WO2019140641A1 - 信息处理方法、系统、云处理设备以及计算机程序产品 - Google Patents

信息处理方法、系统、云处理设备以及计算机程序产品 Download PDF

Info

Publication number
WO2019140641A1
WO2019140641A1 PCT/CN2018/073442 CN2018073442W WO2019140641A1 WO 2019140641 A1 WO2019140641 A1 WO 2019140641A1 CN 2018073442 W CN2018073442 W CN 2018073442W WO 2019140641 A1 WO2019140641 A1 WO 2019140641A1
Authority
WO
WIPO (PCT)
Prior art keywords
template
image information
text
information
image
Prior art date
Application number
PCT/CN2018/073442
Other languages
English (en)
French (fr)
Inventor
廉士国
南一冰
Original Assignee
深圳前海达闼云端智能科技有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 深圳前海达闼云端智能科技有限公司 filed Critical 深圳前海达闼云端智能科技有限公司
Priority to CN201880000059.7A priority Critical patent/CN110291527B/zh
Priority to PCT/CN2018/073442 priority patent/WO2019140641A1/zh
Publication of WO2019140641A1 publication Critical patent/WO2019140641A1/zh

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00Pattern recognition

Definitions

  • the present application relates to the field of information processing technologies, and in particular, to an information processing method, system, cloud processing device, and computer program product.
  • insurance field--the insurance company realizes the risk assessment of the vehicle through the relevant information of the vehicle obtained from the vehicle network, the calculation of the vehicle premium, and the online loss determination. And other related business processing; traffic management field - the public security traffic management department analyzes the driving condition of the vehicle through the relevant information of the vehicle obtained from the vehicle network, timely discovers and manages the traffic violation, and correspondingly responds to the traffic accident deal with.
  • the system After the user uploads the specified system such as the vehicle related documents, the system needs to identify the file content. However, in the prior art, the accuracy of the file content identification is low.
  • the embodiment of the present application provides an information processing method, system, cloud processing device, and computer program product, which improve the recognition efficiency and accuracy of image information.
  • an embodiment of the present application provides an information processing method, including:
  • the embodiment of the present application further provides an information processing system, including:
  • An acquiring unit configured to acquire image information collected by the terminal
  • a matching unit configured to match the first template to the image information based on a template matching relationship
  • An extracting unit configured to extract a text area in the image information according to the first template
  • An identification unit for identifying characters in the text area.
  • the embodiment of the present application further provides a cloud processing device, where the device includes a processor and a memory; the memory is configured to store an instruction, when the instruction is executed by the processor, to cause the device to execute The method of any of the first aspects.
  • the embodiment of the present application further provides a computer program product, which can be directly loaded into an internal memory of a computer and includes software code, and the computer program can be implemented by being loaded and executed by a computer, as in the first aspect. Any of the methods described.
  • the information processing method, the system, the cloud processing device, and the computer program product provided by the embodiment of the present application process the image information collected by the acquired terminal by using the template matching relationship, and match the image information to the first template, and according to the first
  • the template extracts the text area in the image information, and finally recognizes the text in the text area.
  • the image information is matched to the template before the image information is recognized, and the image information is identified by the content of the template.
  • the recognition efficiency and accuracy of image information are improved, and the problem of low accuracy of document content recognition in the prior art is solved.
  • the difficult problem of "template matching" and "text recognition” is solved, and the sample recognition data can be continuously improved by manually intervening the sample data obtained.
  • FIG. 1 is a flowchart of an embodiment of an information processing method according to an embodiment of the present application
  • FIG. 2 is a schematic diagram of a first scenario provided by an embodiment of the present application.
  • FIG. 3 is a schematic diagram of a second scenario provided by an embodiment of the present application.
  • FIG. 4 is another flowchart of an embodiment of an information processing method according to an embodiment of the present disclosure.
  • FIG. 5 is another flowchart of an embodiment of an information processing method according to an embodiment of the present disclosure.
  • FIG. 6 is a schematic structural diagram of an embodiment of an information processing system according to an embodiment of the present disclosure.
  • FIG. 7 is another schematic structural diagram of an embodiment of an information processing system according to an embodiment of the present disclosure.
  • FIG. 8 is another schematic structural diagram of an embodiment of an information processing system according to an embodiment of the present disclosure.
  • FIG. 9 is a schematic structural diagram of an embodiment of a cloud processing device according to an embodiment of the present disclosure.
  • the word “if” as used herein may be interpreted as “when” or “when” or “in response to determining” or “in response to detecting.”
  • the phrase “if determined” or “if detected (conditions or events stated)” may be interpreted as “when determined” or “in response to determination” or “when detected (stated condition or event) “Time” or “in response to a test (condition or event stated)”.
  • the embodiment of the present application provides an information processing method, which is configured to initially identify an image collected by a terminal, and then perform content identification, thereby improving recognition efficiency.
  • FIG. 1 is provided in the embodiment of the present application. As shown in FIG. 1 , the information processing method in this embodiment may specifically include the following steps:
  • the user first uses the terminal to collect image information
  • the process of collecting may be using the camera or sensor of the terminal to collect image information of the quotation, the maintenance list, etc.
  • the collection method may be scanning the quotation and the maintenance list.
  • the illumination may be too bright, the illumination is too dark, the focus is not focused, the camera or the surface of the sensor is stained, etc., therefore, it is preferable to use the specified application installed in the terminal to collect image information
  • the specified application an image acquisition area can be displayed, for example, a capture frame.
  • the prompt information can also be output to ensure that the user can collect high-quality, high-definition image information.
  • terminals involved in the embodiments of the present application may include, but are not limited to, a personal computer (PC), a personal digital assistant (PDA), a wireless handheld device, a tablet computer, and a tablet computer.
  • PC personal computer
  • PDA personal digital assistant
  • Mobile phones MP3 players, MP4 players, etc.
  • the application may be an application (nativeApp) installed on the terminal, or may be a webpage (webApp) of the browser on the terminal, which is not limited by the embodiment of the present application.
  • the terminal After collecting the image information, the terminal can be uploaded to the cloud processing device, so that the cloud processing device can obtain the image information collected by the terminal.
