Unsupervised Attention Regularization Based Domain Adaptation for Oracle Character Recognition

TL;DR

This paper introduces an unsupervised domain adaptation method for oracle character recognition that enhances model robustness and discriminability by enforcing attention consistency and separability, achieving state-of-the-art results.

Contribution

The novel UARN method improves oracle character recognition by incorporating attention regularization based on visual perceptual plausibility, addressing flip-insensitivity and high inter-class similarity issues.

Findings

UARN outperforms previous methods by 8.5% on Oracle-241 dataset.

Enforcing attention consistency improves model robustness to image flips.

Attention separability enhances class discriminability in recognition tasks.

Abstract

The study of oracle characters plays an important role in Chinese archaeology and philology. However, the difficulty of collecting and annotating real-world scanned oracle characters hinders the development of oracle character recognition. In this paper, we develop a novel unsupervised domain adaptation (UDA) method, i.e., unsupervised attention regularization net?work (UARN), to transfer recognition knowledge from labeled handprinted oracle characters to unlabeled scanned data. First, we experimentally prove that existing UDA methods are not always consistent with human priors and cannot achieve optimal performance on the target domain. For these oracle characters with flip-insensitivity and high inter-class similarity, model interpretations are not flip-consistent and class-separable. To tackle this challenge, we take into consideration visual perceptual plausibility when adapting. Specifically, our method enforces attention consistency between the original and flipped images to achieve the model robustness to flipping. Simultaneously, we constrain attention separability between the pseudo class and the most confusing class to improve the model discriminability. Extensive experiments demonstrate that UARN shows better interpretability and achieves state-of-the-art performance on Oracle-241 dataset, substantially outperforming the previously structure-texture separation network by 8.5%.