[PR #1925] feat(ocr): Add character blacklist and line-split options for better accuracy #2729

New Issue

claunia · 2026-01-29T17:23:37Z

claunia commented

2026-01-29 17:23:37 +00:00

Original Pull Request: https://github.com/CCExtractor/ccextractor/pull/1925

State: closed
Merged: Yes

Summary

Add OCR character blacklist (enabled by default) to prevent common misrecognition errors like I → |
Add optional --ocr-line-split mode for multi-line subtitle images
Inspired by subtile-ocr's proven approach

New Options

Option	Default	Description
`--no-ocr-blacklist`	Blacklist ON	Disable the character blacklist (`\|`, `\`, `, `_`, `~`)
`--ocr-line-split`	OFF	Split images into lines, use PSM 7 for each

Test Results (VOBSUB MKV sample)

Metric	Before	After (blacklist)
Pipe `\|` errors	14	0

The blacklist completely eliminates pipe character misrecognition, matching subtile-ocr's accuracy.

Test plan

Build with OCR support
Test VOBSUB extraction with blacklist (default)
Test with --no-ocr-blacklist to verify it can be disabled
Test --ocr-line-split option
Verify help text displays correctly

🤖 Generated with Claude Code

**Original Pull Request:** https://github.com/CCExtractor/ccextractor/pull/1925 **State:** closed **Merged:** Yes --- ## Summary - Add OCR character blacklist (enabled by default) to prevent common misrecognition errors like `I` → `|` - Add optional `--ocr-line-split` mode for multi-line subtitle images - Inspired by subtile-ocr's proven approach ## New Options | Option | Default | Description | |--------|---------|-------------| | `--no-ocr-blacklist` | Blacklist ON | Disable the character blacklist (`\|`, `\`, `` ` ``, `_`, `~`) | | `--ocr-line-split` | OFF | Split images into lines, use PSM 7 for each | ## Test Results (VOBSUB MKV sample) | Metric | Before | After (blacklist) | |--------|--------|-------------------| | Pipe `\|` errors | 14 | **0** | The blacklist completely eliminates pipe character misrecognition, matching subtile-ocr's accuracy. ## Test plan - [x] Build with OCR support - [x] Test VOBSUB extraction with blacklist (default) - [x] Test with `--no-ocr-blacklist` to verify it can be disabled - [x] Test `--ocr-line-split` option - [x] Verify help text displays correctly 🤖 Generated with [Claude Code](https://claude.com/claude-code)

claunia added the pull-request label 2026-01-29 17:23:37 +00:00

claunia closed this issue

2026-01-29 17:23:38 +00:00

Sign in to join this conversation.

1 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: starred/ccextractor#2729