CASIA OpenIR  > 09年以前成果
吴少泓; 王云宽; 孙涛
Source Publication计算机应用
Other AbstractDue to the mutability of unstrained or handwritten digits, most algorithms in previous works either forfeited easy implementation for high accuracy, or vice versa. In this paper, we proposed a novel feature descriptor,distance distribution histogram(DDH) and another adapted shape accumulate histogram(SAH) feature descriptor based on shape context which was not only easy to implement, but also was robust to noise and distortion. To make hybrid features more comprehensive, we also combined other adapted topological features. The new congregated features were complementary as they were formed from different original feature sets extracted by different means. What’s more, they were not complicate. We then employed three support vector machines (SVMs) with different feature vector doing recognition and integrated their results to get the final classification. Best result among several experiments based on different data bases was 99.213%, which demonstrated that our proposed algorithm was very robust and effective.
Keyword特征提取 距离映射直方图 相对链码 形状累积直方图 块分布图 支持向量机
Document Type期刊论文
Recommended Citation
GB/T 7714
吴少泓,王云宽,孙涛. 基于距离分布直方图的数字识别算法[J]. 计算机应用,2012,8(32):2299-2304.
APA 吴少泓,王云宽,&孙涛.(2012).基于距离分布直方图的数字识别算法.计算机应用,8(32),2299-2304.
MLA 吴少泓,et al."基于距离分布直方图的数字识别算法".计算机应用 8.32(2012):2299-2304.
Files in This Item: Download All
File Name/Size DocType Version Access License
基于距离分布直方图的数字识别算法.pdf(941KB)期刊论文作者接受稿开放获取CC BY-NC-SAView Download
Related Services
Recommend this item
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[吴少泓]'s Articles
[王云宽]'s Articles
[孙涛]'s Articles
Baidu academic
Similar articles in Baidu academic
[吴少泓]'s Articles
[王云宽]'s Articles
[孙涛]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[吴少泓]'s Articles
[王云宽]'s Articles
[孙涛]'s Articles
Terms of Use
No data!
Social Bookmark/Share
File name: 基于距离分布直方图的数字识别算法.pdf
Format: Adobe PDF
All comments (0)
No comment.

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.