  • the acquisition of the image collected by the terminal may also collect the positioning information of the terminal, that is, the image information includes the image collected by the terminal and the positioning information of the terminal, and the purpose of collecting the positioning information is to narrow the search template. Quantity. Since different repair shops/4S stores have different geographical locations, correspondingly, the template of the corresponding repair shop/4S shop can be quickly determined according to the geographical location of the terminal.
  • image information of a quotation, a maintenance list, and the like of different repair shops/4S stores are collected in advance, and a predefined template is created.
  • the purpose of setting a predefined template is to process the template content in advance, for example, to divide the area, determine part of the text information, etc., in the process of subsequently recognizing the text, the recognition range can be narrowed, the recognition speed is improved, and the recognition is significantly improved. The effect of efficiency.
  • the image information is compared with the predefined template according to the template matching relationship, and the similarity is determined.
  • the template matching relationship may be a template matching algorithm, and the matching process may be to collect the collected image information one by one.
  • the template matching algorithm may be an image copy detection or a map search method, and the template matching algorithm is used to determine the similarity, for example, using a fast image copy detection algorithm, specifically Firstly, the two-dimensional position information of the image feature points is extracted, and the distance and angle of each feature point and the image center point are calculated, and the number of feature points of each interval is calculated by block, and the binary hash sequence is quantized according to the quantity relationship to form a The level of robust features; then, according to the feature point one-dimensional distribution feature segmentation, the number of feature points in each direction sub-interval is counted, and the secondary image features are formed according to the quantitative relationship.
  • a fast image copy detection algorithm specifically Firstly, the two-dimensional position information of the image feature points is extracted, and the distance and angle of each feature point and the image center point are calculated, and the number of feature points of each interval is calculated by block, and the binary hash sequence is quantized according to the quantity relationship to form a The level of robust features; then, according to the feature point one-dimensional distribution feature
  • a cascading filtering framework is used for copy detection to determine similarity.
  • the similarity can be represented by a number, and the similarity takes a floating point value from 0 to 1. The larger the value, the higher the similarity. For example, the similarity between the acquired image information and the template A is 0.9.
  • the predefined template is determined to be the first template that matches the image information.
  • the first threshold is used to indicate that the similarity between the image information and the predefined template is high.
  • the content corresponding to the image information may be considered to be related to the content of the predefined template. Consistent.
  • the second threshold is used to indicate that the similarity between the image information and the predefined template is low, and the content corresponding to the image information does not match the content of any one of the predefined templates.
  • the cloud processing device sends the image information to the manual module, and the user who manipulates the manual module views the image information, and performs template definition processing on the image information.
  • the defined process includes determining the template name, breaking the template into text regions, and so on.
  • the image information and the first template are sent together to the manual module.
  • the cloud processing device may be considered to have an uncertainty in the image similarity, and needs manual assistance for processing.
  • the cloud processing device sends the image information and the first template to the manual module, and the user who manipulates the manual module views the image information. If the cloud processing device corrects the predefined template for matching the image information, the correct information is returned, if the cloud processing device is an image. If the predefined template of the information matching is incorrect, the template information is processed by the template definition process, and the definition process includes determining the template name, decomposing the template into a text area, and the like.
  • the system will receive the confirmation message returned by the manual template and update the template matching relationship.
  • the purpose of this can be to add a new training set to the algorithm with the help of the artificial, so that the algorithm is self-training and obtain a more accurate matching relationship.
  • the image information is subjected to a tilt correction process according to the first template; and then, according to the predefined extraction region in the first template, the extraction is performed.
  • the text area in the corrected image information is determined, in order to improve the accuracy of the extraction.
  • the process of the tilt correction may obtain a correspondence relationship by comparing key image points between the image information to be recognized and the first template, and then transform the template information according to the correspondence relationship to approximate the template, wherein the first template is
  • the standard forward angle by tilt correction, can adjust the image to be recognized of the non-forward angle to the direction of the first template, which is more advantageous for extracting the text area in the image information.
  • the text area in the image information is extracted. Since the template has previously divided the extraction area in the embodiment of the present application, when extracting the text area in the image information, the image is extracted based on the extracted area in the template. The corresponding text area is cut out in the information.
  • FIG. 2 is a schematic diagram of a first scenario according to an embodiment of the present disclosure.
  • FIG. 3 is a schematic diagram of a second scenario according to an embodiment of the present disclosure. As shown in FIG. 2, the image information before correction is as shown in FIG. 3 , which is Corrected image information.
  • the text in the recognized text area can be completed as follows:
  • the text area is binarized to obtain a first image; specifically, the gray level of the text area can be adjusted to be converted into black and white, and then the white color is removed and the black color is retained to obtain the first image.
  • character segmentation processing is performed on the first image to obtain at least one second image; specifically, the first image is projected in a vertical direction, and each character is distinguished according to the gray value.
  • each second image is subjected to text recognition to obtain a corresponding text.
  • a recognition result and a recognition confidence are given, wherein the recognition confidence is a floating point value from 0 to 1, and the greater the value, the higher the reliability of the recognition, and
  • the cloud processing device sends the character to the manual module, and the manual module provides the recognition result, which can solve the problem that “the upper template can be matched but the detailed text cannot be recognized”.
  • the artificially given recognition result will be used as the annotation data for the corresponding text area for retraining and improving the text recognition (based on the "text area picture" - "corresponding text annotation” data sample pair).
