WO2022100401A1 - 基于图像识别的价格信息处理方法、装置、设备和介质 - Google Patents

基于图像识别的价格信息处理方法、装置、设备和介质 Download PDF

Info

Publication number
WO2022100401A1
WO2022100401A1 PCT/CN2021/125399 CN2021125399W WO2022100401A1 WO 2022100401 A1 WO2022100401 A1 WO 2022100401A1 CN 2021125399 W CN2021125399 W CN 2021125399W WO 2022100401 A1 WO2022100401 A1 WO 2022100401A1
Authority
WO
WIPO (PCT)
Prior art keywords
information
candidate text
price
text information
candidate
Prior art date
Application number
PCT/CN2021/125399
Other languages
English (en)
French (fr)
Inventor
魏莲莲
史逯强
夏继光
戚依楠
Original Assignee
北京沃东天骏信息技术有限公司
北京京东尚科信息技术有限公司
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 北京沃东天骏信息技术有限公司, 北京京东尚科信息技术有限公司 filed Critical 北京沃东天骏信息技术有限公司
Publication of WO2022100401A1 publication Critical patent/WO2022100401A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/06—Buying, selling or leasing transactions
    • G06Q30/0601—Electronic shopping [e-shopping]
    • G06Q30/0611—Request for offers or quotes
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/95—Retrieval from the web
    • G06F16/951—Indexing; Web crawling techniques
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/044—Recurrent networks, e.g. Hopfield networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/06—Buying, selling or leasing transactions
    • G06Q30/0601—Electronic shopping [e-shopping]
    • G06Q30/0641—Shopping interfaces
    • G06Q30/0643—Graphical representation of items or shoppers

Definitions

  • Embodiments of the present disclosure relate to the field of computer technology, and in particular, to a method, apparatus, device, and medium for processing price information based on image recognition.
  • Price information processing is a technology that obtains price information through certain technical means.
  • the common method is to crawl the price information displayed on the page through a specific crawler.
  • the web page structure of different shopping platforms is often different. Therefore, different crawler programs need to be developed to crawl the price information displayed on the page; and when the structure of the web page changes, the crawler program will fail, making the crawler program’s Development and maintenance costs are high; in addition, the price on the page is only the current price of the item. For scenarios such as new product pre-sale, the price information at the time of pre-sale cannot be obtained in advance, resulting in inaccurate price information generated and displayed.
  • Some embodiments of the present disclosure propose an image recognition-based price information processing method, apparatus, device and medium to solve one or more of the technical problems mentioned in the above background art section.
  • some embodiments of the present disclosure provide a method for processing price information based on image recognition, the method comprising: extracting text information from an image of an item display as text recognition information to obtain a text recognition information set; The text identification information that satisfies the first preset condition group is selected from the information set as candidate text information to obtain a candidate text information set.
  • the price-related identifier information set is generated, and the first candidate text information, the second candidate text information set and the third candidate text information are generated.
  • a first distance value set and a second distance value set are generated.
  • candidate price information is generated.
  • the above-mentioned price identification keyword information set and the preset price-related identifier information set generate the first candidate text information, the second candidate text information set and the third candidate text information, including: : The candidate text information whose text content information includes any price tag keyword in the above price tag keyword information set is selected from the candidate text information set as the first candidate text information.
  • the above-mentioned price identification keyword information set and the preset price-related identifier information set generate the first candidate text information, the second candidate text information set and the third candidate text information, including: : Selecting candidate text information containing numerical values in the text content information from the above candidate text information set as the second candidate text information to obtain a second candidate text information set.
  • the above-mentioned price identification keyword information set and the preset price-related identifier information set generate the first candidate text information, the second candidate text information set and the third candidate text information, including: : From the above-mentioned candidate text information set, the candidate text information whose text content information includes any price-related identifier information in the above-mentioned price-related identifier information set is selected as the third candidate text information.
  • generating a first distance value set and a second distance value set based on the above-mentioned first candidate text information, the above-mentioned second candidate text information set and the above-mentioned third candidate text information including the following steps: The distance value between the center coordinate value of the identification frame included in the information and the coordinate value of the identification frame center included in each second candidate text information set in the above-mentioned second candidate text information set is determined as the first distance value, and the first distance value set is obtained; The distance value between the coordinate value of the center of the identification frame included in the third candidate text information and the coordinate value of the center of the identification frame included in each of the second candidate text information in the second candidate text information set is determined as the second distance value, and the second distance value is obtained. gather.
  • generating candidate price information based on the above-mentioned second candidate text information set, the above-mentioned first distance value set and the above-mentioned second distance value set includes the following steps: filtering out the corresponding No. 1 information from the above-mentioned second candidate text information set.
  • a second candidate text information whose distance value satisfies the second preset condition is used as the first candidate price information, and a first candidate price information set is obtained; the corresponding second distance value satisfies the third
  • the first candidate price information of the preset condition is used as the candidate price information.
  • the method further includes: in response to determining that there is candidate text information including any price-related keyword information in the preset price-related keyword information set in the above-mentioned candidate text information set, based on the above-mentioned candidate text information set and the above-mentioned price.
  • a set of relevant keyword information is used to generate fourth candidate text information.
  • generating the fourth candidate text information based on the above-mentioned candidate text information set and the above-mentioned price-related keyword information set comprising: selecting from the above-mentioned candidate text information set that the text content information includes any of the above-mentioned price-related keyword information sets.
  • a candidate text information of price-related keyword information is used as the fourth candidate text information.
  • generating fourth candidate text information based on the above-mentioned candidate text information set and the above-mentioned price-related keyword information set further comprising: combining the center coordinate value of the identification frame included in the above-mentioned fourth candidate text information with the above-mentioned second candidate text information.
  • the distance value of the coordinate value of the center of the identification frame included in each second candidate text information in the set is determined as the third distance value to obtain a third distance value set.
  • generating the fourth candidate text information based on the above-mentioned candidate text information set and the above-mentioned price-related keyword information set further comprising: filtering out a corresponding third distance value from the above-mentioned second candidate text information set that satisfies the fourth prediction.
  • the second candidate text information of the condition is set as price-related text information.
  • generating the fourth candidate text information based on the above-mentioned candidate text information set and the above-mentioned price-related keyword information set further comprising: in response to determining that the difference between the above-mentioned candidate price information and the page price information is within a preset range, generating the fourth candidate text information.
  • the above-mentioned candidate price information is determined as the price information to be presented.
  • generating fourth candidate text information based on the candidate text information set and the price-related keyword information set further includes: generating presentation price information based on the to-be-presented price information and the price-related text information.
