Recurrence-free unconstrained handwritten text recognition using gated fully convolutional network

TL;DR

This paper introduces a recurrence-free, fully convolutional neural network architecture for unconstrained handwritten text recognition, offering a parallelizable alternative to traditional CNN+LSTM models with competitive accuracy.

Contribution

The authors propose a novel gated fully convolutional network architecture that eliminates the need for recurrent layers in handwritten text recognition.

Findings

Achieves competitive results on RIMES and IAM datasets.

Reduces model complexity and training time compared to LSTM-based models.

Provides open-source code for reproducibility.

Abstract

Unconstrained handwritten text recognition is a major step in most document analysis tasks. This is generally processed by deep recurrent neural networks and more specifically with the use of Long Short-Term Memory cells. The main drawbacks of these components are the large number of parameters involved and their sequential execution during training and prediction. One alternative solution to using LSTM cells is to compensate the long time memory loss with an heavy use of convolutional layers whose operations can be executed in parallel and which imply fewer parameters. In this paper we present a Gated Fully Convolutional Network architecture that is a recurrence-free alternative to the well-known CNN+LSTM architectures. Our model is trained with the CTC loss and shows competitive results on both the RIMES and IAM datasets. We release all code to enable reproduction of our experiments: https://github.com/FactoDeepLearning/LinePytorchOCR.