WO2017140218A1 - 信息识别方法及装置 - Google Patents
信息识别方法及装置 Download PDFInfo
- Publication number
- WO2017140218A1 WO2017140218A1 PCT/CN2017/072956 CN2017072956W WO2017140218A1 WO 2017140218 A1 WO2017140218 A1 WO 2017140218A1 CN 2017072956 W CN2017072956 W CN 2017072956W WO 2017140218 A1 WO2017140218 A1 WO 2017140218A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- information
- sub
- server
- result
- identify
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Ceased
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
- G06V30/40—Document-oriented image-based pattern recognition
- G06V30/42—Document-oriented image-based pattern recognition based on the type of document
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/10—Image acquisition
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
- G06V30/10—Character recognition
Definitions
- the present application relates to the field of optical character recognition, and in particular to an information recognition method and apparatus for performing multi-dimensional information cross-validation using optical character recognition technology.
- OCR Optical Character Recognition
- OCR can be applied to the review of identity documents or the entry of electronic card information.
- IDs or e-cards usually have multiple dimensions of information such as name (can be pinyin or English), expiration date, number, etc., and the information can be cross-validated, that is, information of multiple dimensions needs to be maintained.
- the information of one of the dimensions does not correspond to the information of other dimensions, it means that the identity document or the electronic card may be forged.
- information of multiple dimensions is input into the background of the computer by the OCR device to verify the correspondence between the information.
- the computer processes the information of each dimension one by one, and the entire information recognition takes a long time.
- the embodiment of the present application provides an information identification method with less occupation time.
- an information identification method includes:
- the result of scanning the original file information is determined according to the results of the identification of the received sub-information by each server.
- the embodiment of the present application further provides an information identifying apparatus, including:
- An acquisition module configured to acquire scan information obtained by optically scanning the original file information
- a splitting module configured to split the scan information into at least two sub-information
- a distribution module configured to send the at least two pieces of sub-information to different server ends, to separately identify the received sub-information by each server;
- a processing module configured to determine a result of scanning the original file information according to a result that each server separately identifies the received sub-information.
- the received sub-information is separately identified by each server, so that the time taken to identify the scan information can be reduced, thereby improving the efficiency of information recognition.
- FIG. 1 is a flowchart of an information identification method according to an embodiment of the present application
- FIG. 2 is a schematic structural diagram of an information identification apparatus according to an embodiment of the present application.
- S100 Acquire scan information obtained by optically scanning the original file information.
- the original document information may specifically refer to identity document information and electronic card information.
- Identity documents specific, for example, ID cards, student ID cards, military officers and so on.
- the information of these identification documents may have information of multiple dimensions such as name, identification number, and effective date. Information from multiple dimensions can be cross-validated to verify or identify the identity of the holder.
- Electronic cards specific, such as credit cards, debit cards, membership cards, and the like. These electronic cards also have information of multiple dimensions such as name and identification number. Information from multiple dimensions can be cross-validated to verify or identify the identity of the holder.
- the optical lens is prepared for information collection.
- the optical character recognition device determines the shape of the character in the original document information, that is, by detecting the dark and bright patterns, and then converts the shape into a corresponding electronic signal, thereby obtaining the scan information.
- S200 Split the scan information into at least two sub-information.
- splitting the scan information into at least two pieces of sub-information includes:
- the scan information is split into at least two sub-informations according to the location of the cluster of characters in the original file information.
- information of different dimensions is usually arranged at different positions of the original file.
- ID card the ID number and name are usually arranged in different positions of the ID card.
- the template may be used to treat the information of different dimensions of the ID card differently, thereby splitting the ID card information into the ID number information and the name information.
- the same processing method can be used to split the scan information into at least two sub-information.
- splitting the scan information into at least two pieces of sub-information specifically includes:
- the scan information is split into at least two sub-informations according to the type of characters in the original file information.
- information of different dimensions can be represented by the same character type.
- the character type of the ID card number is a numeric or alphabetic character; and the character type of the name is a Chinese character or an English alphabet. Except for the ID number of the ID card number, the characters are all numbers. Therefore, you can find the 18-character character starting with the number as the number information, and the Chinese character or the letter leading to the trailing blank part as the name information.
- information identification or auditing information of different dimensions of the ID card may be treated differently according to different character types, thereby splitting the ID card information into ID number information and name information.
- the same processing method can be used to split the scan information into at least two sub-information.
- S300 Send the at least two pieces of sub-information to different servers to identify the received sub-information by each server.
- the size and weight of the optical character recognition device are often subject to the convenience of operation of the device, thereby having an important adverse effect on its hardware. Specifically, it can be expressed that the operation speed cannot be increased with higher hardware specifications.
- the at least two pieces of sub-information are sent to different server terminals to separately identify the received sub-information by each server, thereby reducing the time taken for identifying the scan information. Therefore, the efficiency of information recognition can be improved.
- the optical character recognition device can be networked with other servers having higher hardware specifications, and the sub-information can be sent to these servers with higher computing speeds, so that the sub-information can be quickly identified.
- the method further includes:
- determining to identify the sub-information error when the recognition result of the at least one algorithm is not passed, determining to identify the sub-information error; when all the recognition results of the algorithm are uncertain, determining to identify the sub-information error; when all the algorithms are identified If there is no failure in the result, and at least one of the recognition results of all the algorithms is passed, it is determined that the identification of the sub-information is correct.
- the sub-information may be identified according to a preset algorithm, and a confidence probability corresponding to the recognition result may be generated. Confidence probabilities can be divided into three different levels, specific, such as pass, fail, and uncertain. For example, the name information contains the word " ⁇ ".
- the result of the identification by the server according to an algorithm is “shi”, and the confidence probability is 80%, that is, the recognition result is “passed” according to the algorithm.
- the result of the identification by the server according to an algorithm is “shi”, and the confidence probability is 50%, that is, the recognition result is “unsure” according to the algorithm.
- the result of the identification by the server according to an algorithm is “earth”, and the confidence probability is 30%, that is, the recognition result is “not passed” according to the algorithm.
- the passage, non-pass, and uncertainty can correspond to the range of confidence probabilities.
- the range of confidence probability ⁇ 60% can correspond to the recognition result as “pass”; 60%> confidence probability ⁇ 40% can correspond to the recognition result as “unsure”; 40%> confidence probability can correspond to the recognition result as “not passed” .