  • the information processing method provided by the embodiment of the present application processes the image information collected by the acquired terminal by using the template matching relationship, matches the first template to the image information, and extracts the text area in the image information according to the first template, and finally By recognizing the text in the text area, by using the technical solution provided by the embodiment of the present application, the image information is matched to the template before the image information is recognized, and the image information is identified by the content of the template, thereby improving the recognition efficiency of the image information. Accuracy solves the problem of low accuracy in document content recognition in the prior art. At the same time, by adaptively introducing manual intervention, the difficult problem of "template matching" and "text recognition” is solved, and the sample recognition data can be continuously improved by manually intervening the sample data obtained.
  • FIG. 4 is another flowchart of an embodiment of an information processing method according to an embodiment of the present disclosure. As shown in FIG. 4, the information processing method provided by the embodiment of the present application may further include the following steps:
  • a font library is pre-stored in the cloud processing device. For example, a large number of component nouns are stored in a text library for vehicle maintenance, and the character library is used for text based on the similarity of the string in practical applications. Make corrections, search the defined font library for the most similar to the current text, and replace the recognized text with the most similar text. For example, when the recognized text is "front bumper", it can be corrected to "front bumper” by the text library.
  • the text when the text recognition confidence is less than the third threshold, the text is sent to the manual module; the artificial module assists in recognizing the text, and gives the text recognition result or the correct text information, and then, manually
  • the module sends the text recognition result or the correct text information to the cloud processing device, and the cloud processing device receives the text information returned by the manual module, and uses the text and the corresponding text information as a new training sample to update the text recognition training set. Used to update the text recognition algorithm.
  • FIG. 5 is another flowchart of the information processing method provided by the embodiment of the present application, as shown in FIG. 5 .
  • the information processing method provided by the embodiment of the present application may further include the following steps:
  • the image information fails to match the first template, the image information is sent to the manual module.
  • the new template is manually added, and the template matching relationship is adjusted according to the template.
  • the method provided by the embodiment of the present application it is equivalent to adding more samples, and the algorithm is performed by more samples. Training can help improve the accuracy and accuracy of the algorithm.
  • the participation of the artificial module can help improve various databases and template libraries, and can also increase the recognition efficiency of the algorithm while assisting the artificial intelligence algorithm to make recognition and judgment, and can continuously improve the recognition efficiency and accuracy.
  • FIG. 6 is a schematic structural diagram of an embodiment of an information processing system according to an embodiment of the present disclosure.
  • the acquisition unit 11 includes a retrieval unit 11, an extraction unit 13, and an identification unit 14.
  • the obtaining unit 11 is configured to acquire image information collected by the terminal.
  • the matching unit 12 is configured to match the first template to the image information based on the template matching relationship.
  • the extracting unit 13 is configured to extract a text area in the image information according to the first template.
  • the identification unit 14 is configured to recognize characters in the text area.
  • the matching unit 12 is specifically configured to:
  • the predefined template is the first template that matches the image information.
  • the image information includes a picture collected by the terminal and positioning information of the terminal;
  • matching the first template for image information includes:
  • FIG. 7 is another schematic structural diagram of an embodiment of an information processing system according to an embodiment of the present disclosure. As shown in FIG. 7, the system in this embodiment may further include: a receiving unit 15 and an updating unit 16 on the basis of the foregoing content.
  • the matching unit 12 is further configured to:
  • the image information and the first template are sent together to the manual module;
  • the receiving unit 15 is configured to receive the confirmation information returned by the manual module.
  • the updating unit 16 is configured to update the template matching relationship.
  • the matching unit 12 is further configured to:
  • the image information is sent to the manual module
  • the receiving unit 15 is further configured to receive a second template returned by the manual module
  • the updating unit 16 is configured to update the template matching relationship.
  • the matching unit 12 is further configured to:
  • the similarity is less than the second threshold, it is determined that the predefined template does not match the image information, and the image information is transmitted to the manual module.
  • FIG. 8 is another schematic structural diagram of an embodiment of an information processing system according to an embodiment of the present disclosure. As shown in FIG. 8 , the system in this embodiment may further include: a correction unit 17 on the basis of the foregoing content.
  • the correcting unit 17 is configured to correct the text.
  • the extracting unit 13 is specifically configured to:
  • the text area in the corrected image information is extracted according to a predefined extraction area in the first template.
  • the identification unit 14 is specifically configured to:
  • Text recognition is performed for each second image to obtain a corresponding text.
  • the identification unit 14 is further configured to:
  • FIG. 9 is a schematic structural diagram of an embodiment of a cloud processing device according to an embodiment of the present disclosure.
  • the cloud processing device provided by the embodiment of the present disclosure may specifically include: a processor 21 and a memory 22 .
  • the memory 21 is used to store instructions that, when executed by the processor 22, cause the device to perform any of the methods shown in FIGS. 1 through 5.
  • the embodiment of the present application further provides a computer program product, which can be directly loaded into the internal memory of the computer and contains software code. After the computer program is loaded and executed by the computer, any method as shown in FIG. 1 to FIG. 5 can be implemented. .
  • the information processing system, the cloud processing device, and the computer program product of this embodiment may be used to implement the technical solution of the method embodiment shown in FIG. 1 to FIG. 5, and the implementation principle and technical effects are similar, and details are not described herein again.
  • the aforementioned program can be stored in a computer readable storage medium.
  • the program when executed, performs the steps including the foregoing method embodiments; and the foregoing storage medium includes various media that can store program codes, such as a ROM, a RAM, a magnetic disk, or an optical disk.
  • the device embodiments described above are merely illustrative, wherein the units illustrated as separate components may or may not be physically separate, and the components displayed as units may or may not be physical units, ie may be located in one place. Or it can be distributed to at least two network elements. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the embodiment. Those of ordinary skill in the art can understand and implement without undue creative work.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Artificial Intelligence (AREA)
  • Character Input (AREA)

