21 June 2013 New baseline correction algorithm for text-line recognition with bidirectional recurrent neural networks
Author Affiliations +
J. of Electronic Imaging, 22(2), 023028 (2013). doi:10.1117/1.JEI.22.2.023028
Abstract
Many preprocessing techniques have been proposed for isolated word recognition. However, recently, recognition systems have dealt with text blocks and their compound text lines. In this paper, we propose a new preprocessing approach to efficiently correct baseline skew and fluctuations. Our approach is based on a sliding window within which the vertical position of the baseline is estimated. Segmentation of text lines into subparts is, thus, avoided. Experiments conducted on a large publicly available database (Rimes), with a BLSTM (bidirectional long short-term memory) recurrent neural network recognition system, show that our baseline correction approach highly improves performance.
© 2013 SPIE and IS&T
Olivier Morillot, Laurence Likforman-Sulem, Emmanuèle Grosicki, "New baseline correction algorithm for text-line recognition with bidirectional recurrent neural networks," Journal of Electronic Imaging 22(2), 023028 (21 June 2013). https://doi.org/10.1117/1.JEI.22.2.023028
JOURNAL ARTICLE
12 PAGES


SHARE
Back to Top