- a plurality of algorithms may be used to jointly identify the sub-information to improve the sub-information recognition efficiency. For a sub-information, when the recognition result of the at least one algorithm is not passed, determining to identify the sub-information error; when all the recognition results of the algorithm are uncertain, determining to identify the sub-information error; when all the algorithms are identified If there is no failure in the result, and at least one of the recognition results of all the algorithms is passed, it is determined that the identification of the sub-information is correct.
- P represents that the recognition result of the sub-information is "pass"
- the determination rule of the sub-information 1 can be determined according to the determination rule of the sub-information identification, and the recognition result of the sub-information 1 is correct.
- the determination result of the sub-information 1 can be determined to be erroneous based on the determination rule of the sub-information identification.
- a plurality of algorithms can be used to improve the accuracy of the recognition.
- the algorithm A1, the algorithm B1, and the algorithm B2 both consider that the recognition result is "passed", the accuracy or credibility of the recognition result is relatively high.
- the algorithm A1, the algorithm B1, and the algorithm B2 make up for further processing when the recognition result of the sub-information 1 is "undefined” by the algorithm A2 and the algorithm A3, without re-scanning the sub-information 1, thereby improving the processing efficiency.
- S400 Determine a result of scanning the original file information according to the results of the identification of the received sub-information by each server.
- the result of the server identifying the received sub-information indicates the result of a confidence probability of the sub-information.
- A1 A2 A3 B1 B2 result Sub-information 1 P U U P P correct Sub-information 2 P P U U P correct Sub-information 3 P U P U P correct Sub information 4 U P P U U correct Scan result correct
- P represents that the recognition result of the sub-information is "pass"
- the mutual authentication between the sub-information can also be used to improve the confirmation of the degree of confidence in the result of scanning the original file information.
- the information of each dimension in the original document information for example, the number information and the name information in the identity document information are related to each other. Therefore, cross-validation between sub-information of different dimensions can be performed to confirm the result of scanning the original document information.
- the identity document information as an example, that is, the number information and the name information in the original file information have a backup in the database in the background of the computer.
- the database behind the computer here can be the identity database about the security department.
- identifying the obtained number information and identifying the obtained name information it may be verified as interrelated information in a database in the background of the computer, that is, identifying the obtained number information and identifying the obtained name information in a database in the background of the computer
- identifying the obtained number information and identifying the obtained name information in a database in the background of the computer
- the sub information includes a plurality of information elements
- the method further includes:
- the above identification information is still taken as an example to illustrate.
- the sub-information of the name information and the number information is sent to different servers to identify the received sub-information by each server. It is assumed that the number information as the sub information is transmitted to the first server to cause the first server to recognize the received number information according to different algorithms.
- the recognition results of all the algorithms are all uncertain, it is determined that the identification result of the information element is uncertain. That is, for a certain number "1" in the number information, assuming that the number is recognized as "7", the confidence probability of all the algorithms for the result is considered to be: 60%> confidence probability ⁇ 40%. Then, it is determined that the recognition result of the number is uncertain.
- the recognition result of all the algorithms does not pass, and at least one of the recognition results passes, it is determined that the recognition result of the sub information is passed. For a number "1" in the number information, it is assumed that the number is recognized as "1". All algorithms consider the confidence probability of the result to be ⁇ 40%, and some algorithms consider the confidence probability of the result to be ⁇ 60%. Then, it is determined that the recognition result of the number is passed.
- the number information For the number information, if the identification result of at least one of the number information is not passed, the number information is The recognition result is not passed.
- the recognition result of the number information is uncertain.
- each information element in the sub-information is identified by using a plurality of algorithms. For each information element in the sub-information, if the information element has no recognition result of “not passing”, Some algorithms in the algorithm consider that the recognition result of the information element is "passed”. When some algorithms think that the recognition result of the information element is "unsure”, according to the rule, the recognition result of the information element can be considered as "passed”. Thus, it is not necessary to re-identify the information element in such a situation.
- the method further includes:
- the original document information is again optically scanned to re-identify the sub-information identifying the error.
- the identity document information is still taken as an example.
- the identity document information is again optically scanned to re-identify the wrong number information.
- the efficiency of information recognition can be improved.
- the original file information is identity document information or electronic card information;
- the at least two sub-information includes name information and number information.
- the corresponding portion in the original document information is re-scanned.
- the name portion of the identity card is re-scanned to re-acquire the name sub-information. Since only the sub-information identifying the error in the original document information is re-identified, it is possible to save time and improve the information recognition efficiency with respect to rescanning and recognizing the original document information scanning result.
- the ID card reviewer uses a scanner with OCR to scan the ID card.
- the scanner splits the ID card information into at least the ID number and name.
- the ID number and name are sent to the server that is networked with the scanner for identification.
- These servers can be configured with high-quality computing resources relative to the scanner.
- a variety of algorithms can be used to improve the efficiency of ID card identification.
- a plurality of algorithms may be used, so that when an algorithm of a plurality of algorithms cannot judge the recognition result of the ID number, the recognition rate of the ID number is improved by various algorithms.
- the present application further provides an information identification apparatus 1 including:
- the acquiring module 11 is configured to acquire scan information obtained by optically scanning the original file information.
- a splitting module 12 configured to split the scan information into at least two sub-information
- a distribution module 13 is configured to send the at least two pieces of sub-information to different server ends, to separately identify the received sub-information by each server;
- the processing module 14 is configured to determine a result of scanning the original file information according to the results of the identification of the received sub-information by each server.
- the splitting module 12 is configured to:
- the scan information is split into at least two sub-informations according to the location of the cluster of characters in the original file information.
- the splitting module 12 is configured to:
- the scan information is split into at least two sub-informations according to the type of characters in the original file information.
- the distribution module 13 is configured to send the at least two pieces of sub-information to different servers to identify the received sub-information by each server. Specifically for:
- determining to identify the sub-information error when the recognition result of the at least one algorithm is not passed, determining to identify the sub-information error; when all the recognition results of the algorithm are uncertain, determining to identify the sub-information error; when all the algorithms are identified If there is no failure in the result, and at least one of the recognition results of all the algorithms is passed, it is determined that the identification of the sub-information is correct.
