Vision framework

Apple's high-performance computer vision framework for on-device image analysis, face tracking, and text recognition.

Vision streamlines complex computer vision tasks by providing a unified API for Apple's hardware-accelerated machine learning chips. It handles core requirements like face detection (including 2D and 3D landmarks), barcode scanning (supporting 17 symbologies), and optical character recognition (OCR) in 18 languages. Developers use it to track objects across video frames, classify scenes, and perform body-pose estimation with 19 distinct keypoints. By running locally on the Neural Engine, Vision ensures low-latency execution and user privacy without requiring a network connection.

https://developer.apple.com/documentation/vision