Deep learning approach for Chinese handwritten character recognition

Handwritten Chinese character recognition has been a long-standing problem in Computer Vision, in part due to the complex structure and large character set of the Chinese Language. Thanks to the rise of deep learning, not only has the state of the art improved considerably in this area, but also the field is drawing in users from many fields. In this survey, Chinese handwritten character recognition (HCCR) methods are reviewed based on deep learning methodologies, providing a comprehensive view of recent developments in Chinese HCCR. We survey a large number of techniques, systems and data sets, challenges and future research needs to improve recognition accuracy and efficiency.