fast_paddle_ocr 0.0.5
fast_paddle_ocr: ^0.0.5 copied to clipboard
High-performance real-time OCR plugin for Flutter using NCNN and PP-OCRv5. Supports live camera preview, flash control, and photo capture with integrated text recognition.
0.0.5 #
- Massive performance optimization: Decoupled camera rendering from OCR processing to guarantee smooth 60 FPS preview.
- Advanced Dual-Thread Architecture: Separated Detection (DBNet) and Recognition (CRNN) into dedicated background threads.
- Introduced IoU (Intersection over Union) Bounding Box Tracker for ultra-fast 15-20 FPS box responsiveness.
- Extreme battery & thermal efficiency: Throttled the heavy Recognition model to ~3 FPS while maintaining fluid UI.
- Smart Dynamic CPU Allocation: NCNN threads are automatically distributed based on device core count (30:70 ratio).
- Native C++ Text Filtering: Automatically cleans OCR output to strictly alphanumeric characters and dots.
- License updated to Creative Commons Legal Code.
0.0.4 #
- Added detect-only bounding box preview in photo mode (no OCR until capture).
- Full sensor frame capture: saved images now use the complete camera resolution.
- OCR processing limited to overlay/scan area only for accurate text detection.
- Bounding boxes and text labels drawn on full-frame output image.
- Added full-screen image viewer with pinch-to-zoom in example app.
- Fixed
ocrFromImageto save annotated image as separate file (no overwrite).
0.0.3 #
- Added
abiFiltersdocumentation to significantly reduce Android APK size. - Added Platform Support table to README.md.
- Updated homepage and repository links in pubspec.yaml.
0.0.2 #
- Added flash (torch) control and camera toggle features.
- Optimized camera polling mechanism for real-time OCR.
- Improved Android NDK camera stability.
0.0.1 #
- Initial release of the Flutter Paddle OCR plugin.
- Integrated C++ NCNN backend for offline OCR processing.
- Added real-time text scanning and high-resolution photo capture functionality.