  • some embodiments of the present disclosure provide a price information processing device based on image recognition, the device includes: an extraction unit configured to extract text information from an item display image as text recognition information to obtain a text recognition information set a screening unit, configured to screen out the text identification information that satisfies the first preset condition group from the above-mentioned text identification information set as candidate text information to obtain a candidate text information set; the first generating unit is configured to respond to determining the above-mentioned text identification information
  • the candidate text information set contains candidate text information including any price tag keyword information in the preset price tag keyword information set, based on the above candidate text information set, the above price tag keyword information set and the preset price related tags A symbol information set to generate a first candidate text information set, a second candidate text information set and a third candidate text information set; a second generating unit is configured to be based on the first candidate text information set, the second candidate text information set and the above-mentioned first candidate text information set.
  • the first generating unit is further configured to: select from the above-mentioned candidate text information set candidate text information whose text content information includes any price identification keyword in the above-mentioned price identification keyword information set as the first candidate text. information.
  • the first generating unit is further configured to: filter out candidate text information including a numerical value in the text content information from the above candidate text information set as the second candidate text information to obtain the second candidate text information set.
  • the first generating unit is further configured to: filter out candidate text information whose text content information includes any price-related identifier information in the above-mentioned price-related identifier information set from the above-mentioned candidate text information set as the third candidate. text information.
  • the second generating unit is further configured to: compare the coordinate value of the center of the identification frame included in the above-mentioned first candidate text information with the coordinate value of the center of the identification frame included in each of the second candidate text information in the above-mentioned second candidate text information set.
  • the distance value is determined as the first distance value, and the first distance value set is obtained;
  • the distance value of the frame center coordinate value is determined as the second distance value, and a second distance value set is obtained.
  • the third generating unit is further configured to: filter out the second candidate text information whose first distance value satisfies the second preset condition from the above-mentioned second candidate text information set as the first candidate price information, and obtain: The first candidate price information set; the first candidate price information whose corresponding second distance value satisfies the third preset condition is selected from the above-mentioned first candidate price information set as the candidate price information.
  • the apparatus further includes: in response to determining that there is candidate text information including any price-related keyword information in the preset price-related keyword information set in the above-mentioned candidate text information set, based on the above-mentioned candidate text information set and the above-mentioned price.
  • a set of relevant keyword information is used to generate fourth candidate text information.
  • the apparatus is further configured to: filter out candidate text information whose text content information includes any price-related keyword information in the above-mentioned price-related keyword information set from the above-mentioned candidate text information set as the fourth candidate text information.
  • the device further includes: determining the distance value between the coordinate value of the center of the identification frame included in the above-mentioned fourth candidate text information and the coordinate value of the center of the identification frame included in each second candidate text information in the above-mentioned second candidate text information set as: The third distance value is obtained, and the third distance value set is obtained.
  • the apparatus further includes: filtering out the second candidate text information whose third distance value satisfies the fourth preset condition from the above-mentioned second candidate text information set as the price-related text information.
  • the apparatus further includes: in response to determining that the difference between the candidate price information and the page price information is within a preset range, determining the candidate price information as the price information to be presented.
  • the apparatus further includes: generating presentation price information based on the above-mentioned price information to be presented and the above-mentioned price-related text information.
  • some embodiments of the present disclosure provide an electronic device, comprising: one or more processors; a storage device on which one or more programs are stored, when one or more programs are stored by one or more The processor executes, causing one or more processors to implement the method described in any implementation manner of the above first aspect.
  • some embodiments of the present disclosure provide a computer-readable medium on which a computer program is stored, wherein, when the program is executed by a processor, the method described in any implementation manner of the above-mentioned first aspect is implemented.
  • the accuracy of the price information generated by the image recognition-based price information processing methods of some embodiments of the present disclosure is improved.
  • the reason why the generated price information is not accurate is that the price information crawled by the crawler program often does not match the actual price in scenarios such as new product pre-sale.
  • the image recognition-based price information processing methods of some embodiments of the present disclosure obtain the text recognition information set in the item display image. Then, the text recognition information that meets the preset conditions is filtered out according to the preset conditions. In actual situations, the distance between the display position of price information and the display position of price identification keyword information or price-related identifier information is generally close. price information.
  • this method determines the price information of the item through the text identification information set in the item display image, which can better characterize the price corresponding to the item in scenarios such as new product pre-sale. Thus, the accuracy of the generated price information is improved.
  • FIG. 1 is a schematic diagram of an application scenario of an image recognition-based price information processing method according to some embodiments of the present disclosure
  • FIG. 2 is a flowchart of some embodiments of an image recognition-based price information processing method according to the present disclosure
  • 3 is a text recognition information set in some embodiments of the image recognition-based price information processing method according to the present disclosure
  • FIG. 4 is a flow chart of other embodiments of the image recognition-based price information processing method according to the present disclosure.
  • FIG. 5 is a schematic structural diagram of some embodiments of an image recognition-based price information processing apparatus according to the present disclosure
  • FIG. 6 is a schematic structural diagram of an electronic device suitable for implementing some embodiments of the present disclosure.
  • FIG. 1 is a schematic diagram of an application scenario of an image recognition-based price information processing method according to some embodiments of the present disclosure.
  • the computing device 101 may extract text information from the item display image 102 as text identification information to obtain a text identification information set 103 . Then, the computing device 101 may screen out the text identification information that satisfies the first preset condition group from the text identification information set 103 as candidate text information to obtain the candidate text information set 104 .
  • the computing device 101 may respond to determining that the candidate text information set 104 contains candidate text information including any price identification keyword information in the preset price identification keyword information set 105, based on the candidate text information set 104, the above The price identifies the keyword information set 105 and the preset price-related identifier information set 106 to generate the first candidate text information 107 , the second candidate text information set 108 and the third candidate text information 109 . Then, the computing device 101 may generate a first distance value set 110 and a second distance value set 111 based on the above-mentioned first candidate text information 107 , the above-mentioned second candidate text information set 108 and the above-mentioned third candidate text information 109 . Finally, the computing device 101 may generate candidate price information 112 based on the second candidate text information set 108 , the first distance value set 110 , and the second distance value set 111 .
  • the above computing device 101 may be hardware or software.
  • the computing device When the computing device is hardware, it can be implemented as a distributed cluster composed of multiple servers or terminal devices, or can be implemented as a single server or a single terminal device.
  • a computing device When a computing device is embodied as software, it may be installed in the hardware devices listed above. It can be implemented, for example, as multiple software or software modules for providing distributed services, or as a single software or software module. There is no specific limitation here.
  • FIG. 1 is merely illustrative. There may be any number of computing devices depending on implementation needs.
