摘要
分析了非限制自由手写粘连数字串的下边缘特性后,利用句法分析的判别手段对粘连字串的个数进行了判断,在不限制书写长度的情况下,取得了87.5%以上的判断正确率.该算法在银行支票手写账号自动识别系统中得到了有效的应用.
With the use of syntax analysis and waveform analysis, the length estimation algorithm of unconstrained handwritten connected numeral strings was proposed. Length estimation algorithm was the necessary preprocessing method to the segmentation and recognition of connected numeral strings. The correct rate of the algorithm reached above 87.5% at the free written style, and the algorithm has been incorporated into a practical bank check recognition system.
出处
《北京邮电大学学报》
EI
CAS
CSCD
北大核心
2004年第3期63-67,共5页
Journal of Beijing University of Posts and Telecommunications
基金
国家"863计划"项目(2001AA114080)
教育部重点科研项目(02029)
关键词
边缘特征
句法分析
粘连字串
个数判断
Edge detection
Feature extraction
Image processing
Waveform analysis
