What problem does it solve?
It helps iOS apps reliably recognize text, detect faces, scan barcodes, and derive masks or regions from images and live camera frames, without writing bespoke computer-vision pipelines.
Core Features & Use Cases
- OCR (Text Recognition): Extract printed or camera-captured text using the modern async Vision requests and confidence-scored results.
- Face & Capture Analysis: Detect faces, landmarks, and capture quality for selecting or guiding better photos.
- Barcode & Document Scanning: Read barcodes (including QR) and extract document text with layout-aware structure, plus camera scanning via VisionKit.
- Segmentation, Tracking, and ML Inference: Generate person/object masks, track objects across frames, and run custom Core ML models through Vision.
- Use Case: When users point a camera at printed medication labels during a self-check workflow, apply text recognition for key fields, scan a QR for context, and use structured confidence thresholds to drive the next UI step.
Quick Start
Use the vision-framework skill to design a modern iOS 18+ OCR and barcode detection flow by calling Vision requests with async/await on a CGImage or CMSampleBuffer and converting bounding boxes to UIKit coordinates for display.