  • the image recognition-based price information processing method includes the following steps:
  • Step 201 extracting text information from the item display image as text identification information to obtain a text identification information set.
  • the execution body of the image recognition-based price information processing method can use OCR (Optical Character Recognition, Optical Character Recognition) technology, EAST (An Efficient and Accurate Scene Text) Detector, scene text detection) algorithm, CNN (Convolutional Neural Networks, convolutional neural network) model, etc., are obtained by extracting text from the above-mentioned item display images.
  • OCR Optical Character Recognition, Optical Character Recognition
  • EAST An Efficient and Accurate Scene Text
  • scene text detection scene text detection
  • CNN Convolutional Neural Networks, convolutional neural network
  • Step 202 Screen out the text identification information that satisfies the first preset condition group from the text identification information set as candidate text information to obtain a candidate text information set.
  • the above-mentioned execution subject may select text identification information that satisfies the first preset condition group from the text identification information set 301 as candidate text information to obtain a candidate text information set.
  • the first preset condition group may include, but is not limited to, at least one of the following: the area of the identification frame corresponding to the text identification information is greater than or equal to 1 cm 2 , and the center coordinate value of the identification frame corresponding to the text identification information is a positive number.
  • the above-mentioned identification frame may be used to frame the text information in the above-mentioned displayed picture.
  • the center coordinate value of the above-mentioned identification frame is the coordinate value in the image coordinate system.
  • the image coordinate system may be a coordinate system established with the upper left corner of the image as the origin, the line parallel to the wide side of the image as the horizontal axis, and the line parallel to the long side of the image as the vertical axis.
  • Step 203 in response to determining that there is candidate text information including any price identification keyword information in the preset price identification keyword information set in the candidate text information set, based on the candidate text information set, the price identification keyword information set and the preset price identification keyword information set.
  • the price-related identifier information set is generated, and the first candidate text information, the second candidate text information set and the third candidate text information are generated.
  • the execution body in response to determining that the candidate text information set contains candidate text information including any price identification keyword information in the preset price identification keyword information set, the execution body based on the candidate text information set, the price identification key
  • the word information set and the preset price-related identifier information set are used to generate the first candidate text information, the second candidate text information set and the third candidate text information.
  • the above price identification keyword information set may include but not limited to at least one of the following: “arrival price”, “store celebration price”, “newcomer price”, “school price”, “preferential price”, “event price”, “ Coupon After Price”, “Crazy Price", “Buy Price”, “Sale Price”, “Great Price”, “Surprise Price”, “Carnival Price”, “Limited Time Price”.
  • the above price-related identifier information set may include, but is not limited to, at least one of the following: “yuan”, “ ⁇ ”, “$”. The following steps can be included:
  • the BF (Brute Force, violence) algorithm is used to screen out the candidate text information containing any price identification keyword in the above-mentioned price-representing keyword information set from the above-mentioned candidate text information set as the first candidate text information.
  • the candidate text information containing the text value is selected from the above candidate text information set as the second candidate text information to obtain the second candidate text information set.
  • the candidate text information containing any price-related identifier information in the above-mentioned price-related identifier information set is selected from the above-mentioned candidate text information set by the Sunday algorithm as the third candidate text information.
  • Step 204 based on the first candidate text information, the second candidate text information set and the third candidate text information, generate a first distance value set and a second distance value set.
  • the above-mentioned execution body generates the first distance value set and the second distance value set based on the first candidate text information, the second candidate text information set and the third candidate text information, which may include the following steps:
  • the first step is to determine the coordinate value of the upper left corner of the identification frame corresponding to the above-mentioned first candidate text information and the upper left corner of the identification frame corresponding to each second candidate text information in the above-mentioned second candidate text information set by the distance formula between the two points.
  • a distance value between the coordinate values of to generate a first distance value, and obtain a first distance value set.
  • the distance value can be Euclidean distance and so on.
  • the coordinate value of the upper left corner of the identification box corresponding to the third candidate text information and the upper left corner of the identification box corresponding to each second candidate text information in the second candidate text information set are determined by the distance formula between the two points.
  • a distance value between the coordinate values of to generate a second distance value, and obtain a second distance value set.
  • Step 205 Generate candidate price information based on the second candidate text information set, the first distance value set, and the second distance value set.
  • the above-mentioned execution body generates candidate price information based on the second candidate text information set, the first distance value set, and the second distance value set.
  • the second candidate text information whose corresponding first distance value and second distance value are respectively the largest value in the above-mentioned first distance value set and the above-mentioned second distance value set may be selected from the above-mentioned second candidate text information set as a candidate. price information.
  • the accuracy of the price information generated by the image recognition-based price information processing methods of some embodiments of the present disclosure is improved.
  • the reason why the generated price information is not accurate is that the price information crawled by the crawler program often does not match the actual price in scenarios such as new product pre-sale.
  • the image recognition-based price information processing methods of some embodiments of the present disclosure obtain the text recognition information set in the item display image. Then, the text recognition information that meets the preset conditions is filtered out according to the preset conditions. In actual situations, the distance between the display position of price information and the display position of price identification keyword information or price-related identifier information is generally close. price information.
  • this method determines the price information of the item through the text identification information set in the item display image, which can better characterize the price corresponding to the item in scenarios such as new product pre-sale. Thus, the accuracy of the generated price information is improved.
  • FIG. 4 shows a flow 400 of other embodiments of the image recognition-based price information processing method.
  • the process 400 of the image recognition-based price information processing method includes the following steps:
  • Step 401 Extract text information from the item display image as text identification information to obtain a text identification information set.
  • the above-mentioned executive body can identify and extract text information from the above-mentioned item display image through an RNN (Recurrent Neural Network) model and an LSTM (Long Short-Term Memory, long short-term memory network) model to A collection of text recognition information is generated.
  • the above-mentioned text information may include, but is not limited to, at least one of the following: text content information, text confidence value, and identification box size information.
  • the above-mentioned identification frame size information may include, but is not limited to, at least one of the following: a central coordinate value of the identification frame, a length value of the identification frame, and a width value of the identification frame.
  • the above text identification information set may be [[six-speed wind, 0.29652, (80, 80), 40, 45], [five-leaf high-efficiency and energy-saving wind, 0.99917, (240, 80), 100, 45]] .
  • Step 402 Screen out the text identification information that satisfies the first preset condition group from the text identification information set as candidate text information to obtain a candidate text information set.
  • the above-mentioned execution body may screen out the text identification information that satisfies the first preset condition group from the text identification information set as candidate text information to obtain the candidate text information set.