Abstract

本申请实施例提供一种信息处理方法、系统、云处理设备以及计算机程序产品,涉及信息处理技术领域,在一定程度上提高了对图像信息的识别效率与准确性。本申请实施例提供的信息处理方法,包括:获取终端采集的图像信息;基于模板匹配关系,为所述图像信息匹配第一模板;根据所述第一模板,提取所述图像信息中的文字区域;识别所述文字区域内的文字。

Description

信息处理方法、系统、云处理设备以及计算机程序产品 技术领域
本申请涉及信息处理技术领域,尤其涉及一种信息处理方法、系统、云处理设备以及计算机程序产品。
背景技术
随着车联网相关技术的不断成熟,传感器技术、移动通信技术、大数据技术和智能计算技术等均开始与车联网深度融合。在市场需求带动下,区别于传统的交通系统,车联网更加注重车与车、车与路、车与人之间的交互通信,可以说车联网的出现重新定义了车辆交通的运行方式。
现有技术中,基于车联网的应用主要体现在如下几个领域:保险领域—保险公司通过从车联网中获取到的车辆的相关信息,实现对车辆的风险评估,车辆保费计算、在线定损等相关业务处理;交通管理领域—公安交通管理部门通过从车联网中获取到的车辆的相关信息,对车辆的行驶状况进行分析,对交通违法行为进行及时发现和治理,以及对交通事故进行相应处理。
用户将于车辆相关的单据等文件上传指定系统后,系统需要对文件内容进行识别,然而,现有技术中,对文件内容识别的准确率较低。
发明内容
本申请实施例提供一种信息处理方法、系统、云处理设备以及计算机程序产品,提高了对图像信息的识别效率与准确性。
第一方面,本申请实施例提供了一种信息处理方法,包括:
获取终端采集的图像信息;
基于模板匹配关系,为所述图像信息匹配第一模板;
根据所述第一模板,提取所述图像信息中的文字区域;
识别所述文字区域内的文字。
第二方面,本申请实施例还提供了一种信息处理系统,包括:
获取单元,用于获取终端采集的图像信息;
匹配单元,用于基于模板匹配关系,为所述图像信息匹配第一模板;
提取单元,用于根据所述第一模板,提取所述图像信息中的文字区域;
识别单元,用于识别所述文字区域内的文字。
第三方面,本申请实施例还提供了一种云处理设备,所述设备包括处理器以及存储器;所述存储器用于存储指令,所述指令被所述处理器执行时,使得所述设备执行如第一方面中任一种所述的方法。
第四方面,本申请实施例还提供了一种计算机程序产品,可直接加载到计算机的内部存储器中,并含有软件代码,所述计算机程序经由计算机载入并执行后能够实现如第一方面中任一种所述的方法。
本申请实施例提供的信息处理方法、系统、云处理设备以及计算机程序产品,通过基于模板匹配关系,对获取到的终端采集的图像信息进行处理,为图像信息匹配第一模板,并根据第一模板提取图像信息中的文字区域,最后识别文字区域内的文字,通过采用本申请实施例提供的技术方案,在对图像信息识别之前先为图像信息匹配模板,依靠模板的内容对图像信息进行识别,提高了对图像信息的识别效率与准确性,解决了现有技术中对文件内容识别的准确率较低的问题。同时通过自适应地引入人工介入,解决较难识别的“模板匹配”、“文字识别”问题,并且能够通过人工介入获取的样本数据来持续提升文字识别算法。
附图说明
为了更清楚地说明本申请实施例或现有技术中的技术方案,下面将对实施例或现有技术描述中所需要使用的附图作一简单地介绍,显而易见地,下面描述中的附图是本申请的一些实施例,对于本领域普通技术人员来讲,在不付出创造性劳动性的前提下,还可以根据这些附图获得其他的附图。
图1为本申请实施例提供的信息处理方法实施例的流程图;
图2为本申请实施例提供的第一场景示意图;
图3为本申请实施例提供的第二场景示意图;
图4为本申请实施例提供的信息处理方法实施例的另一流程图;
图5为本申请实施例提供的信息处理方法实施例的另一流程图;
图6为本申请实施例提供的信息处理系统实施例的结构示意图;
图7为本申请实施例提供的信息处理系统实施例的另一结构示意图;
图8为本申请实施例提供的信息处理系统实施例的另一结构示意图;
图9为本申请实施例提供的云处理设备实施例的结构示意图。
具体实施方式
为使本申请实施例的目的、技术方案和优点更加清楚,下面将结合本申请实施例中的附图,对本申请实施例中的技术方案进行清楚、完整地描述,显然,所描述的实施例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。
在本申请实施例中使用的术语是仅仅出于描述特定实施例的目的,而非旨在限制本申请。在本申请实施例和所附权利要求书中所使用的单数形式的“一种”、“所述”和“该”也旨在包括多数形式,除非上下文清楚地表示其他含义。
应当理解,本文中使用的术语“和/或”仅仅是一种描述关联对象的关联关系,表示可以存在三种关系,例如,A和/或B,可以表示:单独存在A,同时存在A和B,单独存在B这三种情况。另外,本文中字符“/”,一般表示前后关联对象是一种“或”的关系。
取决于语境,如在此所使用的词语“如果”可以被解释成为“在……时”或“当……时”或“响应于确定”或“响应于检测”。类似地,取决于语境,短语“如果确定”或“如果检测(陈述的条件或事件)”可以被解释成为“当确定时”或“响应于确定”或“当检测(陈述的条件或事件)时”或“响应于检测(陈述的条件或事件)”。
随着家用汽车的普及,汽车数量不断增加,车辆出现事故的几率也在不断的增加,这给保险行业的工作带来了巨大压力。通常情况下,车辆出险的具体操作是由保险理赔业务员将维修厂/4S店的报价单、维修清单等录入理赔系统中。由于报价单、维修清单等多为纸质材料,由人工手写相关的内容,因此,就需要人工手动将报价单、维修清单等录入理赔系统。而手动录入的工作量大、很耗时、也耗人力。由于车联网相关技术,将保险公司与汽车联系起来,使得用户自行将报价单、维修清单等上传至理赔系统成为可能。但是,由于不同的维修厂/4S店的报价单、维修清单等均会不同,使用现有技术中的方法,在将报价单、维修清单等录入理赔系统后,对报价单、维修清单等中的内容进行识别的识别效率很低。因此,本申请实施例提供了一种信息处理方法,通过模板来初步对终端采集到的图像进行识别后,再进行内容的识别,提高了识别效率,具体的,图1为本申请实施例提供的信息处理方法实施例的流程图,如图1所示,本实施例的信息处理方法,具体可以包括如下步骤:
101、获取终端采集的图像信息。
在本申请实施例中,首先由用户使用终端采集图像信息,采集的过程 可以是使用终端的摄像头或者传感器采集报价单、维修清单等内容的图像信息,采集的方式可以是扫描报价单、维修清单等,还可以是为报价单、维修清单等进行拍照。
