Keyword spotting for self-training of BLSTM NN based handwriting recognition systems

作者：

Highlights：

• Self-training for BLSTM NN Recognizers for handwritten text.

• The proposed approach is magnitudes faster than the reference system.

• This approach reduces the need for labeled training data.

摘要

Highlights•Self-training for BLSTM NN Recognizers for handwritten text.•The proposed approach is magnitudes faster than the reference system.•This approach reduces the need for labeled training data.

论文关键词：Document retrieval,Keyword spotting,Handwriting recognition,Neural networks,Semi-supervised learning

论文评审过程：Available online 10 July 2013.

论文官网地址：https://doi.org/10.1016/j.patcog.2013.06.030