  • the above-mentioned first preset condition group may include, but is not limited to, at least one of the following: the text confidence value included in the text identification information is greater than or equal to the preset credibility value, and the text identification information included in the identification frame length value is greater than or equal to the preset identification The frame length value and the marker frame width value are greater than or equal to the preset marker frame width value.
  • the above-mentioned preset reliability value may be 0.1.
  • the above-mentioned preset identification frame length value may be 15 (pixels).
  • the above-mentioned preset width value of the identification frame may be 15 (pixels).
  • Step 403 from the candidate text information set, select the candidate text information whose text content information includes any price identification keyword in the price identification keyword information set as the first candidate text information.
  • the above-mentioned execution body may filter out candidate text information including any price identification keyword in the price identification keyword information set in the text content information from the candidate text information set as the first candidate text information.
  • the above-mentioned price identification keyword information may include, but is not limited to, at least one of the following items: “new product price”, “matching price”, “experience price”, “authentic price”, “new price”, “cost-effective price”, “ Brand Price”, “Exclusive Price”.
  • the above-mentioned executive body may use the KMP (Knuth-Morris-Pratt, pattern matching) algorithm to filter out candidate texts whose text content information includes any price identification keyword in the price identification keyword information set from the above candidate text information set information as the first candidate text information.
  • KMP Knuth-Morris-Pratt, pattern matching
  • Step 404 screening out candidate text information including a numerical value in the text content information from the candidate text information set as the second candidate text information to obtain a second candidate text information set.
  • the above-mentioned execution body may filter out candidate text information including any price identification keyword in the price identification keyword information set in the text content information from the candidate text information set in various ways, as the first candidate text information .
  • the above-mentioned execution body may filter candidate text information including any price identification keyword in the price identification keyword information set in the text content information from the candidate text information set through the Rabin-Karp algorithm as the first candidate text information.
  • Step 405 Screen out candidate text information that includes any price-related identifier information in the price-related identifier information set in the text content information from the candidate text information set as the third candidate text information.
  • the above-mentioned execution body may filter out any price in the set of price-related identifier information contained in the text content information from the set of candidate text information through any of the algorithms mentioned in steps 202, 403, and 404.
  • the candidate text information of the relevant identifier information serves as the third candidate text information.
  • the above price-related identifier information set may include at least one of the following items: "yuan”, "$", " ⁇ ", and "£".
  • Step 406 Determine the distance value between the center coordinate value of the identification frame included in the first candidate text information and the center coordinate value of the identification frame included in each second candidate text information in the second candidate text information set as the first distance value, and obtain the first distance value. A collection of distance values.
  • the execution body may determine the center coordinate value of the identification frame included in the first candidate text information and the identification frame included in each second candidate text information in the second candidate text information set by using a distance formula between two points The distance value of the center coordinate value is used to generate the first distance value, and the first distance value set is obtained.
  • Step 407 Determine the distance value between the center coordinate value of the identification frame included in the third candidate text information and the coordinate value of the identification frame center included in each second candidate text information in the second candidate text information set as the second distance value, and obtain the first A collection of two distance values.
  • the execution body may determine the center coordinate value of the identification frame included in the third candidate text information and the identification frame included in each second candidate text information in the second candidate text information set by using a distance formula between two points The distance value of the center coordinate value is used to generate the second distance value, and the second distance value set is obtained.
  • Step 408 Screen out the second candidate text information whose corresponding first distance value satisfies the second preset condition from the second candidate text information set as the first candidate price information to obtain the first candidate price information set.
  • the above-mentioned execution body may filter out the second candidate text information whose corresponding first distance value satisfies the second preset condition from the second candidate text information set as the first candidate price information, and obtain the first candidate price collection of information.
  • the above-mentioned second preset condition may also be that the first distance value is less than or equal to the average value of each first distance value in the above-mentioned first distance value set.
  • Step 409 Screen out the first candidate price information whose corresponding second distance value satisfies the third preset condition from the first candidate price information set as the candidate price information.
  • the above-mentioned execution body may filter out the first candidate price information whose corresponding second distance value satisfies the third preset condition from the first candidate price information set as the candidate price information.
  • the above-mentioned third preset condition may be that the second distance value is equal to the smallest second distance value in the above-mentioned second distance value set.
  • the above-mentioned execution body may, based on the candidate text information set and the price-related keyword information set, and generate the fourth candidate text information.
  • the above price-related keyword information set may include at least one of the following items: "coupon”, “discount”, “immediate discount”, “full discount”.
  • the above-mentioned execution body may determine the distance value between the center coordinate value of the identification box included in the fourth candidate text information and the center coordinate value of the identification box included in each second candidate text information in the second candidate text information set as the third value.
  • the distance value is obtained, and the third distance value set is obtained.
  • the distance value between the center coordinate value of the identification frame included in the above-mentioned fourth candidate text information and the center coordinate of the identification frame included in each second candidate text information in the above-mentioned second candidate text information set can be determined by the distance formula between the two points .
  • the above-mentioned execution body may filter out the second candidate text information whose corresponding third distance value satisfies the fourth preset condition from the second candidate text information set as the price-related text information.
  • the above-mentioned fourth preset condition may be that the third distance value is the same as the minimum value among the above-mentioned third distance values.
  • the execution entity may determine the candidate price information as the price information to be presented in response to determining that the difference between the candidate price information and the page price information is within a preset range.
  • the above-mentioned preset range may be [0, 100].
  • the above page price information may be obtained by crawling from the display page of the corresponding item through a crawler program.
  • the execution entity may generate presentation price information based on the price information to be presented and the price-related text information.
  • the above-mentioned price information to be presented may be "399 yuan”, and the above-mentioned price-related text information may be "full 300 yuan-40 yuan”. Therefore, the generated presentation price information may be "359 yuan”.
  • the above-mentioned price information to be presented may be "99 yuan”, and the above-mentioned price-related text information may be "full 200 yuan-40 yuan”. Therefore, the generated presentation price information may be "99 yuan”.
  • text identification information with a low text confidence value can be understood as insufficiently accurate text information identification.
  • a smaller value of the length of the identification box and the value of the width of the identification box can be understood as the area occupied by the text information inside the identification box is small.
  • text information is generally punctuation marks, etc.
  • the generation impact is small, so the obtained text identification information is initially screened through the first preset condition group, and the text identification information whose text confidence value, mark frame length value and mark frame width value do not meet the conditions is filtered out. The consumption of computing resources is reduced.
  • the set of text identification information identified from the item display image contains price identification keywords
  • the text identification information of either numerical value or price-related identifier information has a greater impact on the final price generation. Therefore, the first candidate text information, the second candidate text information set and the third candidate text information are obtained through conditional screening. In this way, information that has little influence on the final price generation is further filtered out, and the consumption of computing resources is reduced.
