AttentionHTR: Handwritten Text Recognition Based on Attention Encoder-Decoder Networks

TL;DR

This paper introduces an attention-based encoder-decoder model for handwritten text recognition that leverages transfer learning from scene text models to improve data efficiency, validated on multiple datasets.

Contribution

It presents a novel end-to-end HTR system using transfer learning from scene text models, combining ResNet and bidirectional LSTM with attention mechanisms.

Findings

Effective on multi-writer dataset Imgur5K

Improved recognition accuracy over baseline models

Provides open-source code and pre-trained models

Abstract

This work proposes an attention-based sequence-to-sequence model for handwritten word recognition and explores transfer learning for data-efficient training of HTR systems. To overcome training data scarcity, this work leverages models pre-trained on scene text images as a starting point towards tailoring the handwriting recognition models. ResNet feature extraction and bidirectional LSTM-based sequence modeling stages together form an encoder. The prediction stage consists of a decoder and a content-based attention mechanism. The effectiveness of the proposed end-to-end HTR system has been empirically evaluated on a novel multi-writer dataset Imgur5K and the IAM dataset. The experimental results evaluate the performance of the HTR framework, further supported by an in-depth analysis of the error cases. Source code and pre-trained models are available at https://github.com/dmitrijsk/AttentionHTR.