Abstract:Although great success has been achieved in online handwritten Chinese text recognition (OLHCTR), most existing methods based on over-segmentation or long short-term memory are inefficient and not parallelizable. Moreover, n-gram language models and beam search algorithm were commonly adopted by many existing systems as a part of post-processing, resulting in extremely low speed and large footprint. To this end, we propose a fast, accurate and compact approach for OLHCTR. The proposed method consists of a global and local relationship network (GLRNet) and a Transformer-based language model (TransLM). A novel feature extraction mechanism, which alternately learns global and local dependencies of input trajectories, is proposed in GLRNet for the recognition of online texts. Based on the output of GLRNet, TransLM captures contextual information through Transformer encoder and further improves the recognition accuracy. The recognition and language modelling are always treated as two separate parts. However, the two components of our methods are jointly optimized, which ensures the optimal performance of the whole model. Furthermore, the non-recurrence design improves the parallelization and efficiency of our method, and the parameterized TransLM avoids the large footprint to store the probabilities of n-grams. The experiments on CASIA-OLHWDB2.0-2.2 and ICDAR2013 competition dataset show that our method achieves state-of-the-art performances with the fastest speed and the smallest footprint. Especially in the situation with language model, our method exhibits 2 times to 130 times acceleration compared with existing methods.

Online Handwritten Mongolian Character Recognition Using CMA-MOHR and Coordinate Processing

Online Handwritten Mongolian Word Recognition Using MWRCNN and Position Maps

Online Mongolian Handwriting Recognition Based on Encoder–Decoder Structure with Language Model

Online Handwriting Mongolia Words Recognition Based on Multiple Classifiers

End-to-End Model for Offline Handwritten Mongolian Word Recognition

Sub-Word Based Mongolian Offline Handwriting Recognition

Online Handwriting Mongolia Words Recognition Based on HMM Classifier

End-to-End Model Based on Bidirectional LSTM and CTC for Online Handwritten Mongolian Word Recognition

Online Handwritten Mongolian Word Recognition Using a Novel Sliding Window Method with Recurrent Neural Networks.

A new dataset for mongolian online handwritten recognition

Online Handwriting Mongolia Words Recognition with Recurrent Neural Networks

A New Database for Online Handwritten Mongolian Word Recognition.

Hmmrf: A Stochastic Model For Offline Handwritten Chinese Character Recognition

Segmentation-Free Printed Traditional Mongolian Ocr Using Sequence To Sequence With Attention Model

DNN-HMM for Large Vocabulary Mongolian Offline Handwriting Recognition

Towards Fast, Accurate and Compact Online Handwritten Chinese Text Recognition

End-to-End Model Based on Bidirectional LSTM and CTC for Segmentation-free Traditional Mongolian Recognition

Multi-Task Learning Based Traditional Mongolian Words Recognition.

Grapheme Segmentation Based Mongolian Handwriting Recognition

Hidden Markov random field based approach for off-line handwritten Chinese character recognition

Offline Mongolian Handwriting Recognition Based on Data Augmentation and Improved ECA-Net