  • the numerical value near the text identification information including price identification keyword information and the text identification information including price-related identifier information is often price information. Based on this, further screening is performed by distance to generate candidate price information. This method is compared to using a crawler program to crawl the price information in the page as price information. It can better adapt to the identification and generation of price information in scenarios such as new product pre-sale. Thus, the accuracy of the generated price information is improved.
  • the present disclosure provides some embodiments of a price information processing apparatus based on image recognition, and these apparatus embodiments are similar to those of the method embodiments shown in FIG. 2 .
  • the apparatus can be specifically applied to various electronic devices.
  • the image recognition-based price information processing apparatus 500 in some embodiments includes: an extraction unit 501 , a screening unit 502 , a first generation unit 503 , a second generation unit 504 and a third generation unit 505 .
  • the extracting unit 501 is configured to extract text information from the item display image as text identification information to obtain a text identification information set.
  • the screening unit 502 is configured to screen out the text identification information that satisfies the first preset condition group from the text identification information set as candidate text information to obtain the candidate text information set.
  • the first generating unit 503 is configured to, in response to determining that there is candidate text information including any price tag keyword information in the preset price tag keyword information set in the above candidate text information set, based on the above candidate text information set, the above The price identification keyword information set and the preset price-related identifier information set are used to generate the first candidate text information, the second candidate text information set and the third candidate text information.
  • the second generating unit 504 is configured to generate a first distance value set and a second distance value set based on the first candidate text information, the second candidate text information set and the third candidate text information.
  • the third generating unit 505 is configured to generate candidate price information based on the second candidate text information set, the first distance value set and the second distance value set.
  • the first generating unit 503 is further configured to: filter out the text content information from the candidate text information set that contains any price identification keyword in the price identification keyword information set The candidate text information of is used as the first candidate text information.
  • the first generating unit 503 is further configured to: filter out candidate text information containing numerical values in the text content information from the above-mentioned candidate text information set as the second candidate text information, and obtain the first candidate text information. Two candidate text information sets.
  • the first generating unit 503 is further configured to: filter out the text content information from the candidate text information set that includes any price-related identifier in the price-related identifier information set The candidate text information of the information is used as the third candidate text information.
  • the second generating unit 504 is further configured to: compare the coordinate value of the center of the identification frame included in the above-mentioned first candidate text information with each second candidate in the above-mentioned second candidate text information set The distance value of the center coordinate value of the identification frame included in the text information is determined as the first distance value, and a first distance value set is obtained; The distance value of the center coordinate value of the identification frame included in the second candidate text information is determined as the second distance value, and a second distance value set is obtained.
  • the third generating unit 505 is further configured to: filter out the second candidate text whose corresponding first distance value satisfies the second preset condition from the above-mentioned second candidate text information set The information is used as the first candidate price information to obtain a first candidate price information set; the first candidate price information whose corresponding second distance value satisfies the third preset condition is selected from the first candidate price information set as the candidate price information.
  • the apparatus 500 further includes: a fourth generating unit, configured to respond to determining that any price-related key in the preset price-related keyword information set exists in the above-mentioned candidate text information set
  • the candidate text information of the word information is generated based on the above-mentioned candidate text information set and the above-mentioned price-related keyword information set, and the fourth candidate text information is generated.
  • the fourth generating unit is further configured to: filter out the text content information from the candidate text information set that contains any price-related keyword information in the price-related keyword information set.
  • the candidate text information is used as the fourth candidate text information.
  • the apparatus 500 further includes: combining the coordinate value of the center of the identification frame included in the fourth candidate text information with the identification included in each second candidate text information in the second candidate text information set The distance value of the frame center coordinate value is determined as the third distance value, and a third distance value set is obtained.
  • the apparatus 500 further includes: in response to determining that the difference between the candidate price information and the page price information is within a preset range, determining the candidate price information as the price information to be presented.
  • the apparatus 500 further includes: generating presentation price information based on the above-mentioned price information to be presented and the above-mentioned price-related text information.
  • the units recorded in the apparatus 500 correspond to the respective steps in the method described with reference to FIG. 2 .
  • the display picture corresponding to the item may include full reduction information.
  • the full minus keyword is identified by the fourth generation unit.
  • the full subtraction value is determined.
  • final price information is obtained, and the accuracy of price information generation is improved.
  • FIG. 6 a schematic structural diagram of an electronic device (such as the computing device 101 shown in FIG. 1 ) 600 suitable for implementing some embodiments of the present disclosure is shown.
  • the electronic device shown in FIG. 6 is only an example, and should not impose any limitation on the function and scope of use of the embodiments of the present disclosure.
  • an electronic device 600 may include a processing device (eg, a central processing unit, a graphics processor, etc.) 601 that may be loaded into random access according to a program stored in a read only memory (ROM) 602 or from a storage device 608 Various appropriate actions and processes are executed by the programs in the memory (RAM) 603 . In the RAM 603, various programs and data required for the operation of the electronic device 600 are also stored.
  • the processing device 601, the ROM 602, and the RAM 603 are connected to each other through a bus 604.
  • An input/output (I/O) interface 605 is also connected to bus 604 .
  • I/O interface 605 input devices 606 including, for example, a touch screen, touchpad, keyboard, mouse, camera, microphone, accelerometer, gyroscope, etc.; including, for example, a liquid crystal display (LCD), speakers, vibration An output device 607 of a computer, etc.; a storage device 608 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 609.
  • Communication means 609 may allow electronic device 600 to communicate wirelessly or by wire with other devices to exchange data. While FIG. 6 shows electronic device 600 having various means, it should be understood that not all of the illustrated means are required to be implemented or provided. More or fewer devices may alternatively be implemented or provided. Each block shown in FIG. 6 may represent one device, or may represent multiple devices as required.
  • the processes described above with reference to the flowcharts may be implemented as computer software programs.
  • some embodiments of the present disclosure include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the method illustrated in the flowchart.
  • the computer program may be downloaded and installed from the network via the communication device 609, or from the storage device 608, or from the ROM 602.
  • the processing device 601 When the computer program is executed by the processing device 601, the above-mentioned functions defined in the methods of some embodiments of the present disclosure are performed.
  • the computer-readable medium described in some embodiments of the present disclosure may be a computer-readable signal medium or a computer-readable storage medium, or any combination of the above two.
  • the computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus or device, or a combination of any of the above. More specific examples of computer readable storage media may include, but are not limited to, electrical connections with one or more wires, portable computer disks, hard disks, random access memory (RAM), read only memory (ROM), erasable Programmable read only memory (EPROM or flash memory), fiber optics, portable compact disk read only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.
