proveface 1.0.2

Real-time face liveness detection for Flutter. Blink, smile, and head-turn challenges with passive anti-spoofing. Camera permission handled internally.

1.0.2

Bug Fixes

• Fixed fatal crash on Android when a camera frame was processed concurrently with takePicture(), causing CameraDevice was already closed exception.
• Fixed CameraException: buildPreview() was called on a disposed CameraController when session timed out.
• Fixed post-capture validation always failing on Android due to unsupported BGRA8888 format; Android now converts captured frame to NV21 for ML Kit.
• Fixed RangeError in NV21 conversion when decoded image has an odd height (e.g. portrait photos decoded at 640 px wide).

Improvements

• Android camera processing extracted into a dedicated subclass for cleaner platform separation.
• Challenge countdown timer is now always visible once a challenge starts, regardless of transient face-quality interruptions.
• Face swap mid-challenge is now detected and triggers an immediate restart.
• Face loss after completing one or more challenges triggers an instant restart instead of waiting for a grace period.
• Post-capture image validation now detects face obstructions (hand, mask, etc.) consistently on both platforms by processing at stream-equivalent resolution.
• When all challenge checks are disabled in config, the widget proceeds directly to capture without requiring any liveness checks.

1.0.1

• Fixed: OSI-approved MIT license added for pub.dev recognition.
• Fixed: Static analysis lint rules configured in analysis_options.yaml.

1.0.0

• Initial release.
• Blink, smile, turn-left, turn-right, turn-up, turn-down liveness challenges.
• Randomised challenge ordering on every session.
• Passive anti-spoofing via face_anti_spoofing_detector ML model.
• Screen and video replay detection via pixel texture analysis.
• Face quality checks: distance, centering, lighting, sunglasses, mask.
• Camera permission requested automatically — no setup needed in consuming app.
• Permission denied screen with "Open Settings" shortcut.
• Typed ProveFaceResult with base64 JPEG on success.
• Configurable per-challenge timeouts via ProveFaceConfig.
• appBarColor and title customisable from call site.
• Supports Android API 23+ and iOS 12.0+.