# Benchmarking Scene Text Recognition in Devanagari, Telugu and Malayalam

arXiv: https://arxiv.org/abs/2104.04437 · HTML: https://mlsanity.com/paper/2104.04437/

## Key facts

- Cited by 11 papers in the map (data through 14 Oct 2026).
- Builds on 10 of its 28 references that are in the map.
- First posted on arXiv 9 Apr 2021.

## Builds on (most cited first, 10 of 10)

- [Very Deep Convolutional Networks for Large-Scale Image Recognition](https://mlsanity.com/paper/1409.1556/) (2014)
- [Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift](https://mlsanity.com/paper/1502.03167/) (2015)
- [Speech Recognition with Deep Recurrent Neural Networks](https://mlsanity.com/paper/1303.5778/) (2013)
- [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)
- [Synthetic Data for Text Localisation in Natural Images](https://mlsanity.com/paper/1604.06646/) (2016)
- [Virtual Worlds as Proxy for Multi-Object Tracking Analysis](https://mlsanity.com/paper/1605.06457/) (2016)
- [Synthetic Data and Artificial Neural Networks for Natural Scene Text Recognition](https://mlsanity.com/paper/1406.2227/) (2014)
- [Reading Text in the Wild with Convolutional Neural Networks](https://mlsanity.com/paper/1412.1842/) (2014)
- [Generating Synthetic Data for Text Recognition](https://mlsanity.com/paper/1608.04224/) (2016)
- [Unconstrained Scene Text and Video Text Recognition for Arabic Script](https://mlsanity.com/paper/1711.02396/) (2017)

## Built on by (most cited first, 11 of 11)

- [Revisiting Scene Text Recognition: A Data Perspective](https://mlsanity.com/paper/2307.08723/) (2023)
- [Out of Length Text Recognition with Sub-String Matching](https://mlsanity.com/paper/2407.12317/) (2024)
- [Transfer Learning for Scene Text Recognition in Indian Languages](https://mlsanity.com/paper/2201.03180/) (2022)
- [Show Me the World in My Language: Establishing the First Baseline for Scene-Text to Scene-Text Translation](https://mlsanity.com/paper/2308.03024/) (2023)
- [OCR Synthetic Benchmark Dataset for Indic Languages](https://mlsanity.com/paper/2205.02543/) (2022)
- [Bharat Scene Text: A Novel Comprehensive Dataset and Benchmark for Indian Language Scene Text Understanding](https://mlsanity.com/paper/2511.23071/) (2025)
- [Towards Boosting the Accuracy of Non-Latin Scene Text Recognition](https://mlsanity.com/paper/2201.03185/) (2022)
- [IndicSTR12: A Dataset for Indic Scene Text Recognition](https://mlsanity.com/paper/2403.08007/) (2024)
- [MENTOR: Multilingual tExt detectioN TOward leaRning by analogy](https://mlsanity.com/paper/2403.07286/) (2024)
- [Designing Production-Scale OCR for India: Multilingual and Domain-Specific Systems](https://mlsanity.com/paper/2602.16430/) (2026)
- [What Can Low Resource Languages Learn From Each Other?](https://mlsanity.com/paper/2608.27753/) (2026)

## Abstract (arXiv)

Inspired by the success of Deep Learning based approaches to English scene text recognition, we pose and benchmark scene text recognition for three Indic scripts - Devanagari, Telugu and Malayalam. Synthetic word images rendered from Unicode fonts are used for training the recognition system. And the performance is bench-marked on a new IIIT-ILST dataset comprising of hundreds of real scene images containing text in the above mentioned scripts. We use a segmentation free, hybrid but end-to-end trainable CNN-RNN deep neural network for transcribing the word images to the corresponding texts. The cropped word images need not be segmented into the sub-word units and the error is calculated and backpropagated for the the given word image at once. The network is trained using CTC loss, which is proven quite effective for sequence-to-sequence transcription tasks. The CNN layers in the network learn to extract robust feature representations from word images. The sequence of features learnt by the convolutional block is transcribed to a sequence of labels by the RNN+CTC block. The transcription is not bound by word length or a lexicon and is ideal for Indian languages which are highly inflectional. IIIT-ILST dataset, synthetic word images dataset and the script used to render synthetic images are available at http://cvit.iiit.ac.in/research/projects/cvit-projects/iiit-ilst