Face and pose recognition plugin
A public plugin connecting Android and iOS face and pose recognition through one Flutter interface. I normalized platform-specific camera input and errors and reuse the plugin in two personal apps.
Screens
Common inputs and recognition results
The plugin accepts still images and camera streams and detects face and body landmarks separately or together. Helper functions make MediaPipe results available for expression, gaze and pose analysis.
Containing platform differences
I implemented parallel YUV conversion on Android, vImage processing on iOS and front-camera mirroring corrections. Structured errors let Flutter apps handle results through the same interface.
Publication and use in apps
I published the package on pub.dev and integrated it into an exercise counter and an AR camera. Each app uses the plugin’s results for movement decisions or visual effects instead of rebuilding recognition.
Architecture
