Post-OCR Text Cleanup for Research Corpora

Instructions

1. Cleanup Strategy Selection

• Characterize the error-generating DGP before selecting a method. Document source language(s), era, typeface family (Fraktur, Antiqua, typewritten, handwritten), scan DPI, OCR engine, and domain jargon. Each parameter constrains which corrections are plausible and which risk introducing semantic drift.
• Choose between LLM correction, rule-based fixes, or a hybrid pipeline based on error type. LLM correction excels at context-dependent errors (wrong but plausible characters, broken words, missing diacritics). Rule-based fixes handle deterministic patterns (control characters, Unicode normalization, repetition artifacts, whitespace) with zero risk of content alteration. Use rule-based fixes unconditionally for these categories.
• Default to the hybrid approach for research corpora. Run LLM correction first on all pages, then apply deterministic rule fixes on top. This order matters: LLM correction may introduce formatting artifacts that rule fixes clean up, while the reverse order wastes rule-fix effort on text the LLM will rewrite (Machidon & Machidon 2025).
• Pilot-test LLM correction per language before corpus-wide deployment. LLM post-correction effectiveness is highly language-dependent: English achieves 7-58% CER reduction across open models, while Finnish shows negative or near-zero improvement across the same model set (Kanerva et al. 2025). Never assume cross-language transferability.
• Consider whether correction is needed at all. Define the quality threshold before choosing a strategy. The Hill & Hengchen 70-80% quality band (reported in van Strien et al. 2020) marks the critical threshold below which most downstream NLP tasks perform poorly; above 80% quality many tasks (e.g., topic modeling) tolerate residual noise. If the downstream analysis sits comfortably above this band, the risk of correction-introduced errors may outweigh the benefit.

2. LLM-Based Correction

• For most pages, use a small text-only model. The correction input is already text; image understanding is not needed for well-OCR'd pages. A 7-13B parameter model with 4-bit quantization fits in ~4-20GB VRAM and runs on a single GPU. Larger fp16 models (e.g., Llama-3.1-70B at fp16 yielding ~42% CER reduction vs ~39% at 4-bit) gain 2.5-4.7pp but require roughly 3x the memory (132GB vs 43GB) and often a second GPU (Kanerva et al. 2025).
• For severely degraded pages, use multimodal correction. Feeding both the original page image and the OCR text to a correction model can achieve below 1% CER on degraded documents, but doubles GPU cost (Greif et al. 2025). Reserve this for flagged pages, not routine processing.
• Write tight correction prompts. Instruct the model to "fix clear OCR mistakes only: wrong characters, broken words, garbled punctuation, repetition artifacts. Do not translate, modernize, or add anything. Output the corrected text only." Loose prompts invite hallucination.
• Supply socio-cultural context in the prompt. Including document era, publication type, language register, and genre (e.g., "The text is from an English newspaper in the 1800s") meaningfully reduces CER beyond generic correction prompts — the top-performing CLOCR-C configuration achieved over 60% CER reduction on the NCSE dataset using a modular prompt that combines expert framing, recovery instructions, publication context, text-type context, and anti-overgeneration instructions (Bourne 2024). Misleading or mismatched context degrades performance, so use the real document metadata.
• Add language-specific instructions. For Polish, explicitly mention diacritics restoration (ą, ć, ę, ł, ń, ó, ś, ź, ż). For Korean, mention hangul integrity and hanja preservation. The correction model needs to know which character set to favor.
• Mitigate hallucination with constrained decoding. Constrained decoding techniques — beam search with CER-based re-ranking, sequence-level similarity re-ranking, and token-level Constrained Beam Search that interpolates the model's distribution with a character-similarity distribution — enforce fidelity between input and output and prevent plausible-but-fabricated substitutions (Sastre et al. 2025). Prefer token-level CBS with dynamic α if model logits are accessible; otherwise fall back to beam search with CER re-selection. This matters because WER can worsen even when CER improves: fine-tuning alone in Sastre et al. left CER roughly flat (0.314→0.321) while WER jumped from 0.633 to 0.821, a failure mode constrained decoding directly addresses.
• Use worked prompt templates and a provenance schema. See reference/prompt-templates-and-schema.md for a minimal constrained-decoding-friendly baseline prompt, a Bourne-style socio-cultural-context prompt, and a span-level JSONL provenance schema (per Guo & Wei 2026 §3.2/§3.3).
• Strip LLM overgeneration with alignment-based post-processing. Llama-family models routinely prepend "Here is the corrected text:" or append error-by-error explanations. Without post-hoc trimming (character-level local alignment of output against input, keeping only the aligned region), Llama-3-8B scored -74.1% CER; with trimming, +7.3% (Kanerva et al. 2025). Gemma and GPT-4o are largely unaffected but the step is cheap and should be applied universally.
• Disable chain-of-thought for correction tasks. Reasoning modes add latency without improving transcription fidelity. Use low-temperature sampling or greedy decoding for deterministic output.
• Tune segment length for corpus-scale processing. Short segments (50-100 words) score notably worse CER% across models; 200-300 words appears optimal for page-level correction (Kanerva et al. 2025). When splitting long documents, use a stride that preserves left context (left-uncorrected-concatenate parallelizes cleanly; left-corrected-concatenate is sequential but slightly better at segment boundaries).
• Track all changes with edit-distance metrics. Compute Levenshtein distance and change ratio (edit distance / original length) per page. Flag pages where the correction model altered more than 10% of characters for manual review — high change ratios may indicate hallucination rather than correction. This 10% threshold is an operational heuristic; calibrate against your pilot evaluation.

3. Rule-Based Fixes

• Apply deterministic fixes in a fixed order. (1) Control character removal, (2) zero-width and invisible Unicode character removal, (3) NFKC Unicode normalization, (4) consecutive character repetition collapse, (5) standalone symbol line removal, (6) whitespace normalization. This ordering prevents interactions between fixes.
• Tune repetition collapse thresholds to the corpus. The default of collapsing runs of 4+ identical characters to 3 works for most scripts but may need adjustment for languages with legitimate long character sequences or for documents with intentional formatting patterns.
• Rule-based diacritics restoration is viable for some languages. For Polish, rule-based approaches (removing word breaks, rejecting case-changing corrections, restoring diacritical characters replaced with visually similar ASCII) are competitive with LLM-based correction and more predictable (Ogrodniczuk 2022).
• Generate synthetic OCR errors for training when ground truth is scarce. Glyph-similarity-based synthetic corruption (feature-matched character confusions) produces more realistic training data than random-injection baselines, and outperforms in low-resource languages (Guan & Greene 2024).
• Preserve the raw text alongside every cleaned version. Rule-based fixes are deterministic and reversible, but downstream researchers may prefer different normalization choices. Store both raw and cleaned text at every stage.

4. Quality Diagnostics and Metrics

• Move beyond CER/WER as the sole quality measure. Character-level