- the collecting module is further configured to:
- the original document information is again optically scanned to re-identify the sub-information identifying the error.
- the original file information is identity document information or electronic card information.
- the received sub-information is separately identified by each server, so that the time taken to identify the scan information can be reduced, thereby improving the efficiency of information recognition.
- embodiments of the present invention can be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or a combination of software and hardware. Moreover, the invention can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) including computer usable program code.
- computer-usable storage media including but not limited to disk storage, CD-ROM, optical storage, etc.
- the computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable numerical processing device to operate in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture comprising the instruction device.
- the apparatus implements the functions specified in one or more blocks of a flow or a flow and/or block diagram of the flowchart.
- These computer program instructions can also be loaded onto a computer or other programmable numerical processing device such that a series of operational steps are performed on a computer or other programmable device to produce computer-implemented processing for execution on a computer or other programmable device.
- the instructions provide steps for implementing the functions specified in one or more of the flow or in a block or blocks of a flow diagram.
- a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.
- processors CPUs
- input/output interfaces network interfaces
- memory volatile and non-volatile memory
- the memory may include non-persistent memory, random access memory (RAM), and/or non-volatile memory in a computer readable medium, such as read only memory (ROM) or flash memory.
- RAM random access memory
- ROM read only memory
- Memory is an example of a computer readable medium.
- Computer readable media includes both permanent and non-persistent, removable and non-removable media.
- Information storage can be implemented by any method or technology.
- the information can be computer readable instructions, numerical structures, modules of the programs, or other values.
- Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read only memory. (ROM), electrically erasable programmable read only memory (EEPROM), flash memory or other memory technology, CD-ROM, digital versatile disc (DVD) or other optical storage, magnetic cassette, magnetic tape storage or other magnetic storage device or any other non-transportable media that can be used for storage Calculate information accessed by the device.
- computer readable media does not include temporary storage of computer readable media, such as modulated numerical signals and carrier waves.
- embodiments of the present application can be provided as a method, system, or computer program product.
- the present application can take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment in combination of software and hardware.
- the application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) including computer usable program code.
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Artificial Intelligence (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- General Engineering & Computer Science (AREA)
