# Measuring Annotation Efficiency for Handwritten Devanagari Recognition: Sample-Complexity Curves for Four Pretraining Regimes

arXiv: https://arxiv.org/abs/2609.16859 · HTML: https://mlsanity.com/paper/2609.16859/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 9 of its 22 references that are in the map.
- First posted on arXiv 15 Sep 2026.

## Builds on (most cited first, 9 of 9)

- [How transferable are features in deep neural networks?](https://mlsanity.com/paper/1411.1792/) (2014)
- [SimMIM: A Simple Framework for Masked Image Modeling](https://mlsanity.com/paper/2111.09886/) (2021)
- [An End-to-End Trainable Neural Network for Image-based Sequence Recognition and Its Application to Scene Text Recognition](https://mlsanity.com/paper/1507.05717/) (2015)
- [The Shape of Learning Curves: a Review](https://mlsanity.com/paper/2103.10948/) (2021)
- [How to Choose Pretrained Handwriting Recognition Models for Single Writer Fine-Tuning](https://mlsanity.com/paper/2305.02593/) (2023)
- [Few Shots Are All You Need: A Progressive Few Shot Learning Approach for Low Resource Handwritten Text Recognition](https://mlsanity.com/paper/2107.10064/) (2021)
- [Evaluating Synthetic Pre-Training for Handwriting Processing Tasks](https://mlsanity.com/paper/2304.01842/) (2023)
- [One-shot Compositional Data Generation for Low Resource Handwritten Text Recognition](https://mlsanity.com/paper/2105.05300/) (2021)
- [Self-supervised Pre-training of Text Recognizers](https://mlsanity.com/paper/2405.00420/) (2024)

## Abstract (arXiv)

To train handwritten text recognition systems we need word images and their corresponding transcriptions, and these transcriptions are produced manually. For a script that can be read by only a small number of specialists, this manual transcription is a limitation, because the trained models are supposed to save the time of those same specialists. A relevant question therefore arises: how many transcriptions are needed before a recogniser becomes useful, and how much of that cost can pretraining remove? In this study the answer is measured directly for handwritten Devanagari. We keep the recogniser, optimiser and evaluation protocol the same and change only the number of real transcribed words used for fine-tuning across nine budgets from 10 to 4,000 and four initialisation regimes, with six seeds at every point. The resulting curves are then converted into annotation-equivalent terms. A CER of 0.50 is reached by supervised synthetic pretraining using only 81 transcribed words, whereas random initialisation requires 355, which gives a label multiplier of 4.40 [3.56, 4.99]. There is a zero-shot reference point as well: with no real transcribed words at all, this pretraining is worth about 136 of them. This advantage gets smaller as the target accuracy improves, and at the most demanding target we measure, it cannot be distinguished from no saving at all. A fourth arm in which only the encoder is transferred separates the effect of the pretraining method from that of transfer scope, and masked image modelling is observed to transfer negatively over a bounded range of budgets. We emphasise that the scarcity in this study is constructed by subsampling a large corpus.