DropRegion Training of Inception Font Network for High-Performance Chinese Font Recognition

TL;DR

This paper introduces DropRegion, a data augmentation method, combined with an enhanced Inception font network, to improve Chinese font recognition accuracy despite limited labeled samples and complex character structures.

Contribution

The paper proposes DropRegion for generating diverse training samples and an improved Inception font network with CCCP and global average pooling for better CFR performance.

Findings

DropRegion effectively increases training data variability.

The enhanced IFN achieves higher recognition accuracy.

Experimental results confirm the approach's effectiveness.

Abstract

Chinese font recognition (CFR) has gained significant attention in recent years. However, due to the sparsity of labeled font samples and the structural complexity of Chinese characters, CFR is still a challenging task. In this paper, a DropRegion method is proposed to generate a large number of stochastic variant font samples whose local regions are selectively disrupted and an inception font network (IFN) with two additional convolutional neural network (CNN) structure elements, i.e., a cascaded cross-channel parametric pooling (CCCP) and global average pooling, is designed. Because the distribution of strokes in a font image is non-stationary, an elastic meshing technique that adaptively constructs a set of local regions with equalized information is developed. Thus, DropRegion is seamlessly embedded in the IFN, which enables end-to-end training; the proposed DropRegion-IFN can be used for high performance CFR. Experimental results have confirmed the effectiveness of our new approach for CFR.