From Press to Pixels: Evolving Urdu Text Recognition

TL;DR

This paper compares traditional OCR and LLMs for Urdu newspaper text recognition, introducing a new dataset and showing LLMs' effectiveness with limited data, while addressing layout and image quality challenges.

Contribution

It introduces the Urdu Newspaper Benchmark (UNB) dataset and demonstrates the effectiveness of fine-tuned LLMs for Urdu OCR in low-resource settings.

Findings

Gemini-2.5-Pro achieves WER 0.133 on UNB

Fine-tuning GPT-4o improves WER by 6.13% with limited samples

Super-resolution enhances OCR accuracy by 50%

Abstract

This paper presents a comparative analysis of Large Language Models (LLMs) and traditional Optical Character Recognition (OCR) systems on Urdu newspapers, addressing challenges posed by complex multi-column layouts, low-resolution scans, and the stylistic variability of the Nastaliq script. To handle these challenges, we fine-tune YOLOv11x models for article- and column-level text block extraction and train a SwinIR-based super-resolution module that enhances image quality for downstream text recognition, improving accuracy by an average of 50%. We further introduce the Urdu Newspaper Benchmark (UNB), a manually annotated dataset for Urdu OCR comprising 829 paragraph images with a total of 9,982 sentences. Using UNB and the OpenITI corpus, we conduct a systematic comparison between traditional CNN+RNN-based OCR systems and modern LLMs, presenting detailed insertion, deletion, and substitution error analyses alongside character-level confusion patterns. We find that Gemini-2.5-Pro achieves the best performance on UNB (WER 0.133), while fine-tuning GPT-4o on just 500 in-domain samples yields a 6.13% absolute WER improvement, demonstrating the adaptability of LLMs to low-resource, morphologically complex scripts like Urdu. The UNB dataset and fine-tuned models are publicly available at https://github.com/sameearif/urdu-newspaper-benchmark.