Camera, PhotoKit, image pipelines, and Vision
Image work is a pipeline: acquire, authorize, decode, orient, transform, store, thumbnail, analyze, display, and delete. Each stage has different memory, privacy, and device evidence.
Design the capability
- Prefer system pickers when they satisfy selection. Treat limited-library access as a first-class state.
- Capture through a device adapter and preserve orientation, color, metadata policy, and stable attachment identity.
- Generate bounded thumbnails instead of decoding every full-resolution image into scrolling views.
- Run Vision analysis off the UI path, expose confidence and correction, and keep local text recognition visible to the user.
Validation boundary
No camera, real photo library, limited-library prompt, image orientation matrix, Vision request, or physical-device memory run occurred.
Series navigation
- Previous: Part 100: MapKit, Core Location, geocoding, and WeatherKit
- Next: Part 102: Audio, speech, recording, and interruptions
- Series index: Zero to iOS Hero