文章检索

  • 检索
  • 检索词:
  • 高级检索
您是今天第 1948位访问者
您是第 4908615 位访问者
高仿真光敏印章盖印印文的自动识别
Automatic recognition of high-simulation photosensitive seal stamping
投稿时间:2019-05-13  修订日期:2019-07-03
中文关键词:高仿真光敏印章  盖印印文  自动识别  卷积神经网络
英文关键词:high simulation photosensitive seal  stamp impression  auto recognition  convolutional neural network
基金项目:财政部基本科研业务费2018JB022
作者单位E-mail
张倩 中国人民公安大学 490793861@qq.com 
韩星周 公安部物证鉴定中心  
摘要点击次数: 97
全文下载次数: 0
中文摘要:
      目的:实现高仿真光敏印章印文的自动识别,探究训练样本量、网络模型对识别准确率的影响。方法:通过扫描打印伪造法、拓印设计伪造法制备两枚高仿光敏印章,盖印3000枚印文作为训练样本,盖印30枚印文作为测试样本,利用卷积神经网络四种模型实现高仿真光敏印章印文的鉴别。结果:四种网络模型均能得到100%的识别准确率。结论:实验结果表明针对高仿真光敏印章印文识别任务,卷积神经网络能作为一种可行的方法为检验提供辅助参考,综合分析四种网络模型,Resnet50是最优选择。
英文摘要:
      Objective: To realize the automatic recognition of high simulation photosensitive seals and explore the influence of training sample size and network model on the recognition accuracy. Methods: Two high imitation photosensitive seals were prepared by scanning, printing and rubbing design forgery. 3000 seals were stamped as training samples and 30 seals were stamped as test samples. Four models of convolutional neural network were used to identify the high simulation photosensitive seals. Result: All the four network models can get 100% recognition accuracy. Conclusion: The experimental results show that the convolutional neural network can be used as a feasible method to provide auxiliary reference for the test and comprehensively analyze the four network models, Resnet50 is the best choice for the task of high-simulation photosensitive seal printing.
引用本文:
DOI:
学科分类代码:
  查看/发表评论  下载PDF阅读器

分享按钮