Stroke Sequence-Dependent Deep Convolutional Neural Network for Online Handwritten Chinese Character Recognition

In this paper, we propose a novel model, named Stroke Sequence-dependent Deep Convolutional Neural Network (SSDCNN), using the stroke sequence information and eight-directional features for Online Handwritten Chinese Character Recognition (OLHCCR). On one hand, SSDCNN can learn the representation of Online Handwritten Chinese Character (OLHCC) by incorporating the natural sequence information of the strokes. On the other hand, SSDCNN can incorporate eight-directional features in a natural way. Firstly, SSDCNN takes the stroke sequence as input by transforming them into stacks of feature maps according to their writing order. And then, the fixed length stroke sequence dependent representations of OLHCC are derived via a series of convolution and max-pooling operations. Thirdly, stroke sequence dependent representation is combined with the eight-directional features via a number of fully connected neural network layers. Finally, the softmax classifier is used as recognizer. In order to train SSDCNN, we divide the process of training into two stages: 1) The training data is used to pre-train the whole architecture until the performance tends to converge. 2) Fully-connected neural network which is used to combine the stroke sequence-dependent representation with eight-directional features and softmax layer are further trained. Experiments were conducted on the OLHCCR competition tasks of ICDAR 2013. Results show that, SSDCNN can reduce the recognition error by 50% ∗ Corresponding author Email addresses: [email protected] (Baotian Hu), [email protected] (Xin Liu), [email protected] (Xiangping Wu), [email protected] (Qingcai Chen) Preprint submitted to Elsevier October 14, 2016 ar X iv :1 61 0. 04 05 7v 1 [ cs .C V ] 1 3 O ct 2 01 6 (5.13% vs 2.56%) compared to the model which only use eight-directional features. The proposed SSDCNN achieves 97.44% accuracy which reduces the recognition error by about 1.9% compared with the best submitted system on ICDAR2013 competition. These results indicate that SSDCNN can exploit the stroke sequence information to learn high-quality representation of OLHCC. It also shows that the learnt representation and the classical eight-directional features complement each other within the SSDCNN architecture.