由于用户在采集图像信息的过程中,可能会出现光照过亮、光照过暗、不聚焦、摄像头或者传感器表面有污渍等情况,因此,优选使用安装在终端中的指定应用来采集图像信息,在指定应用中,可以显示图像采集区域,例如,一个采集框。为了进一步的提高图像识别效率,在指定应用中,还可以输出提示信息,确保用户能够采集到高质量、高清晰度的图像信息。
需要说明的是,本申请实施例中所涉及的终端可以包括但不限于个人计算机(Personal Computer,PC)、个人数字助理(Personal Digital Assistant,PDA)、无线手持设备、平板电脑(Tablet Computer)、手机、MP3播放器、MP4播放器等。
可以理解的是,应用可以是安装在终端上的应用程序(nativeApp),或者还可以是终端上的浏览器的一个网页程序(webApp),本申请实施例对此不进行限定。
终端在采集图像信息后,可以上传至云处理设备中,使得云处理设备可以获取到终端采集的图像信息。
在本申请实施例中,在获取终端采集的图像的同时,还可以采集终端的定位信息,即图像信息包括终端采集的图片和终端的定位信息,采集定位信息的目的,在于可以缩小检索模板的数量。由于不同的维修厂/4S店具有不同的地理位置,相应的,可以根据终端的地理位置,快速确定对应的维修厂/4S店的模板。
102、基于模板匹配关系,为图像信息匹配第一模板。
在本申请实施例中,会预先采集不同的维修厂/4S店的报价单、维修清单等的图像信息,制作成预定义模板。设定预定义模板的目的在于可以预 先对模板内容进行处理,例如,划分区域、确定部分文字信息等,在后续识别文字的过程中,能够缩小识别范围,提高识别速度,起到显著的提高识别效率的效果。
具体的,首先,根据模板匹配关系将图像信息与预定义模板进行对比,确定相似度;在本申请实施例中,模板匹配关系可以是模板匹配算法,匹配过程可以是将采集到的图像信息逐一与预定义模板进行对比,在一个具体的实现过程中,模板匹配算法可以是图像拷贝检测或以图搜图的方法,通过模板匹配算法来确定相似度,例如,采用快速图像拷贝检测算法,具体的,首先提取图像特征点的二维位置信息,通过计算各个特征点与图像中心点的距离、角度,分块统计各区间的特征点数量,依据数量关系量化生成二值哈希序列,构成一级鲁棒特征;然后,根据特征点一维方向分布特征分块统计各方向子区间特征点数量,依据数量关系构成二级图像特征。最后,拷贝检测时采用级联式过滤框架,确定相似度。在一个具体的实现过程中,相似度可以用数字表示,相似度从0到1取浮点数值,值越大相似度越高,例如,采集到的图像信息与模板A的相似度为0.9。
其次,当相似度大于或者等于第一阈值时,确定预定义模板为与图像信息匹配的第一模板。在本申请实施例中,第一阈值用于表示图像信息与预定义模板的相似度高,当相似度大于或者等于第一阈值时,即可以认为图像信息对应的内容与预定义模板的内容相一致。
当相似度小于第二阈值时,确定预定义模板与图像信息不匹配,将图像信息发送至人工模块。在本申请实施例中,第二阈值用于表示图像信息与预定义模板的相似度低,图像信息对应的内容与任意一个预定义模板的内容均不匹配,当相似度小于第二阈值时,即可以认为预定义模板没有与图像信息相匹配的,则需要人工协助进行处理,云处理设备将图像信息发送至人工模块,操控人工模块的用户则查看图像信息,对图像信息进行模 板定义处理,定义的过程包括,确定模板名称、将模板分解成文字区域等。
当相似度大于或者等于第二阈值,且小于第一阈值时,将图像信息以及第一模板一并发送至人工模块。在本申请实施例中,当相似度大于或者等于第二阈值,且小于第一阈值时,即可以认为云处理设备对图像相似度的识别存在不确定性,需要人工协助进行处理。云处理设备将图像信息以及第一模板发送至人工模块,操控人工模块的用户则查看图像信息,如果云处理设备为图像信息匹配的预定义模板正确,则返回正确信息,如果云处理设备为图像信息匹配的预定义模板不正确,则,对图像信息进行模板定义处理,定义的过程包括,确定模板名称、将模板分解成文字区域等。
系统将接收人工模板返回的确认消息,以及更新模板匹配关系。这样做的目的,可以由人工的辅助下,为算法增加新的训练集,使得算法自训练,得到更加精准的匹配关系。
103、根据第一模板,提取图像信息中的文字区域。
在本申请实施例中,当确定了第一模板后,为了提高提取的准确性,首先,根据第一模板对图像信息进行倾斜校正处理;然后,根据第一模板中预定义的提取区域,提取校正后的图像信息中的文字区域。
具体的,倾斜校正的过程可以通过比对待识别的图像信息和第一模板间的关键特征点来获得对应关系,然后基于此对应关系对待识别图像信息做变换来逼近模板,其中,第一模板是标准正向角度,通过倾斜校正,可以将非正向角度的待识别图像调整为第一模板的方向,更有利于提取图像信息中的文字区域。在倾斜校正之后,提取图像信息中的文字区域,由于在本申请实施例中,模板已经预先划分了提取区域,因此,在提取图像信息中的文字区域时,基于模板中的提取区域,从图像信息中切分出对应的文字区域。图2为本申请实施例提供的第一场景示意图,图3为本申请实施例提供的第二场景示意图,如图2所示,其为校正前的图像信息,如图3 所示,其为校正后的图像信息。
104、识别文字区域内的文字。
在本申请实施例中,识别文字区域内的文字可以通过如下方式来完成:
首先,对文字区域进行二值化处理得到第一图像;具体的,可以对文字区域的灰度进行调整,使其转化为黑色和白色,然后,将白色去掉,黑色保留,得到第一图像。
其次,对第一图像进行字符分割处理,得到至少一个第二图像;具体的,将第一图像在垂直方向做投影,根据灰度值区分每一个字符。
最后,对每个第二图像进行文字识别,得到对应的文字。具体的,在识别过程中,对于每个字符的识别,都会给出识别结果和识别置信度,其中,识别置信度从0到1取浮点数值,值越大识别的可靠性越高,并且当置信度低于某个阈值时,识别结果不可靠,则云处理设备将该字符发送至人工模块,由人工模块辅助给出识别结果,可以解决“能够匹配上模板但无法识别出细节文字”的问题。此人工给出的识别结果将作为对对应文字区域的标注数据,用于重新训练提升文字识别(基于“文字区域图片”-“对应的文字标注”数据样本对)。
本申请实施例提供的信息处理方法,通过基于模板匹配关系,对获取到的终端采集的图像信息进行处理,为图像信息匹配第一模板,并根据第一模板提取图像信息中的文字区域,最后识别文字区域内的文字,通过采用本申请实施例提供的技术方案,在对图像信息识别之前先为图像信息匹配模板,依靠模板的内容对图像信息进行识别,提高了对图像信息的识别效率与准确性,解决了现有技术中对文件内容识别的准确率较低的问题。同时通过自适应地引入人工介入,解决较难识别的“模板匹配”、“文字识别”问题,并且能够通过人工介入获取的样本数据来持续提升文字识别算法。