- Character Discrimination (AREA)
Abstract
本申请公开了一种信息识别方法及装置,所述方法包括:获取针对原始文件信息进行光学扫描得到的扫描信息;将所述扫描信息拆分为至少两个子信息;将所述至少两个子信息发送到不同的服务端,以通过各服务端分别对接收到的子信息进行识别;根据各服务端分别对接收到的子信息识别的结果,确定对原始文件信息扫描的结果。在本申请提供的实施例中,通过各服务端分别对接收到的子信息进行识别,从而可以减少对扫描信息进行识别所消耗的时间,从而可以提高信息识别的效率。
Description
本申请要求2016年02月18日递交的申请号为201610091533.4、发明名称为“信息识别方法及装置”的中国专利申请的优先权,其全部内容通过引用结合在本申请中。
本申请涉及光学字符识别领域,尤其涉及一种应用光学字符识别技术进行多维信息交叉验证的信息识别方法及装置。
OCR(Optical Character Recognition,光学字符识别)是指电子设备(例如扫描仪或数码相机)扫描衬底上的字符,通过检测暗、亮的模式确定字符的形状,然后用字符识别方法将形状翻译成计算机文字的过程。
OCR可以应用于身份证件的审核或者电子卡信息的录入。这些身份证件或电子卡通常有姓名(可以是拼音或英文)、有效期、号码等多个维度的信息,而且这些信息是可以交叉验证的,也就是,多个维度的信息需保持对应关系。当其中某个维度的信息与其他维度的信息不能对应时,则意味着身份证件或电子卡可能是伪造的。
在实现现有技术过程中,发明人发现现有技术中至少存在如下问题:
现有的信息识别方法中,多个维度的信息被OCR设备输入到计算机后台进行信息之间对应关系的验证。计算机后台逐一处理每个维度的信息,整个信息识别占用的时间长。
因此,需要提供一种信息识别占用时间少的技术方案。
发明内容
本申请实施例提供一种占用时间少的信息识别方法。
具体的,一种信息识别方法,包括:
获取针对原始文件信息进行光学扫描得到的扫描信息;
将所述扫描信息拆分为至少两个子信息;
将所述至少两个子信息发送到不同的服务端,以通过各服务端分别对接收到的子信息进行识别;
根据各服务端分别对接收到的子信息识别的结果,确定对原始文件信息扫描的结果。
本申请实施例还提供一种信息识别装置,包括:
采集模块,用于获取针对原始文件信息进行光学扫描得到的扫描信息;
拆分模块,用于将所述扫描信息拆分为至少两个子信息;
分发模块,用于将所述至少两个子信息发送到不同的服务端,以通过各服务端分别对接收到的子信息进行识别;
处理模块,用于根据各服务端分别对接收到的子信息识别的结果,确定对原始文件信息扫描的结果。
本申请实施例提供的信息识别方法和装置,至少具有如下有益效果:
在本申请提供的实施例中,通过各服务端分别对接收到的子信息进行识别,从而可以减少对扫描信息进行识别所消耗的时间,从而可以提高信息识别的效率。
此处所说明的附图用来提供对本申请的进一步理解,构成本申请的一部分,本申请的示意性实施例及其说明用于解释本申请,并不构成对本申请的不当限定。在附图中:
图1为本申请实施例提供的信息识别方法流程图;
图2为本申请实施例提供的信息识别装置的结构示意图。
为使本申请的目的、技术方案和优点更加清楚,下面将结合本申请具体实施例及相应的附图对本申请技术方案进行清楚、完整地描述。显然,所描述的实施例仅是本申请一部分实施例,而不是全部的实施例。基于本申请中的实施例,本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例,都属于本申请保护的范围。
请参照图1,计算机进行信息识别时,可以按照以下步骤进行:
S100:获取针对原始文件信息进行光学扫描得到的扫描信息。
在本申请应用的场景中,原始文件信息可以特别的指,身份证件信息、电子卡信息。身份证件,具体的,例如,身份证、学生证、军官证等等。这些身份证件的信息上可以具有姓名、识别号码、有效日期等多个维度的信息。可以利用多个维度的信息相互交叉验证,用以对持有者的身份进行审核或识别。电子卡,具体的,例如信用卡、借记卡、会员卡等等。这些电子卡,同样具有姓名、识别号码等多个维度的信息。可以利用多个维度的信息相互交叉验证,用以对持有者的身份进行审核或识别。
将这些原始文件信息放置在光学字符识别设备处,计算机可以利用光学字符识别设
备的光学镜头进行信息的采集。光学字符识别设备将原始文件信息中的字符,也就是通过检测暗、亮的模式确定字符的形状,然后转化为对应的电子信号,从而获得扫描信息。
S200:将所述扫描信息拆分为至少两个子信息。
进一步的,在本申请提供的一种实施例中,将所述扫描信息拆分为至少两个子信息,具体包括:
根据原始文件信息中字符聚簇的位置,将扫描信息拆分为至少两个子信息。
在原始文件信息中,不同维度的信息通常布置在原始文件的不同位置。具体的,以身份证为例,身份证号码和姓名通常布置在身份证的不同位置。在进行信息识别或审核时,可以使用模板,将身份证的不同维度的信息区别对待,从而将身份证信息拆分为身份证号码信息和姓名信息。同样,对于其他身份证件信息和电子卡信息,可以采用同样的处理方式,将所述扫描信息拆分为至少两个子信息。
进一步的,将所述扫描信息拆分为至少两个的子信息,具体包括:
根据原始文件信息中字符的类型,将扫描信息至少拆分为两个子信息。
在原始文件信息中,不同维度的信息可以以同的字符类型表示。具体的,以身份证为例,身份证号码的字符类型为数字或字母型字符;而姓名的字符类型为汉字或英文字母的类型。除去身份证号码的尾数可能是字母外,其余字符均为数字,因此,可以查找以数字开头的18位字符作为号码信息,而将以汉字或字母开头到尾端空白部分作为姓名信息。在进行信息识别或审核时,可以根据不同的字符类型,将身份证的不同维度的信息区别对待,从而将身份证信息拆分为身份证号码信息和姓名信息。同样,对于其他身份证件信息和电子卡信息,可以采用同样的处理方式,将所述扫描信息拆分为至少两个子信息。
S300:将所述至少两个子信息发送到不同的服务端,以通过各服务端分别对接收到的子信息进行识别。
应当指出的是,光学字符识别设备的尺寸、重量往往受制于该设备操作的便利性,从而对其硬件带来重要的不利影响。具体的可以表现在,不能以较高的硬件规格来提高运算速度。在本申请提供的实施例中,将所述至少两个子信息发送到不同的服务端,以通过各服务端分别对接收到的子信息进行识别,从而可以减少对扫描信息进行识别所消耗的时间,从而可以提高信息识别的效率。具体的,可以将光学字符识别设备与其他具有较高硬件规格的服务器联网,将这些子信息发送到这些具有较高运算速度的服务器,从而可以快速进行子信息的识别。
进一步的,在本申请提供的又一实施例中,所述方法还包括:
将所述至少两个子信息发送到不同的服务端,以通过各服务端分别对接收到的子信息进行识别,具体包括:
将所述至少两个子信息发送到不同的服务端,以使服务端根据不同的算法对接收到的子信息进行识别;其中,服务端通过每个算法对子信息进行识别的结果包括通过、不通过、不确定;
针对一个子信息,当至少一个算法的识别结果为不通过时,则确定识别该子信息错误;当所有算法的识别结果均为不确定时,则确定识别该子信息错误;当所有算法的识别结果中没有不通过,并且,所有算法的识别结果中至少有一个为通过时,则确定识别该子信息识别正确。
当服务器接收到子信息时,可以根据预设的算法对子信息进行识别,并且可以生成对应识别结果的置信概率。置信概率可以分为三个不同的层次,具体的,例如通过、不通过、不确定。例如,姓名信息中包含“士”字。服务器根据某算法识别的结果为“士”字,置信概率为80%,也就是根据该算法认为该识别结果为“通过”。或者,服务器根据某算法识别的结果为“士”字,置信概率为50%,也就是根据该算法认为该识别结果为“不确定”。再或者,服务器根据某算法识别的结果为“土”字,置信概率为30%,也就是根据该算法认为该识别结果为“不通过”。应当指出的是,这里的通过、不通过、不确定可以与置信概率的范围相对应。例如,置信概率≥60%的范围可以对应识别结果为“通过”;60%>置信概率≥40%可以对应识别结果为“不确定”;40%>置信概率可以对应识别结果为“不通过”。