  • a computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.
  • a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code therein. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the foregoing.
  • a computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device .
  • Program code embodied on a computer readable medium may be transmitted using any suitable medium including, but not limited to, electrical wire, optical fiber cable, RF (radio frequency), etc., or any suitable combination of the foregoing.
  • the client and server can use any currently known or future developed network protocol such as HTTP (HyperText Transfer Protocol) to communicate, and can communicate with digital data in any form or medium Communication (eg, a communication network) interconnects.
  • HTTP HyperText Transfer Protocol
  • Examples of communication networks include local area networks (“LAN”), wide area networks (“WAN”), the Internet (eg, the Internet), and peer-to-peer networks (eg, ad hoc peer-to-peer networks), as well as any currently known or future development network of.
  • the above-mentioned computer-readable medium may be included in the above-mentioned electronic device; or may exist alone without being assembled into the electronic device.
  • the above-mentioned computer-readable medium carries one or more programs, and when the above-mentioned one or more programs are executed by the electronic device, the electronic device: extracts text information from the item display image as text identification information, and obtains a set of text identification information .
  • the text identification information that satisfies the first preset condition group is selected from the above text identification information set as candidate text information to obtain a candidate text information set.
  • Computer program code for carrying out operations of some embodiments of the present disclosure may be written in one or more programming languages, including object-oriented programming languages—such as Java, Smalltalk, C++, or a combination thereof, Also included are conventional procedural programming languages - such as the "C" language or similar programming languages.
  • the program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer, or entirely on the remote computer or server.
  • the remote computer may be connected to the user's computer through any kind of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (eg, using an Internet service provider to via Internet connection).
  • LAN local area network
  • WAN wide area network
  • each block in the flowchart or block diagrams may represent a module, segment, or portion of code that contains one or more logical functions for implementing the specified functions executable instructions.
  • the functions noted in the blocks may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved.
  • each block of the block diagrams and/or flowchart illustrations, and combinations of blocks in the block diagrams and/or flowchart illustrations can be implemented in dedicated hardware-based systems that perform the specified functions or operations , or can be implemented in a combination of dedicated hardware and computer instructions.
  • the units described in some embodiments of the present disclosure may be implemented by means of software, and may also be implemented by means of hardware.
  • the described unit may also be provided in the processor, for example, it may be described as: a processor includes an extraction unit, a screening unit, a first generation unit, a second generation unit and a third generation unit.
  • the names of these units do not constitute a limitation on the unit itself under certain circumstances, for example, the screening unit may also be described as "screening out the text identification information that satisfies the first preset condition group from the text identification information set. As the candidate text information, obtain the unit of the candidate text information set".