由于人工书写存在连字、简笔、不规整等情况,对于文字的识别正确性具有一定的误差率,因此,在前述内容的基础上,本申请实施例还可以具有如下方式来解决这个问题,具体的,图4为本申请实施例提供的信息处理方法实施例的另一流程图,如图4所示,本申请实施例提供的信息处理方法,还可以包括如下步骤:
105、对文字进行校正。
在本申请实施例中,云处理设备中会预先存储文字库,例如,针对车辆维修的文字库中存储大量零部件名词,在实际应用中基于字符串的相似度的方式,利用文字库对文字进行校正,在已定义的文字库中搜索与当前文字最相似的,并用最相似的文字替代识别出的文字。例如,当识别出的文字是“车前保脸杠”时,通过文字库可以将其校正为“车前保险杠”。
在本发明实施例中,当文字识别置信度小于第三阈值时,发送所述文字至人工模块;由人工模块辅助对文字进行识别,并给出文字识别结果或者正确的文字信息,然后,人工模块将文字识别结果或者正确的文字信息发送至云处理设备,云处理设备接收人工模块返回的文字信息,将该文字以及对应的文字信息一并作为新的训练样本,更新文字识别训练集,以用于后续更新文字识别算法。通过采用本申请实施例的技术方案,进一步的提高识别图像信息中文字的准确率。
在前述内容的基础上,本申请实施例提供的技术方案中,还可以包括如下步骤,具体的,图5为本申请实施例提供的信息处理方法实施例的另一流程图,如图5所示,本申请实施例提供的信息处理方法,还可以包括如下步骤:
106、当未能为图像信息匹配第一模板时,将图像信息发送至人工模块。
107、接收人工模块返回的第二模板。
108、更新模板匹配关系。
可以理解的是,人工增加的新的模板,并根据模板相应的调整模板匹配关系,对于本申请实施例提供的方法来说,相当于增加了更多的样本,通过更多的样本对算法进行训练,能够有利于提高算法的精度和准确性。
在本申请实施例中,人工模块的参与,可以帮助完善各种数据库,以及模板库,在辅助人工智能算法做出识别判断的同时还能增加对算法的训练样本,能够不断的提高识别效率和准确性。
为了实现前述内容中的方法,本申请实施例还提供一种信息处理系统,图6为本申请实施例提供的信息处理系统实施例的结构示意图,如图6所示,本实施例的系统可以包括:获取单元11、匹配单元12、提取单元13和识别单元14。
获取单元11,用于获取终端采集的图像信息。
匹配单元12,用于基于模板匹配关系,为图像信息匹配第一模板。
提取单元13,用于根据第一模板,提取图像信息中的文字区域。
识别单元14,用于识别文字区域内的文字。
在一个具体的实现过程中,匹配单元12,具体用于:
根据模板匹配关系将图像信息与预定义模板进行对比,确定相似度;
当相似度大于或者等于第一阈值时,确定预定义模板为与图像信息匹配的第一模板。
在一个具体的实现过程中,图像信息包括终端采集的图片和终端的定位信息;
基于模板匹配关系,为图像信息匹配第一模板包括:
基于定位信息和模板匹配关系,为所采集的图片匹配第一模板。
图7为本申请实施例提供的信息处理系统实施例的另一结构示意图,如图7所示,本实施例的系统在前述内容的基础上,还可以包括:接收单元15和更新单元16。
在一个具体的实现过程中,匹配单元12,还用于:
根据模板匹配关系将图像信息与预定义模板进行对比,确定相似度;
当相似度小于第一阈值,且大于或者等于第二阈值时,将图像信息以及第一模板一并发送至人工模块;
接收单元15,用于接收人工模块返回的确认信息。
更新单元16,用于更新模板匹配关系。
在另一个具体的实现过程中,匹配单元12,还用于:
当未能为图像信息匹配第一模板时,将图像信息发送至人工模块;
接收单元15,还用于接收人工模块返回的第二模板;
更新单元16,用于更新模板匹配关系。
在另一个具体的实现过程中,匹配单元12,还用于:
当相似度小于第二阈值时,确定预定义模板与图像信息不匹配,将图像信息发送至人工模块。
图8为本申请实施例提供的信息处理系统实施例的另一结构示意图,如图8所示,本实施例的系统在前述内容的基础上,还可以包括:校正单元17。
校正单元17,用于对文字进行校正。
在一个具体的实现过程中,提取单元13,具体用于:
根据第一模板对图像信息进行倾斜校正处理;
根据第一模板中预定义的提取区域,提取校正后的图像信息中的文字区域。
在一个具体的实现过程中,识别单元14,具体用于:
对文字区域进行二值化处理得到第一图像;
对第一图像进行字符分割处理,得到至少一个第二图像;
对每个第二图像进行文字识别,得到对应的文字。
在一个具体的实现过程中,识别单元14,还用于:
当文字识别置信度小于第三阈值时,发送所述文字至人工模块;
接收所述人工模块返回的文字信息;
更新文字识别训练集。
图9为本申请实施例提供的云处理设备实施例的结构示意图,如图9所示,本申请实施例提供的云处理设备,具体可以包括:处理器21以及存储器22。
其中,存储器21用于存储指令,指令被处理器22执行时,使得设备执行如图1至图5所示任意一种方法。
本申请实施例还提供一种计算机程序产品,可直接加载到计算机的内部存储器中,并含有软件代码,计算机程序经由计算机载入并执行后能够实现如图1至图5所示任意一种方法。
本实施例的信息处理系统、云处理设备以及计算机程序产品,可以用于执行图1至图5所示方法实施例的技术方案,其实现原理和技术效果类似,此处不再赘述。
本领域普通技术人员可以理解:实现上述各方法实施例的全部或部分步骤可以通过程序指令相关的硬件来完成。前述的程序可以存储于一计算机可读取存储介质中。该程序在执行时,执行包括上述各方法实施例的步骤;而前述的存储介质包括:ROM、RAM、磁碟或者光盘等各种可以存储程序代码的介质。
以上所描述的装置实施例仅仅是示意性的,其中作为分离部件说明的单元可以是或者也可以不是物理上分开的,作为单元显示的部件可以是或者也可以不是物理单元,即可以位于一个地方,或者也可以分布到至少两个网络单元上。可以根据实际的需要选择其中的部分或者全部模块来实现本实施例方案的目的。本领域普通技术人员在不付出创造性的劳动的情况 下,即可以理解并实施。
最后应说明的是:以上各实施例仅用以说明本申请的技术方案,而非对其限制;尽管参照前述各实施例对本申请进行了详细的说明,本领域的普通技术人员应当理解:其依然可以对前述各实施例所记载的技术方案进行修改,或者对其中部分或者全部技术特征进行等同替换;而这些修改或者替换,并不使相应技术方案的本质脱离本申请各实施例技术方案的范围。

