jedarden
|
8d6a1a07df
|
docs(pdftract-372e): finalize watermark and background separation research note v1.0
- Added Section 2: Combined Watermark Scoring Algorithm with signal definitions, pseudo-code, threshold tuning, and weight overrides
- Added Section 4: Font-Based Signals (font size, color, weight/family)
- Added Section 11: Text Output Mode behavior (pre/post Phase 7)
- Added Section 12: Edge Cases (stamps vs watermarks, raster watermarks, form profile override, reading-order interaction)
- Added Section 13: Validation Corpus with empirical baseline results
- Expanded Section 10 with WatermarkSignals struct containing individual signal scores
- File grows from 198 to 546 lines
Closes: pdftract-372e
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
|
2026-05-24 10:33:37 -04:00 |
|