Have convolutions already made recurrence obsolete for unconstrained handwritten text recognition ?

TL;DR

This paper investigates whether convolutional neural networks can replace recurrent networks in unconstrained handwritten text recognition, comparing their performance and training efficiency on standard and augmented datasets.

Contribution

It provides an experimental comparison between convolutional architectures and traditional CNN+BLSTM models for handwritten text recognition, including analysis on augmented data.

Findings

Convolutional architectures achieve competitive accuracy with recurrent models.

Convolutional models train faster and support better parallelism.

Augmented data with printed grids impacts recognition performance.

Abstract

Unconstrained handwritten text recognition remains an important challenge for deep neural networks. These last years, recurrent networks and more specifically Long Short-Term Memory networks have achieved state-of-the-art performance in this field. Nevertheless, they are made of a large number of trainable parameters and training recurrent neural networks does not support parallelism. This has a direct influence on the training time of such architectures, with also a direct consequence on the time required to explore various architectures. Recently, recurrence-free architectures such as Fully Convolutional Networks with gated mechanisms have been proposed as one possible alternative achieving competitive results. In this paper, we explore convolutional architectures and compare them to a CNN+BLSTM baseline. We propose an experimental study regarding different architectures on an offline handwriting recognition task using the RIMES dataset, and a modified version of it that consists of augmenting the images with notebook backgrounds that are printed grids.