Handwritten Chinese character recognition by joint classification and similarity ranking

CASIA OpenIR > 多模态人工智能系统全国重点实验室 > 模式分析与学习

	Handwritten Chinese character recognition by joint classification and similarity ranking
	Cheng Cheng 1; Xu-Yao Zhang2 ; Xiao-Hu Shao 1; Xiang-Dong Zhou 1
	2016
会议名称	International Conference on Frontiers in Handwriting Recognition (ICFHR)
会议录名称	International Conference on Frontiers in Handwriting Recognition (ICFHR)
会议日期	October 23-26
会议地点	Shenzhen, China
摘要	Deep convolutional neural networks (DCNN) have recently achieved state-of-the-art performance on handwritten Chinese character recognition (HCCR). However, most of DCNN models employ the softmax activation function and minimize cross-entropy loss, which may loss some inter-class information. To cope with this problem, we demonstrate a small but consistent advantage of using both classification and similarity ranking signals as supervision. Specifically, the presented method learns a DCNN model by maximizing the inter-class variations and minimizing the intra-class variations, and simultaneously minimizing the cross-entropy loss. In addition, we also review some loss functions for similarity ranking and evaluate their performance. Our experiments demonstrate that the presented method achieves state-of-the-art accuracy on the well-known ICDAR 2013 offline HCCR competition dataset.
关键词	Hccr
文献类型	会议论文
条目标识符	http://ir.ia.ac.cn/handle/173211/12470
专题	多模态人工智能系统全国重点实验室_模式分析与学习
通讯作者	Cheng Cheng
作者单位	1.重庆绿色科学研究院 2.中科院自动化所
推荐引用方式 GB/T 7714	Cheng Cheng,Xu-Yao Zhang,Xiao-Hu Shao,et al. Handwritten Chinese character recognition by joint classification and similarity ranking[C],2016.