Implementing Passport OCR with CameraX × ML Kit
summary
Describes implementing passport OCR on Android using CameraX 1.1.0 preview image analysis together with ML Kit on‑device Latin text recognition to capture frames and run OCR. It extracts the MRZ using regex, manages orientation via dynamic targetRotation and correct ImageProxy lifecycle, and applies common-character mappings and validation to auto‑fix or present corrections; device‑specific failures are logged and retried based on tests across Galaxy, Pixel and Xperia devices.
AI-written summary · about
chapters (9)
- 0:00OCR introduction
- 4:19CameraX and ML Kit
- 7:29Dependencies and manifest
- 10:40Preview and camera setup
- 12:47Image analysis and passport
- 15:37Screen rotation and accuracy
- 17:45MRZ extraction and UI
- 18:49Accuracy improvements and validation
- 21:00Post-release monitoring
related talks
39mNow smile (also in 3D)! Exploring AR, ML and Camera related APIs on Android
29mAndroid Jetpack: Understand the CameraX camera-support library
12mML Kit: Turnkey APIs to use on-device ML in mobile apps
30mMachine Learning in Android Apps with ML Kit
34mML Kit x Material Design: Design Patterns for Mobile Machine Learning
33mImprove (and Simplify) Data Capture With Google’s ML Kit
24mReact Native Ignite: Building an AI app
38mML Kit: Machine Learning for Mobile with Firebase