  • exemplary types of hardware logic components include: Field Programmable Gate Arrays (FPGAs), Application Specific Integrated Circuits (ASICs), Application Specific Standard Products (ASSPs), Systems on Chips (SOCs), Complex Programmable Logical Devices (CPLDs) and more.
  • FPGAs Field Programmable Gate Arrays
  • ASICs Application Specific Integrated Circuits
  • ASSPs Application Specific Standard Products
  • SOCs Systems on Chips
  • CPLDs Complex Programmable Logical Devices

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Finance (AREA)
  • Accounting & Taxation (AREA)
  • General Engineering & Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Software Systems (AREA)
  • Databases & Information Systems (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Evolutionary Computation (AREA)
  • Biomedical Technology (AREA)
  • Mathematical Physics (AREA)
  • Biophysics (AREA)
  • Artificial Intelligence (AREA)
  • General Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Computational Linguistics (AREA)
  • Health & Medical Sciences (AREA)
  • Development Economics (AREA)
  • Economics (AREA)
  • Marketing (AREA)
  • Strategic Management (AREA)
  • General Business, Economics & Management (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

本公开的实施例公开了基于图像识别的价格信息处理方法、装置、设备和介质。该方法的一具体实施方式包括：从物品展示图像中提取文本信息作为文本识别信息，得到文本识别信息集合；从文本识别信息集合中筛选出满足第一预设条件组的文本识别信息作为候选文本信息；响应于确定候选文本信息集合中存在包括预设的价格标识关键字信息集合中任一价格标识关键字信息的候选文本信息，生成第一候选文本信息、第二候选文本信息集合和第三候选文本信息；生成第一距离值集合和第二距离值集合；基于第二候选文本信息集合、第一距离值集合和第二距离值集合，生成候选价格信息。该实施方式提高了生成的价格信息的准确度。

Description

基于图像识别的价格信息处理方法、装置、设备和介质 技术领域
本公开的实施例涉及计算机技术领域，具体涉及基于图像识别的价格信息处理方法、装置、设备和介质。
背景技术
价格信息处理，是通过一定的技术手段获取价格信息的一项技术。目前，常用的方法是通过特定的爬虫程序爬取页面中显示的价格信息。
然而，当采用上述方式进行价格信息处理时，经常会存在如下技术问题：
第一，不同的购物平台的网页页面结构往往不相同，因此，需要开发不同的爬虫程序爬取页面中显示的价格信息；并且，当网页结构发生变化时会导致爬虫程序失效，使得爬虫程序的开发以及维护成本较高；此外，页面中的价格仅为物品的当前价格，对于新品预售等场景，无法提前获取预售时的价格信息，从而，导致生成以及显示的价格信息不够准确。
发明内容
本公开的内容部分用于以简要的形式介绍构思，这些构思将在后面的具体实施方式部分被详细描述。本公开的内容部分并不旨在标识要求保护的技术方案的关键特征或必要特征，也不旨在用于限制所要求的保护的技术方案的范围。
本公开的一些实施例提出了基于图像识别的价格信息处理方法、装置、设备和介质，来解决以上背景技术部分提到的技术问题中的一项或多项。
第一方面，本公开的一些实施例提供了一种基于图像识别的价格信息处理方法，该方法包括：从物品展示图像中提取文本信息作为文 本识别信息，得到文本识别信息集合；从上述文本识别信息集合中筛选出满足第一预设条件组的文本识别信息作为候选文本信息，得到候选文本信息集合。响应于确定上述候选文本信息集合中存在包括预设的价格标识关键字信息集合中任一价格标识关键字信息的候选文本信息，基于上述候选文本信息集合、上述价格标识关键字信息集合和预设的价格相关标识符信息集合，生成第一候选文本信息、第二候选文本信息集合和第三候选文本信息。基于上述第一候选文本信息、上述第二候选文本信息集合和上述第三候选文本信息，生成第一距离值集合和第二距离值集合。基于上述第二候选文本信息集合、上述第一距离值集合和上述第二距离值集合，生成候选价格信息。
可选的，基于上述候选文本信息集合、上述价格标识关键字信息集合和预设的价格相关标识符信息集合，生成第一候选文本信息、第二候选文本信息集合和第三候选文本信息，包括：从上述候选文本信息集合中筛选出文本内容信息中包含上述价格标识关键字信息集合中任一价格标识关键字的候选文本信息作为第一候选文本信息。
可选的，基于上述候选文本信息集合、上述价格标识关键字信息集合和预设的价格相关标识符信息集合，生成第一候选文本信息、第二候选文本信息集合和第三候选文本信息，包括：从上述候选文本信息集合中筛选出文本内容信息中包含数值的候选文本信息作为第二候选文本信息，得到第二候选文本信息集合。
可选的，基于上述候选文本信息集合、上述价格标识关键字信息集合和预设的价格相关标识符信息集合，生成第一候选文本信息、第二候选文本信息集合和第三候选文本信息，包括：从上述候选文本信息集合中筛选出文本内容信息中包含上述价格相关标识符信息集合中任一价格相关标识符信息的候选文本信息作为第三候选文本信息。
可选的，基于上述第一候选文本信息、上述第二候选文本信息集合和上述第三候选文本信息，生成第一距离值集合和第二距离值集合，包括以下步骤：将上述第一候选文本信息包括的标识框中心坐标值与上述第二候选文本信息集合中每个第二候选文本信息包括的标识框中心坐标值的距离值确定为第一距离值，得到第一距离值集合；将上述 第三候选文本信息包括的标识框中心坐标值与上述第二候选文本信息集合中每个第二候选文本信息包括的标识框中心坐标值的距离值确定为第二距离值，得到第二距离值集合。
可选的，基于上述第二候选文本信息集合、上述第一距离值集合和上述第二距离值集合，生成候选价格信息，包括以下步骤：从上述第二候选文本信息集合中筛选出对应的第一距离值满足第二预设条件的第二候选文本信息作为第一候选价格信息，得到第一候选价格信息集合；从上述第一候选价格信息集合中筛选出对应的第二距离值满足第三预设条件的第一候选价格信息作为候选价格信息。
可选的，方法还包括：响应于确定上述候选文本信息集合中存在包括预设的价格相关关键字信息集合中任一价格相关关键字信息的候选文本信息，基于上述候选文本信息集合和上述价格相关关键字信息集合，生成第四候选文本信息。
可选的，基于上述候选文本信息集合和上述价格相关关键字信息集合，生成第四候选文本信息，包括：从上述候选文本信息集合中筛选出文本内容信息包含上述价格相关关键字信息集合中任一价格相关关键字信息的候选文本信息作为第四候选文本信息。
可选的，基于上述候选文本信息集合和上述价格相关关键字信息集合，生成第四候选文本信息，还包括：将上述第四候选文本信息包括的标识框中心坐标值与上述第二候选文本信息集合中每个第二候选文本信息包括的标识框中心坐标值的距离值确定为第三距离值，得到第三距离值集合。
可选的，基于上述候选文本信息集合和上述价格相关关键字信息集合，生成第四候选文本信息，还包括：从上述第二候选文本信息集合中筛选出对应的第三距离值满足第四预设条件的第二候选文本信息作为价格相关文本信息。
可选的，基于上述候选文本信息集合和上述价格相关关键字信息集合，生成第四候选文本信息，还包括：响应于确定上述候选价格信息与页面价格信息的差值在预设范围内，将上述候选价格信息确定为待呈现价格信息。
