Make the first step approachable.
My CV describes a broader GEMID project involving ML image classification, inclusion and damage analysis, and Three.js-based 3D visualization. The supplied source snapshot does not include those components. GEMID’s broader ambition is AI-assisted gemstone identification. The reviewed prototype focuses on the mobile experience around that idea: explaining the workflow, acquiring an image, and preparing for a later analysis integration.
02 / The approach
Introduce, capture, review.
Registration and login lead into a short onboarding sequence. The home screen offers capture, history, settings, and help. The camera path requests permission before opening the device camera, and an image preview provides a checkpoint.
03 / The interaction
A useful pause before analysis.
The selected image stays visible before the next action. This establishes a place for future image-quality guidance, retakes, and processing feedback. The illustrated history uses clearly labelled demo entries to show a populated component state.
04 / The engineering
Build the application foundation.
Expo Router structures the client. Axios connects account requests to an Express backend, with MongoDB and Mongoose for user records. The backend includes password hashing and JWT issuance. Expo Image Picker supports capture; the document-picker path needs compatibility work.
THE TOOLKIT
What this version demonstrates
Scan Now is a placeholder alert in this archive. Model inference, prediction results, and interactive 3D rendering are not included. Profile and settings screens are partial local implementations. No identification accuracy or completed scan results are claimed.
Have a workflow worth improving?
Let’s discuss it ↗