Radical Analysis Network for Zero-Shot Learning in Printed Chinese Character Recognition

Abstract

Chinese characters have a huge set of character categories, more than 20,000 and the number is still increasing as more and more novel characters continue being created. However, the enormous characters can be decomposed into a compact set of about 500 fundamental and structural radicals. This paper introduces a novel radical analysis network (RAN) to recognize printed Chinese characters by identifying radicals and analyzing two-dimensional spatial structures among them. The proposed RAN first extracts visual features from input by employing convolutional neural networks as an encoder. Then a decoder based on recurrent neural networks is employed, aiming at generating captions of Chinese characters by detecting radicals and two-dimensional structures through a spatial attention mechanism. The manner of treating a Chinese character as a composition of radicals rather than a single character class largely reduces the size of vocabulary and enables RAN to possess the ability of recognizing unseen Chinese character classes, namely zero-shot learning.

DenseRAN for Offline Handwritten Chinese Character Recognition

“…We adopt a coverage based spatial attention model built in the decoder to detect the radicals and internal two-dimensional structures simultaneously [14], [15]. Compared with [18] focusing on printed Chinese character recognition, DenseRAN focuses on HCCR which is much more difficult due to the diversity of writing styles.…”

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

“…When the training set contains 3255 character classes, TRAN achieves a character accuracy of 60.37% which is a relatively pleasant performance compared with traditional recognition systems as they can not recognize unseen Chinese character classes which means their accuracies are definitely 0%. The performance of recognizing unseen Chinese character classes is not as good as the performance presented in [22] because the handwritten Chinese characters are much more ambiguous compared with printed Chinese characters due to the large handwriting variations.…”

“…For each predicted radical, a coverage based attention model [21] built in the decoder scans the entire input sequence and chooses the most relevant part to describe a segmented radical or a two-dimensional structure between radicals. Our proposed TRAN is related to our previous work [22] with two main differences: 1) [22] focused on the application of RAN on printed Chinese character recognition while this paper focuses on handwritten Chinese character recognition. It is interesting to investigate the performance of RAN on handwritten Chinese character recognition as handwritten characters are much more ambiguous due to the diversity of writing styles.…”

Maximum Entropy Regularization and Chinese Text Recognition

“…Thus character-based recognition model is prone to overfit. Radical-based methods [16,29,30,35] convert Chinese characters to radicals as the basic class to simplify the character structure and decrease the number of classes, thus reducing the demand for training data. They are mentioned here to prove the large demand for data if we use character-based recognition.…”