v1.1.0

🎯 Oriented Bounding Box (OBB) Support

This release adds comprehensive support for oriented bounding boxes as exclusion masks, enabling more precise masking of rotated objects during video stabilization.

✨ New Features

• xywha box format: Define boxes with explicit rotation angles [x_center, y_center, width, height, angle_degrees]
• Smart mask creation: Automatic rotation detection and optimal masking strategy for all box formats
• Enhanced box transformation: Transform boxes between any format combination while preserving rotation information

🔧 Improvements

• More precise exclusion masks for rotated objects (only masks actual object area, not enclosing rectangle)
• Improved feature detection efficiency around rotated objects
• Fast rectangular masking preserved for axis-aligned boxes

📖 Usage Example

import numpy as np

# Define oriented boxes with rotation angles
obb_boxes = np.array([
    [100, 150, 50, 30, 45],   # 45° rotation
    [200, 200, 60, 40, 90],   # 90° rotation
])

stabilizer.set_ref_frame(ref_frame, obb_boxes, box_format='xywha')

🛡️ Compatibility

• ✅ Fully backward compatible - existing code using xywh format works without changes
• ✅ All 26 tests passing (20 existing + 6 new)
• ✅ Zero security vulnerabilities (CodeQL analysis)

📝 Documentation

• Updated README with detailed box format explanations and practical examples
• Added comprehensive test coverage for OBB functionality

Full Changelog: v1.0.1...v1.1.0