Handwriting Recognition with Large Multidimensional Long Short-Term Memory Recurrent Neural Networks

  title={Handwriting Recognition with Large Multidimensional Long Short-Term Memory Recurrent Neural Networks},
  author={Paul Voigtlaender and Patrick Doetsch and Hermann Ney},
  journal={2016 15th International Conference on Frontiers in Handwriting Recognition (ICFHR)},
Multidimensional long short-term memory recurrent neural networks achieve impressive results for handwriting recognition. However, with current CPU-based implementations, their training is very expensive and thus their capacity has so far been limited. We release an efficient GPU-based implementation which greatly reduces training times by processing the input in a diagonal-wise fashion. We use this implementation to explore deeper and wider architectures than previously used for handwriting… CONTINUE READING
Highly Cited
This paper has 26 citations. REVIEW CITATIONS

9 Figures & Tables

Extracted Numerical Results

  • We used a smoothed word-based trigram language model with a perplexity of 420 on the training set and an out of vocabulary rate of 4% on the development set.
  • During the training, we measure the CTC objective function value and the label error rate, i.e. the lowest character error rate of the network itself without lexicon or language model, on a holdout set of 10% of the training data.



Citations per Year

Citation Velocity: 15

Averaging 15 citations per year over the last 2 years.

Learn more about how we calculate this metric in our FAQ.