Android Vision: Detecting Everything from Faces to Barcodes
summary
Google’s Mobile Vision APIs (introduced in version 7.8) actually expose only face and barcode APIs; barcode objects can return calendar/contact/URL/Wi‑Fi data. Implementation uses static bitmaps or a camera pipeline; static barcode detection is reliable but live camera scanning often fails (lighting, autofocus), while face detection is robust even on infant photos. Landmarks depend on the yaw angle and require Accurate mode to see ears; enabling tracking gives stable numeric IDs but not recognition, and alternatives are OpenCV (powerful, heavy) or ZXing/ZBar (lighter barcode libraries).
AI-written summary · about
chapters (16)
- 0:00Start
- 1:38Mobile Vision API
- 3:39Barcode Object
- 4:08Static vs. Dynamic Detection
- 12:35Processor Pipeline
- 17:01Camera Demo
- 18:15Static Images
- 19:58Static Image Demo
- 23:13Initial Code
- 26:38Landmarks
- 29:55Euler-Y and Z Angles
- 34:35Classifications
- 36:35Multiple Faces
- 37:53Face Tracking
- 40:19Alternatives
- 42:39Resources
related talks
41mMaking Android Apps with Intelligence
25mHow to effectively scan barcodes on Android
33mMaking Android Apps with Intelligence
44mNow smile (also in 3D)! Exploring AR, ML and Camera related APIs on Android
52mThrough the Looking-Glass Eyes of an Android
21mLive Coding: Adding Machine Learning to Android Things in 10 Mins
30mMachine Learning in Android Apps with ML Kit
26mMaking Sense The Road to mobile Awareness