Trajectory-based Radical Analysis Network for Online Handwritten Chinese Character Recognition

@article{Zhang2018TrajectorybasedRA,
  title={Trajectory-based Radical Analysis Network for Online Handwritten Chinese Character Recognition},
  author={Jian-shu Zhang and Yixing Zhu and Jun Du and Lirong Dai},
  journal={2018 24th International Conference on Pattern Recognition (ICPR)},
  year={2018},
  pages={3681-3686}
}
Recently, great progress has been made for online handwritten Chinese character recognition due to the emergence of deep learning techniques. However, previous research mostly treated each Chinese character as one class without explicitly considering its inherent structure, namely the radical components with complicated geometry. In this study, we propose a novel trajectory-based radical analysis network (TRAN) to firstly identify radicals and analyze two-dimensional structures among radicals… 
View With Semantic Reader
8 Citations
Background Citations
4
Methods Citations
1

Figures and Tables from this paper

8 Citations

Citation Type

Citation Type

Radical analysis network for learning hierarchies of Chinese characters
Multi-Scale Attention with Dense Encoder for Handwritten Mathematical Expression Recognition
TLDR
This study improves the encoder by employing densely connected convolutional networks as they can strengthen feature extraction and facilitate gradient propagation especially on a small training set and presents a novel multi-scale attention model which is employed to deal with the recognition of math symbols in different scales and restore the fine-grained details dropped by pooling operations.
Track, Attend, and Parse (TAP): An End-to-End Framework for Online Handwritten Mathematical Expression Recognition
TLDR
The proposed end-to-end framework does not require the explicit symbol segmentation and a predefined expression grammar for parsing and demonstrates the strong complementarity between offline information with static-image input and online information with ink-trajectory input by blending a fully convolutional networks-based watcher into TAP.
A Comparative Study of Attention-Based Encoder-Decoder Approaches to Natural Scene Text Recognition
TLDR
The results show that attention on a two-dimensional feature map can yield better performance than one-dimensional one and an RNN based decoder performs better than CNN based one and it is more effective and efficient for CNN-DBLSTM-CTC based approaches to leverage an explicit language model to boost recognition accuracy.
Distinguishing Oracle Variants Based on the Isomorphism and Symmetry Invariances of Oracle-Bone Inscriptions
TLDR
The method proposed in this paper has achieved great results in the recognition of oracle variant characters with a priori knowledge introduced to integrate multi-domain methods to identify oracle variants.
Image Purification Technique for Myanmar OCR Applying Skew Angle Detection and Free Skew
TLDR
The ability of the proposed method to effectively implement global and local skew angle detection, free skew and line segmentation in different handwritten and digital text images of the Myanmar character set with high accuracies confirms the robustness of the technique, its reliability and its suitability for application in many other related languages.
Discrete Memory Addressing Variational Autoencoder for Visual Concept Learning
TLDR
A discrete memory addressing VAE model is proposed for explicitly memorizing and reasoning about shared primitives in images and derives a proper variational lower bound which can be optimized efficiently.
Cross-Modal Prototype Learning for Zero-Shot Handwritten Character Recognition

References

SHOWING 1-10 OF 34 REFERENCES
Radical Analysis Network for Zero-Shot Learning in Printed Chinese Character Recognition
TLDR
A novel radical analysis network (RAN) to recognize printed Chinese characters by identifying radicals and analyzing two-dimensional spatial structures among them by employing convolutional neural networks as an encoder and decoder.
RAN: Radical analysis networks for zero-shot learning of Chinese characters
TLDR
The Radical Analysis Networks (RAN) that recognize Chinese characters by identifying radicals and analyzing 2D spatial structures between them are introduced and test a simple implementation on experiments of recognizing printed Chinese characters with seen and unseen classes.
Drawing and Recognizing Chinese Characters with Recurrent Neural Network
TLDR
Under the RNN framework, a conditional generative model with character embedding is proposed for automatically drawing recognizable Chinese characters and can be recognized by the discriminative RNN model with high accuracy.
Radical-Based Chinese Character Recognition via Multi-Labeled Learning of Deep Residual Networks
TLDR
The radical-level composition of Chinese characters is considered, and a deep residual network with multi-labeled learning is proposed to detect position-dependent radicals, which enables the recognition of novel characters without training samples if the characters are composed of radicals appearing in training samples.
A GRU-Based Encoder-Decoder Approach with Attention for Online Handwritten Mathematical Expression Recognition
TLDR
A novel end-to-end approach based on the encoder-decoder framework with the attention mechanism for online handwritten mathematical expression recognition (OHMER) that can simultaneously accomplish the symbol recognition and structural analysis to output a character sequence in LaTeX format is presented.
Multi-Scale Attention with Dense Encoder for Handwritten Mathematical Expression Recognition
TLDR
This study improves the encoder by employing densely connected convolutional networks as they can strengthen feature extraction and facilitate gradient propagation especially on a small training set and presents a novel multi-scale attention model which is employed to deal with the recognition of math symbols in different scales and restore the fine-grained details dropped by pooling operations.
High performance offline handwritten Chinese character recognition using GoogLeNet and directional feature maps
TLDR
It has been shown that with the proper incorporation with traditional directional feature maps, the proposed single and ensemble HCCR-GoogLeNet models achieve new state of the art recognition accuracy of 96.35% and 96.74%, respectively, outperforming previous best result with significant gap.
Optical recognition of handwritten Chinese characters by hierarchical radical matching method
'Online recognition of Chinese characters: the state-of-the-art
TLDR
This paper reviews the advances in online Chinese character recognition (OLCCR), with emphasis on the research works from the 1990s, in terms of pattern representation, character classification, learning/adaptation, and contextual processing.
DropSample: A New Training Method to Enhance Deep Convolutional Neural Networks for Large-Scale Unconstrained Handwritten Chinese Character Recognition
...
...