About this document

Unconstrained Handwritten Word Recognition Based On Trigrams Using BLSTM by Trung Thành Lê is a document available to read on EtoBox.

This paper presents a method for improving unconstrained handwritten word recognition by utilizing a hybrid approach that combines trigrams and Bidirectional Long Short-Term Memory (BLSTM) networks. The authors propose a novel token passing algorithm that leverages the outputs of two separate recognizers trained on distinct subsets of training data based on trigrams, addressing the challenges posed by insufficient training data and uneven word distribution. The approach aims to enhance recognition accuracy,

Author

Trung Thành Lê

Language

EN