Claims (22)

  1. 一种信息处理方法,其特征在于,包括:
    获取终端采集的图像信息;
    基于模板匹配关系,为所述图像信息匹配第一模板;
    根据所述第一模板,提取所述图像信息中的文字区域;
    识别所述文字区域内的文字。
  2. 根据权利要求1所述的方法,其特征在于,所述基于模板匹配关系,为所述图像信息匹配第一模板,包括:
    根据模板匹配关系将所述图像信息与预定义模板进行对比,确定相似度;
    当所述相似度大于或者等于第一阈值时,确定所述预定义模板为与所述图像信息匹配的第一模板。
  3. 根据权利要求1或2所述的方法,其特征在于,所述方法还包括:
    根据模板匹配关系将所述图像信息与预定义模板进行对比,确定相似度;
    当所述相似度小于第一阈值,且大于或者等于第二阈值时,将所述图像信息以及所述第一模板一并发送至人工模块;
    接收所述人工模块返回的确认信息;
    更新所述模板匹配关系。
  4. 根据权利要求2所述的方法,其特征在于,所述方法还包括:
    当未能为所述图像信息匹配第一模板时,将所述图像信息发送至人工模块;
    接收所述人工模块返回的第二模板;
    更新所述模板匹配关系。
  5. 根据权利要求4所述的方法,其特征在于,所述当未能为所述图像 信息匹配第一模板时,将所述图像信息发送至人工模块,包括:
    当所述相似度小于第二阈值时,确定所述预定义模板与所述图像信息不匹配,将所述图像信息发送至人工模块。
  6. 根据权利要求1所述的方法,其特征在于,所述根据所述第一模板,提取所述图像信息中的文字区域,包括:
    根据所述第一模板对所述图像信息进行倾斜校正处理;
    根据所述第一模板中预定义的提取区域,提取校正后的所述图像信息中的文字区域。
  7. 根据权利要求1所述的方法,其特征在于,所述识别所述文字区域内的文字,包括:
    对所述文字区域进行二值化处理得到第一图像;
    对所述第一图像进行字符分割处理,得到至少一个第二图像;
    对每个所述第二图像进行文字识别,得到对应的文字。
  8. 根据权利要求7所述的方法,其特征在于,所述方法还包括:
    当文字识别置信度小于第三阈值时,发送所述文字至人工模块;
    接收所述人工模块返回的文字信息;
    更新文字识别训练集。
  9. 根据权利要求1所述的方法,其特征在于,所述图像信息包括终端采集的图片和所述终端的定位信息;
    所述基于模板匹配关系,为所述图像信息匹配第一模板包括:
    基于所述定位信息和模板匹配关系,为所采集的图片匹配第一模板。
  10. 根据权利要求1所述的方法,其特征在于,所述方法还包括:
    对所述文字进行校正。
  11. 一种信息处理系统,其特征在于,包括:
    获取单元,用于获取终端采集的图像信息;
    匹配单元,用于基于模板匹配关系,为所述图像信息匹配第一模板;
    提取单元,用于根据所述第一模板,提取所述图像信息中的文字区域;
    识别单元,用于识别所述文字区域内的文字。
  12. 根据权利要求11所述的系统,其特征在于,所述系统还包括:
    校正单元,用于对所述文字进行校正。
  13. 根据权利要求11所述的系统,其特征在于,所述匹配单元,具体用于:
    根据模板匹配关系将所述图像信息与预定义模板进行对比,确定相似度;
    当所述相似度大于或者等于第一阈值时,确定所述预定义模板为与所述图像信息匹配的第一模板。
  14. 根据权利要求11或13所述的系统,其特征在于,所述匹配单元,还用于:
    根据模板匹配关系将所述图像信息与预定义模板进行对比,确定相似度;
    当所述相似度小于第一阈值,且大于或者等于第二阈值时,将所述图像信息以及所述第一模板一并发送至人工模块;
    所述系统还包括:
    接收单元,用于接收所述人工模块返回的确认信息;
    更新单元,用于更新所述模板匹配关系。
  15. 根据权利要求13所述的系统,其特征在于所述匹配单元,还用于:
    当未能为所述图像信息匹配第一模板时,将所述图像信息发送至人工模块;
    所述系统还包括:
    接收单元,用于接收所述人工模块返回的第二模板;
    更新单元,用于更新所述模板匹配关系。
  16. 根据权利要求15所述的系统,其特征在于,所述匹配单元,还用于:
    当所述相似度小于第二阈值时,确定所述预定义模板与所述图像信息不匹配,将所述图像信息发送至人工模块。
  17. 根据权利要求11所述的系统,其特征在于,所述提取单元,具体用于:
    根据所述第一模板对所述图像信息进行倾斜校正处理;
    根据所述第一模板中预定义的提取区域,提取校正后的所述图像信息中的文字区域。
  18. 根据权利要求11所述的系统,其特征在于,所述识别单元,具体用于:
    对所述文字区域进行二值化处理得到第一图像;
    对所述第一图像进行字符分割处理,得到至少一个第二图像;
    对每个所述第二图像进行文字识别,得到对应的文字。
  19. 根据权利要求18所述的系统,其特征在于,所述识别单元,还用于:
    当文字识别置信度小于第三阈值时,发送所述文字至人工模块;
    接收所述人工模块返回的文字信息;
    更新文字识别训练集。
  20. 根据权利要求11所述的系统,其特征在于,所述图像信息包括终端采集的图片和所述终端的定位信息;
    所述基于模板匹配关系,为所述图像信息匹配第一模板包括:
    基于所述定位信息和模板匹配关系,为所采集的图片匹配第一模板。
  21. 一种云处理设备,其特征在于,所述设备包括处理器以及存储器; 所述存储器用于存储指令,所述指令被所述处理器执行时,使得所述设备执行如权利要求1~10中任一种所述的方法。
  22. 一种计算机程序产品,其特征在于,可直接加载到计算机的内部存储器中,并含有软件代码,所述计算机程序经由计算机载入并执行后能够实现如权利要求1~10中任一种所述的方法。
PCT/CN2018/073442 2018-01-19 2018-01-19 信息处理方法、系统、云处理设备以及计算机程序产品 WO2019140641A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
CN201880000059.7A CN110291527B (zh) 2018-01-19 2018-01-19 信息处理方法、系统、云处理设备以及计算机程序产品
PCT/CN2018/073442 WO2019140641A1 (zh) 2018-01-19 2018-01-19 信息处理方法、系统、云处理设备以及计算机程序产品