当子信息的识别结果为“不通过”或“不确定”时,则需要重新对子信息进行识别,增加子信息的识别时间,降低子信息识别效率。在本申请实施例中,可以使用多种算法对子信息进行联合识别,以提高子信息识别效率。针对一个子信息,当至少一个算法的识别结果为不通过时,则确定识别该子信息错误;当所有算法的识别结果均为不确定时,则确定识别该子信息错误;当所有算法的识别结果中没有不通过,并且,所有算法的识别结果中至少有一个为通过时,则确定识别该子信息识别正确。
请参照表1,为使用不同的算法(A1、A2、A3、B1、B2)对不同的子信息进行识别时的举例。
表1
| A1 | A2 | A3 | B1 | B2 | 识别结果 |
| 子信息1 | P | U | U | P | P | 正确 |
| 子信息2 | P | P | U | U | P | 正确 |
| 子信息3 | P | U | P | N | P | 错误 |
| 子信息4 | N | P | P | P | P | 错误 |
其中,P代表子信息的识别结果为“通过”;
U代表子信息的识别结果为“不确定”;
N代表子信息的识别结果为“不通过”。
例如,对于子信息1而言,对于同一个识别结果,算法A2、算法A3认为识别结果“不确定”时,根据子信息识别的确定规则可以确定,子信息1的识别结果正确。而对于子信息3而言,对于同一个识别结果,由于算法B1认为识别结果“不通过”,根据子信息识别的确定规则可以确定子信息1的识别结果错误。在本申请实施例中,由于不同的算法对不同类型的字符在识别时具有不同的准确度,所以使用多种算法可以提高识别的准确度。例如,对于子信息1,由于算法A1、算法B1、算法B2均认为识别结果“通过”,则该识别结果的准确性或者可信度相对较高。同时,算法A1、算法B1、算法B2弥补了算法A2、算法A3对子信息1的识别结果“不确定”时的进一步处理,而不必重新扫描子信息1,从而提高了处理效率。
同样,对于构成子信息的每一个字符可以使用相同的方式,以提高子信息中每一个字符的识别效率。
S400:根据各服务端分别对接收到的子信息识别的结果,确定对原始文件信息扫描的结果。
服务端对接收到的子信息识别的结果表明了子信息的一种置信概率的结果。
当所有子信息的识别结果正确时,可以确定对原始文件信息扫描的结果正确,例如表2。
请参照表2,为使用不同的算法(A1、A2、A3、B1、B2)对不同的子信息进行识别以及最终对原始文件信息扫描的结果确认。
表2
| A1 | A2 | A3 | B1 | B2 | 结果 | |
| 子信息1 | P | U | U | P | P | 正确 |
| 子信息2 | P | P | U | U | P | 正确 |
| 子信息3 | P | U | P | U | P | 正确 |
| 子信息4 | U | P | P | U | U | 正确 |
| 扫描结果 | 正确 |
其中,P代表子信息的识别结果为“通过”;
U代表子信息的识别结果为“不确定”。
在表2中,对于子信息1,算法A1、算法B1、算法B2代表子信息的识别结果为“通过”,算法A2、算法A3代表子信息的识别结果为“不确定”,根据识别规则,子信息1的识别结果最终判定为“正确”。对于子信息2、子信息3、子信息4的情形类似。假设,扫描信息被拆分为子信息1、子信息2、子信息3和子信息4。在本申请实施例中,子信息1、子信息2、子信息3和子信息4的识别结果最终分别被判定为“正确”。当子信息1、子信息2、子信息3和子信息4在计算机后台的数据库中可以交叉验证时,可以确认对原始文件信息扫描的结果为“正确”。
而现有技术中,仅使用一种算法时,例如,仅使用算法A1时,子信息1、子信息2、子信息3的识别结果为“通过”,而子信息4的识别结果为“不确定”。由于子信息4的识别结果无法最终被确认,需要重新对原始文件信息扫描。在本申请实施中,可以看出,当一种算法对于某个子信息的识别结果的置信概率低,也就是说,某个子信息的识别结果无法被确认时,使用多种算法,则有可能对该子信息的识别结果予以确认,从而不必重新对原始文件信息扫描,可以提高信息识别的效率。
进一步的,还可以通过子信息之间的交互验证,来提高对原始文件信息扫描的结果置信程度的确认。
原始文件信息中每一个维度的信息之间具有关联性,例如,身份证件信息中的号码信息和姓名信息是相互关联的。因此,可以通过不同维度的子信息之间进行交叉验证,以确认原始文件信息扫描的结果。仍以身份证件信息为例,也就是说,原始文件信息中的号码信息和姓名信息在计算机后台的数据库中具有备份。这里的计算机后台的数据库可以是有关安全部门的身份数据库。当识别获得的号码信息和识别获得的姓名信息,可以在计算机后台的数据库中验证为相互关联的信息时,也就是说,识别获得的号码信息和识别获得的姓名信息在计算机后台的数据库中相互匹配时,可以确定对原始文件信息扫描的结果正确。
进一步的,在本申请提供的又一实施例中,子信息包括若干信息元,所述方法还包括:
将所述至少两个子信息发送到不同的服务端,以通过各服务端分别对接收到的子信息进行识别,具体包括:
将所述至少两个子信息发送到不同的服务端,以使服务端根据不同的算法对接收到
的子信息进行识别;其中,服务端通过每个算法对子信息进行识别的结果包括通过、不通过、不确定;
针对子信息的每一个信息元:
当至少一个算法的识别结果为不通过时,则确定该信息元的识别结果为不通过;
当所有算法的识别结果均为不确定时,则确定该信息元的识别结果为不确定;
当所有算法的识别结果没有不通过,并且,识别结果至少有一个通过时,则确定该信息元的识别结果为通过;
针对子信息:
当一个子信息中的信息元的识别结果至少有一个不通过时,则该子信息的识别结果为不通过;
当一个子信息中的信息元的识别结果均为不确定时,则确定该子信息的识别结果为不确定;
当一个子信息中的信息元的识别结果均为通过时,则确定该子信息的识别结果为通过。
具体的,仍以上面的身份证件信息为例来说明。将姓名信息和号码信息这两个子信息发送到不同的服务端,以通过各服务端分别对接收到的子信息进行识别。假设,作为子信息的号码信息被发送到第一服务器,以使第一服务器根据不同的算法对接收到的号码信息进行识别。
对于号码信息中的每一个数字,假设,至少一个算法的识别结果为不通过时,则确定该信息元的识别结果为不通过。也就是说,对于号码信息中的某个数字“1”,假设该数字被识别为“7”,至少有一种算法对该结果的置信概率认为小于30%,则确定该数字的识别结果为不通过。
而当所有算法的识别结果均为不确定时,则确定该信息元的识别结果为不确定。也就是,对于号码信息中的某个数字“1”,假设该数字被识别为“7”,所有算法对该结果的置信概率认为:60%>置信概率≥40%。则,确定该数字的识别结果为不确定。
当所有算法的识别结果没有不通过,并且,识别结果至少有一个通过时,则确定子信息的识别结果为通过。对于号码信息中的某个数字“1”,假设该数字被识别为“1”。所有算法认为该结果的置信概率均≥40%,其中,有的算法认为该结果的置信概率≥60%。则,确定该数字的识别结果为通过。
对于号码信息,号码信息中至少有一个数字的识别结果为不通过时,则该号码信息
的识别结果为不通过。
对于号码信息,号码信息中所有数字的识别结果均为不确定时,则该号码信息的识别结果为不确定。
对于号码信息,号码信息中所有数字的识别结果均为通过时,则该号码信息的识别结果为通过。在本申请实施例中,使用多种算法对子信息中的每一个信息元进行识别,对于子信息中的每一个信息元,在该信息元没有识别结果为“不通过”的前提下,多种算法中有的算法认为该信息元的识别结果“通过”,有的算法认为该信息元的识别结果“不确定”时,根据规则,可以认为该信息元的识别结果“通过”。从而不必在此种情形下重新对该信息元进行识别。
进一步的,在本申请提供的又一实施例中,所述方法还包括:
当至少一个子信息识别错误时,再次针对原始文件信息进行光学扫描以便重新对识别错误的子信息进行识别。
