Stroke-Based Autoencoders: Self-Supervised Learners for Efficient Zero-Shot Chinese Character Recognition

TL;DR

This paper introduces a stroke-based autoencoder that models Chinese character morphology using self-supervised learning, enabling effective zero-shot recognition and improved character representations by capturing detailed stroke information.

Contribution

The paper presents a novel stroke-based autoencoder architecture trained to reconstruct stroke sequences, enhancing zero-shot recognition and morphological understanding of Chinese characters.

Findings

Outperforms existing methods in zero-shot recognition accuracy.

Enriches Chinese word embeddings with morphological features.

Effective in recognizing unseen characters based on stroke information.

Abstract

Chinese characters carry a wealth of morphological and semantic information; therefore, the semantic enhancement of the morphology of Chinese characters has drawn significant attention. The previous methods were intended to directly extract information from a whole Chinese character image, which usually cannot capture both global and local information simultaneously. In this paper, we develop a stroke-based autoencoder(SAE), to model the sophisticated morphology of Chinese characters with the self-supervised method. Following its canonical writing order, we first represent a Chinese character as a series of stroke images with a fixed writing order, and then our SAE model is trained to reconstruct this stroke image sequence. This pre-trained SAE model can predict the stroke image series for unseen characters, as long as their strokes or radicals appeared in the training set. We have designed two contrasting SAE architectures on different forms of stroke images. One is fine-tuned on existing stroke-based method for zero-shot recognition of handwritten Chinese characters, and the other is applied to enrich the Chinese word embeddings from their morphological features. The experimental results validate that after pre-training, our SAE architecture outperforms other existing methods in zero-shot recognition and enhances the representation of Chinese characters with their abundant morphological and semantic information.