Meter Value Recognition using Locally Connected Hierarchical Networks.pdfVIP

下载本文档

7
0
约1.49万字
约 6页
2017-04-13 发布于江苏
举报
版权申诉

Meter Value Recognition using Locally Connected Hierarchical Networks.pdf

1、本文档共6页，可阅读全部内容。
2、原创力文档（book118）网站文档一经付费（服务费），不意味着购买了该文档的版权，仅供个人/单位学习、研究之用，不得用于商业用途，未经授权，严禁复制、发行、汇编、翻译或者网络传播等，侵权必究。
3、本站所有内容均由合作方或网友上传，本站不对文档的完整性、权威性及其观点立场正确性做任何保证或承诺！文档内容仅供研究参考，付费前请自行鉴别。如您付费，意味着您自己接受本站规则且自行承担风险，本站不退款、不进行额外附加服务；查看《如何避免下载的几个坑》。如果您已付费下载过本站文档，您可以点击这里二次下载。
4、如文档侵犯商业秘密、侵犯著作权、侵犯人身权等，请点击“版权申诉”（推荐），也可以打举报电话：400-050-0827(电话支持时间：9:00-18:30)。
5、该文档为VIP文档，如果想要下载，成为VIP会员后，下载免费。
6、成为VIP后，下载本文档将扣除1次下载权益。下载后，不支持退款、换文档。如有疑问请联系我们。
7、成为VIP后，您将拥有八大权益，权益包括：VIP文档下载权益、阅读免打扰、文档格式转换、高级专利检索、专属身份标志、高级客服、多端互通、版权登记。
8、VIP文档为合作方或网友上传，每下载1次，网站将根据用户上传文档的质量评分、类型等，对文档贡献者给予高额补贴、流量扶持。如果你也想贡献VIP文档。上传文档

Meter Value Recognition using Locally Connected Hierarchical Networks

Meter Value Recognition using Locally Connected Hierarchical Networks Sven Behnke International Computer Science Institute 1947 Center St., Berkeley, CA, 94704, USA Abstract. This paper describes a two-stage system for the recognition of postage meter values. A feed-forward Neural Abstraction Pyramid is initialized in an unsupervised manner and trained in a supervised fashion to classify an entire digit block. It does not need prior digit segmentation. If the block recognition is not confident enough, a second stage tries to recognize single digits, taking into account the block classifier output for a neighboring digit as context. The system is evaluated on a large database. It can recognize meter values that are hard to read for humans. 1 Introduction Meter stamps are commonly used to mark letters. Postage meters print stamps usually with red ink in the letter’s upper right corner. During automatic mail processing the stamps are recognized in order to compare the postage value to the letter’s weight. Meter value reading consists of three steps: meter stamp detection, value localization, and value recognition. In the following, a system is described that covers only the last step, the recognition of isolated meter values. A block classifier is trained to recognize the entire meter value, with- out prior digit segmentation. It is based on the Neural Abstraction Pyramid architecture [1, 2, 3]. If this classifier cannot make a confident decision, single digit classifiers are queried and combined with its results. A database of Swedish Post meter marks is used that has been collected by Siemens ElectroCom Postautomation GmbH. It contains 5,471 examples, randomly assigned to a training set of size 4,372 and a test set of 1,099. Fig. 1 shows some sample images. As can be seen, the recognition of the meter value is not an easy task. The images have low resolution and low contrast. High variance of the print, the lighting, and the background complicate recognition further.