具体的,仍以身份证件信息为例,当号码信息识别错误时,再次对身份证件信息进行光学扫描以便重新对识别错误的号码信息进行识别。在该过程中,由于只对号码信息,也就是部分扫描信息重新识别,从而可以提高信息识别的效率。进一步的,在本申请提供的又一实施例中,所述原始文件信息为身份证件信息或电子卡信息;
至少两个子信息包括姓名信息和号码信息。
当存在子信息识别错误的情形时,重新扫描原始文件信息中对应的部分。例如对于身份信息的识别而言,当身份证的姓名识别错误时,重新扫描身份证的姓名部分以重新获取姓名子信息。由于仅对原始文件信息中识别错误的子信息重新识别,因而,相对于对原始文件信息扫描结果进行重新扫描并识别可以节省时间,提高信息识别效率。
下面以身份证的识别为例,来介绍一种本申请实施例的具体应用场景。身份证审核人员使用带有OCR的扫描器,对身份证进行扫描。扫描器将身份证信息至少拆分为身份证号码和姓名。身份证号码和姓名被发送到与扫描器联网的服务器进行识别。这些服务器相对于扫描器可以配置优质的运算资源。当身份证号码被正确识别,姓名被正确识别时,可以确定对身份证的扫描无误,以便于对身份证的持有者进行审核。而当身份证号码识别错误时,可以重新扫描身份证的号码部分。
另外还可以使用多种算法,以便提高身份证识别的效率。例如,可以使用多种算法,以便当多种算法中某一算法不能对身份证号码的识别结果进行判断时,通过多种算法提高身份证号码的识别率。
当然,对于身份证的姓名的识别同样适用。
以上是本申请实施例提供的信息识别方法,基于同样的思路,请参照图2,本申请还提供一种信息识别装置1,包括:
采集模块11,用于获取针对原始文件信息进行光学扫描得到的扫描信息;
拆分模块12,用于将所述扫描信息拆分为至少两个子信息;
分发模块13,用于将所述至少两个子信息发送到不同的服务端,以通过各服务端分别对接收到的子信息进行识别;
处理模块14,用于根据各服务端分别对接收到的子信息识别的结果,确定对原始文件信息扫描的结果。
进一步的,在本申请提供的又一实施例中,所述拆分模块12用于:
根据原始文件信息中字符聚簇的位置,将扫描信息拆分为至少两个子信息。
进一步的,在本申请提供的又一实施例中,所述拆分模块12用于:
根据原始文件信息中字符的类型,将扫描信息至少拆分为两个子信息。
进一步的,在本申请提供的又一实施例中,所述分发模块13,用于将所述至少两个子信息发送到不同的服务端,以通过各服务端分别对接收到的子信息进行识别,具体用于:
将所述至少两个子信息发送到不同的服务端,以使服务端根据不同的算法对接收到的子信息进行识别;其中,服务端通过每个算法对子信息进行识别的结果包括通过、不通过、不确定;
针对一个子信息,当至少一个算法的识别结果为不通过时,则确定识别该子信息错误;当所有算法的识别结果均为不确定时,则确定识别该子信息错误;当所有算法的识别结果中没有不通过,并且,所有算法的识别结果中至少有一个为通过时,则确定识别该子信息识别正确。
进一步的,在本申请提供的又一实施例中,所述采集模块还用于:
当至少一个子信息识别错误时,再次针对原始文件信息进行光学扫描以便重新对识别错误的子信息进行识别。
进一步的,在本申请提供的又一实施例中,所述原始文件信息为身份证件信息或电子卡信息。
在本申请提供的实施例中,通过各服务端分别对接收到的子信息进行识别,从而可以减少对扫描信息进行识别所消耗的时间,从而可以提高信息识别的效率。
本领域内的技术人员应明白,本发明的实施例可提供为方法、系统、或计算机程序产品。因此,本发明可采用完全硬件实施例、完全软件实施例、或结合软件和硬件方面的实施例的形式。而且,本发明可采用在一个或多个其中包含有计算机可用程序代码的计算机可用存储介质(包括但不限于磁盘存储器、CD-ROM、光学存储器等)上实施的计算机程序产品的形式。
本发明是参照根据本发明实施例的方法、设备(系统)、和计算机程序产品的流程图和/或方框图来描述的。应理解可由计算机程序指令实现流程图和/或方框图中的每一流程和/或方框、以及流程图和/或方框图中的流程和/或方框的结合。可提供这些计算机程序指令到通用计算机、专用计算机、嵌入式处理机或其他可编程数值处理设备的处理器以产生一个机器,使得通过计算机或其他可编程数值处理设备的处理器执行的指令产生用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的装置。
这些计算机程序指令也可存储在能引导计算机或其他可编程数值处理设备以特定方式工作的计算机可读存储器中,使得存储在该计算机可读存储器中的指令产生包括指令装置的制造品,该指令装置实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能。
这些计算机程序指令也可装载到计算机或其他可编程数值处理设备上,使得在计算机或其他可编程设备上执行一系列操作步骤以产生计算机实现的处理,从而在计算机或其他可编程设备上执行的指令提供用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的步骤。
在一个典型的配置中,计算设备包括一个或多个处理器(CPU)、输入/输出接口、网络接口和内存。
内存可能包括计算机可读介质中的非永久性存储器,随机存取存储器(RAM)和/或非易失性内存等形式,如只读存储器(ROM)或闪存(flash RAM)。内存是计算机可读介质的示例。
计算机可读介质包括永久性和非永久性、可移动和非可移动媒体可以由任何方法或技术来实现信息存储。信息可以是计算机可读指令、数值结构、程序的模块或其他数值。计算机的存储介质的例子包括,但不限于相变内存(PRAM)、静态随机存取存储器(SRAM)、动态随机存取存储器(DRAM)、其他类型的随机存取存储器(RAM)、只读存储器(ROM)、电可擦除可编程只读存储器(EEPROM)、快闪记忆体或其他内存技术、
只读光盘只读存储器(CD-ROM)、数字多功能光盘(DVD)或其他光学存储、磁盒式磁带,磁带磁磁盘存储或其他磁性存储设备或任何其他非传输介质,可用于存储可以被计算设备访问的信息。按照本文中的界定,计算机可读介质不包括暂存电脑可读媒体(transitory media),如调制的数值信号和载波。
还需要说明的是,术语“包括”、“包含”或者其任何其他变体意在涵盖非排他性的包含,从而使得包括一系列要素的过程、方法、商品或者设备不仅包括那些要素,而且还包括没有明确列出的其他要素,或者是还包括为这种过程、方法、商品或者设备所固有的要素。在没有更多限制的情况下,由语句“包括一个……”限定的要素,并不排除在包括所述要素的过程、方法、商品或者设备中还存在另外的相同要素。
本领域技术人员应明白,本申请的实施例可提供为方法、系统或计算机程序产品。因此,本申请可采用完全硬件实施例、完全软件实施例或结合软件和硬件方面的实施例的形式。而且,本申请可采用在一个或多个其中包含有计算机可用程序代码的计算机可用存储介质(包括但不限于磁盘存储器、CD-ROM、光学存储器等)上实施的计算机程序产品的形式。
以上所述仅为本申请的实施例而已,并不用于限制本申请。对于本领域技术人员来说,本申请可以有各种更改和变化。凡在本申请的精神和原理之内所作的任何修改、等同替换、改进等,均应包含在本申请的权利要求范围之内。
Claims (12)
- 一种信息识别方法,其特征在于,包括:获取针对原始文件信息进行光学扫描得到的扫描信息;将所述扫描信息拆分为至少两个子信息;将所述至少两个子信息发送到不同的服务端,以通过各服务端分别对接收到的子信息进行识别;根据各服务端分别对接收到的子信息识别的结果,确定对原始文件信息扫描的结果。
- 如权利要求1所述的方法,其特征在于,将所述扫描信息拆分为至少两个子信息,具体包括:根据原始文件信息中字符聚簇的位置,将扫描信息拆分为至少两个子信息。
- 如权利要求1所述的方法,其特征在于,将所述扫描信息拆分为至少两个的子信息,具体包括:根据原始文件信息中字符的类型,将扫描信息至少拆分为两个子信息。
