sprintliveness 0.0.8

Flutter plugin for face liveness detection. Fully supports Android && Ios platform.

0.0.1

initial pulish

• TODO: Describe initial release.

0.0.2

bug fixed

0.0.3

• Fixed iOS Head Up/Down Movement Detection issue
• Corrected pitch angle handling for iOS front camera by accounting for mirrored coordinate system
• Fixed iOS Left/Right Movement Detection issue
• Corrected yaw angle handling for iOS front camera to complete full axis correction
• Improved liveness action detection reliability on iOS devices
• Documentation updates and improvements

0.0.4

• Removed google_mlkit_face_mesh_detection dependency (lacks iOS implementation)
• Migrated to use google_mlkit_face_detection exclusively for all face detection features
• Fixed MissingPluginException by using FaceDetector with full feature support
• Enabled contours, landmarks, classification, and tracking in FaceDetector
• Removed face_mesh_utils.dart file (no longer needed)
• Updated liveness_screen.dart to use native Face properties directly
• Updated mobile_face_detector_service.dart to use FaceDetector API
• Simplified code by using built-in headEulerAngleX, headEulerAngleY, smilingProbability, and eye probabilities
• All liveness detection features work without face mesh dependency
• Fixed MLKInvalidImage error on iOS by correcting YUV420 format mapping
• Updated InputImageFormat mapping to use InputImageFormat.yuv420 for iOS instead of nv21
• Added platform-specific handling for iOS YUV420 2-plane format (Y + CbCr)
• Resolves app crash when processing camera images on iOS devices
• Fixed "Could not detect a face in the final image" error on iOS
• Removed final face detection on captured image (ML Kit limitation on iOS front camera)
• Liveness challenges verify face presence in real-time, so final image detection is not needed
• Improved reliability by trusting real-time face detection during challenges

0.0.5

• Migrated iOS implementation to native Apple Vision Framework for full arm64 simulator support.
• Fixed architecture compatibility issues on Apple Silicon (M1/M2/M3/M5) Macs.
• Removed requirement to exclude arm64 architecture in Podfile.
• Increased minimum iOS deployment target to 15.5 for better compatibility with modern ML libraries.
• Standardized coordinate mapping between Vision (iOS) and ML Kit (Android) for consistent liveness detection.
• Updated example app with required versioning and deployment settings for App Store submission.
• Improved overall stability and reduced plugin binary size on iOS.