Similar Handwritten Chinese Character Discrimination by Weakly Supervised Learning

TL;DR

This paper introduces a weakly supervised learning approach for discriminating similar handwritten Chinese characters by localizing discriminative regions and classifying them without fixed-size windows, outperforming existing methods.

Contribution

It formulates similar Chinese character recognition as an SVM-based optimization problem and proposes a novel Gradient Context feature descriptor for improved discrimination.

Findings

Outperforms state-of-the-art methods on CASIA dataset

Effectively localizes discriminative regions without fixed-size windows

Handles high variance in size and position of characters

Abstract

Traditional approaches for handwritten Chinese character recognition suffer in classifying similar characters. In this paper, we propose to discriminate similar handwritten Chinese characters by using weakly supervised learning. Our approach learns a discriminative SVM for each similar pair which simultaneously localizes the discriminative region of similar character and makes the classification. For the first time, similar handwritten Chinese character recognition (SHCCR) is formulated as an optimization problem extended from SVM. We also propose a novel feature descriptor, Gradient Context, and apply bag-of-words model to represent regions with different scales. In our method, we do not need to select a sized-fixed sub-window to differentiate similar characters. The unconstrained property makes our method well adapted to high variance in the size and position of discriminative regions in similar handwritten Chinese characters. We evaluate our proposed approach over the CASIA Chinese character data set and the results show that our method outperforms the state of the art.