# Efficient Domain Adaptation for Text Line Recognition via Decoupled Language Models

arXiv: https://arxiv.org/abs/2603.28028 · HTML: https://mlsanity.com/paper/2603.28028/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 25 of its 48 references that are in the map.
- First posted on arXiv 30 Mar 2026.

## Builds on (most cited first, 25 of 25)

- [Adam: A Method for Stochastic Optimization](https://mlsanity.com/paper/1412.6980/) (2014)
- [Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer](https://mlsanity.com/paper/1910.10683/) (2019)
- [BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehension](https://mlsanity.com/paper/1910.13461/) (2019)
- [Generalized Intersection over Union: A Metric and A Loss for Bounding Box Regression](https://mlsanity.com/paper/1902.09630/) (2019)
- [DINO: DETR with Improved DeNoising Anchor Boxes for End-to-End Object Detection](https://mlsanity.com/paper/2203.03605/) (2022)
- [PEGASUS: Pre-training with Extracted Gap-sentences for Abstractive Summarization](https://mlsanity.com/paper/1912.08777/) (2019)
- [Multilingual Denoising Pre-training for Neural Machine Translation](https://mlsanity.com/paper/2001.08210/) (2020)
- [LayoutLM: Pre-training of Text and Layout for Document Image Understanding](https://mlsanity.com/paper/1912.13318/) (2019)
- [ByT5: Towards a token-free future with pre-trained byte-to-byte models](https://mlsanity.com/paper/2105.13626/) (2021)
- [LayoutLMv3: Pre-training for Document AI with Unified Text and Image Masking](https://mlsanity.com/paper/2204.08387/) (2022)
- [MASS: Masked Sequence to Sequence Pre-training for Language Generation](https://mlsanity.com/paper/1905.02450/) (2019)
- [TrOCR: Transformer-based Optical Character Recognition with Pre-trained Models](https://mlsanity.com/paper/2109.10282/) (2021)
- [ProphetNet: Predicting Future N-gram for Sequence-to-Sequence Pre-training](https://mlsanity.com/paper/2001.04063/) (2020)
- [Towards Accurate Scene Text Recognition with Semantic Reasoning Networks](https://mlsanity.com/paper/2003.12294/) (2020)
- [Chargrid: Towards Understanding 2D Documents](https://mlsanity.com/paper/1809.08799/) (2018)
- [DAN: a Segmentation-free Document Attention Network for Handwritten Document Recognition](https://mlsanity.com/paper/2203.12273/) (2022)
- [Rethinking Text Line Recognition Models](https://mlsanity.com/paper/2104.07787/) (2021)
- [DTrOCR: Decoder-only Transformer for Optical Character Recognition](https://mlsanity.com/paper/2308.15996/) (2023)
- [Recognition of Handwritten Chinese Text by Segmentation: A Segment-annotation-free Approach](https://mlsanity.com/paper/2207.14801/) (2022)
- [HWNet v2: An Efficient Word Image Representation for Handwritten Documents](https://mlsanity.com/paper/1802.06194/) (2018)
- [Neural OCR Post-Hoc Correction of Historical Corpora](https://mlsanity.com/paper/2102.00583/) (2021)
- [Faster DAN: Multi-target Queries with Document Positional Encoding for End-to-end Handwritten Document Recognition](https://mlsanity.com/paper/2301.10593/) (2023)
- [OCR Error Correction Using Character Correction and Feature-Based Word Classification](https://mlsanity.com/paper/1604.06225/) (2016)
- [A Tool for Facilitating OCR Postediting in Historical Documents](https://mlsanity.com/paper/2004.11471/) (2020)
- [General Detection-based Text Line Recognition](https://mlsanity.com/paper/2409.17095/) (2024)

## Abstract (arXiv)

Optical character recognition remains critical infrastructure for document digitization, yet state-of-the-art performance is often restricted to well-resourced institutions by prohibitive computational barriers. End-to-end transformer architectures achieve strong accuracy but demand hundreds of GPU hours for domain adaptation, limiting accessibility for practitioners and digital humanities scholars. We present a modular detection-and-correction framework that achieves near-SOTA accuracy with single-GPU training. Our approach decouples lightweight visual character detection (domain-agnostic) from domain-specific linguistic correction using pretrained sequence models including T5, ByT5, and BART. By training the correctors entirely on synthetic noise, we enable annotation-free domain adaptation without requiring labeled target images. Evaluating across modern clean handwriting, cursive script, and historical documents, we identify a critical "Pareto frontier" in architecture selection: T5-Base excels on modern text with standard vocabulary, whereas ByT5-Base dominates on historical documents by reconstructing archaic spellings at the byte level. Our results demonstrate that this decoupled paradigm matches end-to-end transformer accuracy while reducing compute by approximately 95%, establishing a viable, resource-efficient alternative to monolithic OCR architectures.