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

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.

References in corpus (9)

• Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
• Convolutional Sequence to Sequence Learning
• Language Modeling with Gated Convolutional Networks
• Highway Networks
• Layer Normalization
• Quasi-Recurrent Neural Networks
• Batch Renormalization: Towards Reducing Minibatch Dependence in Batch-Normalized Models
• Accurate, Data-Efficient, Unconstrained Text Recognition with Convolutional Neural Networks
• CTCModel: a Keras Model for Connectionist Temporal Classification

Cited by in corpus (7)

• End-to-end Handwritten Paragraph Text Recognition Using a Vertical Attention Network
• DAN: a Segmentation-free Document Attention Network for Handwritten Document Recognition
• Character-Based Handwritten Text Transcription with Attention Networks
• Recurrence-free unconstrained handwritten text recognition using gated fully convolutional network
• Faster DAN: Multi-target Queries with Document Positional Encoding for End-to-end Handwritten Document Recognition
• DANIEL: A fast Document Attention Network for Information Extraction and Labelling of handwritten documents
• EASTER: Efficient and Scalable Text Recognizer