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2018/073442 WO2019140641A1 (zh) 2018-01-19 2018-01-19 信息处理方法、系统、云处理设备以及计算机程序产品

Publications (1)

Publication Number Publication Date
WO2019140641A1 true WO2019140641A1 (zh) 2019-07-25

Family

ID=67301874

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/073442 WO2019140641A1 (zh) 2018-01-19 2018-01-19 信息处理方法、系统、云处理设备以及计算机程序产品

Country Status (2)

Country Link
CN (1) CN110291527B (zh)
WO (1) WO2019140641A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111444397A (zh) * 2020-05-10 2020-07-24 黄菊 一种人工智能数据收集系统
CN111639636A (zh) * 2020-05-29 2020-09-08 北京奇艺世纪科技有限公司 一种字符识别方法及装置

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112001381B (zh) * 2020-09-14 2023-09-19 中国银行股份有限公司 一种智能预填单审核方法及装置
CN113221904A (zh) * 2021-05-13 2021-08-06 北京惠朗时代科技有限公司 一种语义关联文字识别方法及装置

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20090087103A1 (en) * 2007-09-28 2009-04-02 Hitachi High-Technologies Corporation Inspection Apparatus and Method
CN104916034A (zh) * 2015-06-09 2015-09-16 普联软件股份有限公司 一种基于可干预模板的票据识别系统及识别方法
CN107437085A (zh) * 2017-08-09 2017-12-05 厦门商集企业咨询有限责任公司 一种提升ocr识别率的方法、装置及可读存储介质

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105096244B (zh) * 2015-07-02 2019-02-22 北京旷视科技有限公司 图像变换方法及其装置以及图像识别方法及其装置

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20090087103A1 (en) * 2007-09-28 2009-04-02 Hitachi High-Technologies Corporation Inspection Apparatus and Method
CN104916034A (zh) * 2015-06-09 2015-09-16 普联软件股份有限公司 一种基于可干预模板的票据识别系统及识别方法
CN107437085A (zh) * 2017-08-09 2017-12-05 厦门商集企业咨询有限责任公司 一种提升ocr识别率的方法、装置及可读存储介质

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111444397A (zh) * 2020-05-10 2020-07-24 黄菊 一种人工智能数据收集系统
CN111639636A (zh) * 2020-05-29 2020-09-08 北京奇艺世纪科技有限公司 一种字符识别方法及装置

Also Published As

Publication number Publication date
CN110291527B (zh) 2023-03-24
CN110291527A (zh) 2019-09-27

Similar Documents

Publication Publication Date Title
WO2019238063A1 (zh) 文本检测分析方法、装置及设备
WO2019140641A1 (zh) 信息处理方法、系统、云处理设备以及计算机程序产品
WO2019169532A1 (zh) 车牌识别方法及云系统
WO2018233038A1 (zh) 基于深度学习的车牌识别方法、装置、设备及存储介质
WO2018166116A1 (zh) 车损识别方法、电子装置及计算机可读存储介质
US8064703B2 (en) Property record document data validation systems and methods
US20070217692A1 (en) Property record document data verification systems and methods
WO2020038138A1 (zh) 样本标注方法及装置、损伤类别的识别方法及装置
WO2014092979A1 (en) Method of perspective correction for devanagari text
WO2020164278A1 (zh) 一种图像处理方法、装置、电子设备和可读存储介质
CN108323209B (zh) 信息处理方法、系统、云处理设备以及计算机存储介质
CN111414905B (zh) 一种文本检测方法、文本检测装置、电子设备及存储介质
CN112733666A (zh) 一种难例图像的搜集、及模型训练方法、设备及存储介质
CN111222409A (zh) 车辆品牌的标注方法、装置及系统
CN114550051A (zh) 一种车损检测方法、装置、计算机设备及存储介质
CN116612479A (zh) 一种轻量级的票据ocr识别方法及系统
CN112232336A (zh) 一种证件识别方法、装置、设备及存储介质
CN112990142B (zh) 一种基于ocr的视频导图生成方法、装置、设备及存储介质
CN114386013A (zh) 学籍自动认证方法、装置、计算机设备及存储介质
KR102008630B1 (ko) 이미지 인식률 증대 장치 및 방법
Adak et al. Automatic number plate recognition (ANPR) with YOLOv3-CNN
CN115690795A (zh) 简历信息提取方法、装置、电子设备和存储介质
Thaiparnit et al. Tracking vehicles system based on license plate recognition
US11704924B2 (en) Method, apparatus, and computer-readable storage medium for recognizing characters in a digital document
CN115527197B (zh) 车号识别方法、装置、计算机设备以及卡车

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: 18901295

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS (EPO FORM 1205A DATED 18.11.2020)

122 Ep: pct application non-entry in european phase

Ref document number: 18901295

Country of ref document: EP

Kind code of ref document: A1