- 如权利要求1所述的方法,其特征在于,将所述至少两个子信息发送到不同的服务端,以通过各服务端分别对接收到的子信息进行识别,具体包括:将所述至少两个子信息发送到不同的服务端,以使服务端根据不同的算法对接收到的子信息进行识别;其中,服务端通过每个算法对子信息进行识别的结果包括通过、不通过、不确定;针对一个子信息,当至少一个算法的识别结果为不通过时,则确定识别该子信息错误;当所有算法的识别结果均为不确定时,则确定识别该子信息错误;当所有算法的识别结果中没有不通过,并且,所有算法的识别结果中至少有一个为通过时,则确定识别该子信息识别正确。
- 如权利要求1所述的方法,其特征在于,所述方法还包括:当至少一个子信息识别错误时,再次针对原始文件信息进行光学扫描以便重新对识别错误的子信息进行识别。
- 如权利要求1-5中任一项所述的方法,其特征在于,所述原始文件信息为身份证件信息或电子卡信息;至少两个子信息包括姓名信息和号码信息。
- 一种信息识别装置,其特征在于,包括:采集模块,用于获取针对原始文件信息进行光学扫描得到的扫描信息;拆分模块,用于将所述扫描信息拆分为至少两个子信息;分发模块,用于将所述至少两个子信息发送到不同的服务端,以通过各服务端分别对接收到的子信息进行识别;处理模块,用于根据各服务端分别对接收到的子信息识别的结果,确定对原始文件信息扫描的结果。
- 如权利要求7所述的装置,其特征在于,所述拆分模块用于:根据原始文件信息中字符聚簇的位置,将扫描信息拆分为至少两个子信息。
- 如权利要求7所述的装置,其特征在于,所述拆分模块用于:根据原始文件信息中字符的类型,将扫描信息至少拆分为两个子信息。
- 如权利要求7所述的装置,其特征在于,所述分发模块,用于将所述至少两个子信息发送到不同的服务端,以通过各服务端分别对接收到的子信息进行识别,具体用于:将所述至少两个子信息发送到不同的服务端,以使服务端根据不同的算法对接收到的子信息进行识别;其中,服务端通过每个算法对子信息进行识别的结果包括通过、不通过、不确定;针对一个子信息,当至少一个算法的识别结果为不通过时,则确定识别该子信息错误;当所有算法的识别结果均为不确定时,则确定识别该子信息错误;当所有算法的识别结果中没有不通过,并且,所有算法的识别结果中至少有一个为通过时,则确定识别该子信息识别正确。
- 如权利要求7所述的装置,其特征在于,所述采集模块还用于:当至少一个子信息识别错误时,再次针对原始文件信息进行光学扫描以便重新对识别错误的子信息进行识别。
- 如权利要求7-11中任一项所述的装置,其特征在于,所述原始文件信息为身份证件信息或电子卡信息。
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610091533.4 | 2016-02-18 | ||
| CN201610091533.4A CN107092903A (zh) | 2016-02-18 | 2016-02-18 | 信息识别方法及装置 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2017140218A1 true WO2017140218A1 (zh) | 2017-08-24 |
Family
ID=59625567
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/CN2017/072956 Ceased WO2017140218A1 (zh) | 2016-02-18 | 2017-02-06 | 信息识别方法及装置 |
Country Status (3)
| Country | Link |
|---|---|
| CN (1) | CN107092903A (zh) |
| TW (1) | TWI673656B (zh) |
| WO (1) | WO2017140218A1 (zh) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108829464A (zh) * | 2018-06-25 | 2018-11-16 | 腾讯科技(深圳)有限公司 | 服务启动方法、装置、计算机设备及存储介质 |
| CN110287971A (zh) * | 2019-05-22 | 2019-09-27 | 平安银行股份有限公司 | 数据验证方法、装置、计算机设备及存储介质 |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101364267A (zh) * | 2007-08-09 | 2009-02-11 | 富士通株式会社 | 字符识别装置、字符识别方法及计算机产品 |
| CN101542504A (zh) * | 2006-09-08 | 2009-09-23 | 谷歌公司 | 后光学字符识别处理中的形状聚类 |
| CN101739441A (zh) * | 2009-12-01 | 2010-06-16 | 中国建设银行股份有限公司 | 一种图像信息录入的方法以及系统 |
| US20110222771A1 (en) * | 2010-03-11 | 2011-09-15 | Microsoft Corporation | Page layout determination of an image undergoing optical character recognition |
| CN102681986A (zh) * | 2012-05-23 | 2012-09-19 | 董名垂 | 页面即时翻译系统及页面即时翻译方法 |
| CN103617423A (zh) * | 2013-11-26 | 2014-03-05 | 觅林网络科技(上海)有限公司 | 基于色彩参数的图像拆分识别方法 |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6047251A (en) * | 1997-09-15 | 2000-04-04 | Caere Corporation | Automatic language identification system for multilingual optical character recognition |
| US8401293B2 (en) * | 2010-05-03 | 2013-03-19 | Microsoft Corporation | Word recognition of text undergoing an OCR process |
| TWI497425B (zh) * | 2010-11-18 | 2015-08-21 | Alibaba Group Holding Ltd | Method, apparatus and reptile server for digital image recognition |
| KR101860569B1 (ko) * | 2011-09-08 | 2018-07-03 | 삼성전자주식회사 | 문자 및 바코드를 동시에 인식하는 문자 및 바코드 인식 장치 및 그 제어 방법 |
| CN103488984B (zh) * | 2013-10-11 | 2017-04-12 | 瑞典爱立信有限公司 | 基于智能移动设备的二代身份证识别方法及装置 |
| CN104239853B (zh) * | 2014-08-27 | 2018-04-27 | 北京捷通华声语音技术有限公司 | 一种图像的处理方法和装置 |
-
2016
- 2016-02-18 CN CN201610091533.4A patent/CN107092903A/zh active Pending
- 2016-12-28 TW TW105143656A patent/TWI673656B/zh not_active IP Right Cessation
-
2017
- 2017-02-06 WO PCT/CN2017/072956 patent/WO2017140218A1/zh not_active Ceased
