Paper Overview
Field: Computer Vision Authors: Yonatan Haile Medhanie, Yuanhua Ni Published: 2026-04-22 arXiv: 2604.20813
Summary
Transformer-based OCR models have shown strong performance on Latin and CJK scripts, but their application to African syllabic writing systems remains limited. This work presents the first adaptation of TrOCR for printed Tigrinya using the Ge'ez script.
Approach
- Starting from a pre-trained model, the authors extend the byte-level BPE tokenizer to cover 230 Ge'ez characters.
- They introduce Word-Aware Loss Weighting to resolve systematic word-boundary failures that arise when applying Latin-centric BPE conventions to a new script.
- The unmodified model produces no usable output on Ge'ez text.
- After adaptation, the TrOCR-Printed variant achieves 0.22% Character Error Rate (CER) and 97.20% exact match accuracy on a held-out test set of 5,000 synthetic images from the GLOCR dataset.
- An ablation study confirms that Word-Aware Loss Weighting is a critical component, reducing CER by two orders of magnitude compared to vocabulary expansion alone.
- The full pipeline trains in under three hours on a single 8GB consumer GPU.
- Paper: https://arxiv.org/abs/2604.20813
Results
All code, model weights, and evaluation scripts have been publicly released.