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101542504A (zh) * | 2006-09-08 | 2009-09-23 | 谷歌公司 | 后光学字符识别处理中的形状聚类 |
| CN101364267A (zh) * | 2007-08-09 | 2009-02-11 | 富士通株式会社 | 字符识别装置、字符识别方法及计算机产品 |
| CN101739441A (zh) * | 2009-12-01 | 2010-06-16 | 中国建设银行股份有限公司 | 一种图像信息录入的方法以及系统 |
| US20110222771A1 (en) * | 2010-03-11 | 2011-09-15 | Microsoft Corporation | Page layout determination of an image undergoing optical character recognition |
| CN102681986A (zh) * | 2012-05-23 | 2012-09-19 | 董名垂 | 页面即时翻译系统及页面即时翻译方法 |
| CN103617423A (zh) * | 2013-11-26 | 2014-03-05 | 觅林网络科技(上海)有限公司 | 基于色彩参数的图像拆分识别方法 |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108829464A (zh) * | 2018-06-25 | 2018-11-16 | 腾讯科技(深圳)有限公司 | 服务启动方法、装置、计算机设备及存储介质 |
| CN110287971A (zh) * | 2019-05-22 | 2019-09-27 | 平安银行股份有限公司 | 数据验证方法、装置、计算机设备及存储介质 |
| CN110287971B (zh) * | 2019-05-22 | 2023-11-14 | 平安银行股份有限公司 | 数据验证方法、装置、计算机设备及存储介质 |
Also Published As
| Publication number | Publication date |
|---|---|
| TW201732683A (zh) | 2017-09-16 |
| TWI673656B (zh) | 2019-10-01 |
| CN107092903A (zh) | 2017-08-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11210510B2 (en) | Storing anonymized identifiers instead of personally identifiable information | |
| US11349666B2 (en) | Electronically signing and distributing identification data as a service that provides proof of identity, integrity, validity and origin of data for non-repudiation and ID validation methods | |
| US12537702B2 (en) | System and method for fact verification using blockchain and machine learning technologies | |
| US20200159993A1 (en) | Methods, devices and systems for data augmentation to improve fraud detection | |
| AU2017410934B2 (en) | Detecting orientation of textual documents on a live camera feed | |
| CN108388822A (zh) | 一种检测二维码图像的方法和装置 | |
| WO2022142550A1 (zh) | 图像识别方法、装置和存储介质 | |
| CN114612919B (zh) | 一种票据信息处理系统及方法、装置 | |
| Bulatov et al. | Towards a unified framework for identity documents analysis and recognition | |
| US20230132261A1 (en) | Unified framework for analysis and recognition of identity documents | |
| CN105046289A (zh) | 一种文字域类型识别方法和文字域类型识别系统 | |
| US11023764B2 (en) | Method and system for optical character recognition of series of images | |
| WO2017140218A1 (zh) | 信息识别方法及装置 | |
| CN114386013A (zh) | 学籍自动认证方法、装置、计算机设备及存储介质 | |
| CN113936286A (zh) | 图像文本识别方法、装置、计算机设备及存储介质 | |
| US20170228272A1 (en) | System for processing data using different processing channels based on source error probability | |
| CN106250755B (zh) | 用于生成验证码的方法及装置 | |
| WO2021054850A1 (ru) | Способ и система интеллектуальной обработки документа | |
| CN112598029B (zh) | 一种ocr识别对抗样本攻击的方法和装置 | |
| RU2010121726A (ru) | Система и способ аутентификации банковской карты | |
| CN112070092A (zh) | 一种验证码参数的获取方法及装置 | |
| RU2739342C1 (ru) | Способ и система интеллектуальной обработки документа | |
| CN109359274B (zh) | 一种对批量生成的字符串进行识别的方法、装置及设备 | |
| CN116052199A (zh) | 一种基于ai的工程施工档案文件电子化方法及装置 | |
| Shetty et al. | Automated ID document extraction and classification: An integrated approach for efficient information retrieval from identity documents |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 17752650 Country of ref document: EP Kind code of ref document: A1 |
|
| NENP | Non-entry into the national phase |
Ref country code: DE |
|
| 122 | Ep: pct application non-entry in european phase |
Ref document number: 17752650 Country of ref document: EP Kind code of ref document: A1 |