可选的，基于上述候选文本信息集合和上述价格相关关键字信息集合，生成第四候选文本信息，还包括：基于上述待呈现价格信息和上述价格相关文本信息，生成呈现价格信息。
第二方面，本公开的一些实施例提供了一种基于图像识别的价格信息处理装置，装置包括：提取单元，被配置成从物品展示图像中提取文本信息作为文本识别信息，得到文本识别信息集合；筛选单元，被配置成从上述文本识别信息集合中筛选出满足第一预设条件组的文本识别信息作为候选文本信息，得到候选文本信息集合；第一生成单元，被配置成响应于确定上述候选文本信息集合中存在包括预设的价格标识关键字信息集合中任一价格标识关键字信息的候选文本信息，基于上述候选文本信息集合、上述价格标识关键字信息集合和预设的价格相关标识符信息集合，生成第一候选文本信息、第二候选文本信息集合和第三候选文本信息；第二生成单元，被配置成基于上述第一候选文本信息、上述第二候选文本信息集合和上述第三候选文本信息，生成第一距离值集合和第二距离值集合；第三生成单元，被配置成基于上述第二候选文本信息集合、上述第一距离值集合和上述第二距离值集合，生成候选价格信息。
可选的，第一生成单元被进一步配置成：从上述候选文本信息集合中筛选出文本内容信息中包含上述价格标识关键字信息集合中任一价格标识关键字的候选文本信息作为第一候选文本信息。
可选的，第一生成单元被进一步配置成：从上述候选文本信息集合中筛选出文本内容信息中包含数值的候选文本信息作为第二候选文本信息，得到第二候选文本信息集合。
可选的，第一生成单元被进一步配置成：从上述候选文本信息集合中筛选出文本内容信息中包含上述价格相关标识符信息集合中任一价格相关标识符信息的候选文本信息作为第三候选文本信息。
可选的，第二生成单元被进一步配置成：将上述第一候选文本信息包括的标识框中心坐标值与上述第二候选文本信息集合中每个第二候选文本信息包括的标识框中心坐标值的距离值确定为第一距离值，得到第一距离值集合；将上述第三候选文本信息包括的标识框中心坐 标值与上述第二候选文本信息集合中每个第二候选文本信息包括的标识框中心坐标值的距离值确定为第二距离值，得到第二距离值集合。
可选的，第三生成单元被进一步配置成：从上述第二候选文本信息集合中筛选出对应的第一距离值满足第二预设条件的第二候选文本信息作为第一候选价格信息，得到第一候选价格信息集合；从上述第一候选价格信息集合中筛选出对应的第二距离值满足第三预设条件的第一候选价格信息作为候选价格信息。
可选的，装置还包括：响应于确定上述候选文本信息集合中存在包括预设的价格相关关键字信息集合中任一价格相关关键字信息的候选文本信息，基于上述候选文本信息集合和上述价格相关关键字信息集合，生成第四候选文本信息。
可选的，装置被进一步配置成：从上述候选文本信息集合中筛选出文本内容信息包含上述价格相关关键字信息集合中任一价格相关关键字信息的候选文本信息作为第四候选文本信息。
可选的，装置还包括：将上述第四候选文本信息包括的标识框中心坐标值与上述第二候选文本信息集合中每个第二候选文本信息包括的标识框中心坐标值的距离值确定为第三距离值，得到第三距离值集合。
可选的，装置还包括：从上述第二候选文本信息集合中筛选出对应的第三距离值满足第四预设条件的第二候选文本信息作为价格相关文本信息。
可选的，装置还包括：响应于确定上述候选价格信息与页面价格信息的差值在预设范围内，将上述候选价格信息确定为待呈现价格信息。
可选的，装置还包括：基于上述待呈现价格信息和上述价格相关文本信息，生成呈现价格信息。
第三方面，本公开的一些实施例提供了一种电子设备，包括：一个或多个处理器；存储装置，其上存储有一个或多个程序，当一个或多个程序被一个或多个处理器执行，使得一个或多个处理器实现上述第一方面任一实现方式所描述的方法。
第四方面，本公开的一些实施例提供了一种计算机可读介质，其上存储有计算机程序，其中，程序被处理器执行时实现上述第一方面任一实现方式所描述的方法。
本公开的上述各个实施例具有如下有益效果：通过本公开的一些实施例的基于图像识别的价格信息处理方法生成的价格信息的准确度有所提高。具体来说，造成生成的价格信息不够精确的原因在于：采用爬虫程序爬取的价格信息往往与新品预售等场景下的实际价格不符。基于此，本公开的一些实施例的基于图像识别的价格信息处理方法通过获取物品展示图像中的文本识别信息集合。然后，根据预设条件筛选出符合预设条件的文本识别信息。在实际情况中，价格信息的显示位置一般与价格标识关键字信息或价格相关标识符信息的显示位置之间的距离较近，通过距离进行筛选以得到物品在对应的新品预售等场景下的价格信息。由于在新品预售等场景下，物品对应的价格信息往往会预先更新在物品展示图像中，而未同步地在显示页面中进行更新。因此，此种方法通过物品展示图像中的文本识别信息集合确定物品的价格信息，能够较好的表征新品预售等场景下的物品对应的价格。从而，提高了生成的价格信息的准确度。
附图说明
结合附图并参考以下具体实施方式，本公开各实施例的上述和其他特征、优点及方面将变得更加明显。贯穿附图中，相同或相似的附图标记表示相同或相似的元素。应当理解附图是示意性的，元件和元素不一定按照比例绘制。
图1是根据本公开的一些实施例的基于图像识别的价格信息处理方法的一个应用场景的示意图；
图2是根据本公开的基于图像识别的价格信息处理方法的一些实施例的流程图；
图3是根据本公开的基于图像识别的价格信息处理方法的一些实施例中的文本识别信息集合；
图4是根据本公开的基于图像识别的价格信息处理方法的另一些 实施例的流程图；
图5是根据本公开的基于图像识别的价格信息处理装置的一些实施例的结构示意图；
图6是适于用来实现本公开的一些实施例的电子设备的结构示意图。
具体实施方式
下面将参照附图更详细地描述本公开的实施例。虽然附图中显示了本公开的某些实施例，然而应当理解的是，本公开可以通过各种形式来实现，而且不应该被解释为限于这里阐述的实施例。相反，提供这些实施例是为了更加透彻和完整地理解本公开。应当理解的是，本公开的附图及实施例仅用于示例性作用，并非用于限制本公开的保护范围。
另外还需要说明的是，为了便于描述，附图中仅示出了与有关发明相关的部分。在不冲突的情况下，本公开中的实施例及实施例中的特征可以相互组合。
需要注意，本公开中提及的“第一”、“第二”等概念仅用于对不同的装置、模块或单元进行区分，并非用于限定这些装置、模块或单元所执行的功能的顺序或者相互依存关系。
需要注意，本公开中提及的“一个”、“多个”的修饰是示意性而非限制性的，本领域技术人员应当理解，除非在上下文另有明确指出，否则应该理解为“一个或多个”。
本公开实施方式中的多个装置之间所交互的消息或者信息的名称仅用于说明性的目的，而并不是用于对这些消息或信息的范围进行限制。
下面将参考附图并结合实施例来详细说明本公开。
图1是本公开的一些实施例的基于图像识别的价格信息处理方法的一个应用场景的示意图。
在图1的应用场景中，首先，计算设备101可以从物品展示图像102中提取文本信息作为文本识别信息，得到文本识别信息集合103。 然后，计算设备101可以从文本识别信息集合103中筛选出满足第一预设条件组的文本识别信息作为候选文本信息，得到候选文本信息集合104。其次，计算设备101可以响应于确定上述候选文本信息集合104中存在包括预设的价格标识关键字信息集合105中任一价格标识关键字信息的候选文本信息，基于上述候选文本信息集合104、上述价格标识关键字信息集合105和预设的价格相关标识符信息集合106，生成第一候选文本信息107、第二候选文本信息集合108和第三候选文本信息109。然后，计算设备101可以基于上述第一候选文本信息107、上述第二候选文本信息集合108和上述第三候选文本信息109，生成第一距离值集合110和第二距离值集合111。最后，计算设备101可以基于上述第二候选文本信息集合108、上述第一距离值集合110和上述第二距离值集合111，生成候选价格信息112。
需要说明的是，上述计算设备101可以是硬件，也可以是软件。当计算设备为硬件时，可以实现成多个服务器或终端设备组成的分布式集群，也可以实现成单个服务器或单个终端设备。当计算设备体现为软件时，可以安装在上述所列举的硬件设备中。其可以实现成例如用来提供分布式服务的多个软件或软件模块，也可以实现成单个软件或软件模块。在此不做具体限定。
应该理解，图1中的计算设备的数目仅仅是示意性的。根据实现需要，可以具有任意数目的计算设备。
继续参考图2，示出了根据本公开的基于图像识别的价格信息处理方法的一些实施例的流程图200。该基于图像识别的价格信息处理方法，包括以下步骤：
步骤201，从物品展示图像中提取文本信息作为文本识别信息，得到文本识别信息集合。
在一些实施例中，基于图像识别的价格信息处理方法的执行主体(如图1所示的计算设备101)可以通过OCR(Optical Character Recognition，光学字符识别)技术，EAST(An Efficient and Accurate Scene Text Detector，场景文字检测)算法，CNN(Convolutional Neural Networks，卷积神经网络)模型等对上述物品展示图像进行文字提取 得到的。
步骤202，从文本识别信息集合中筛选出满足第一预设条件组的文本识别信息作为候选文本信息，得到候选文本信息集合。
在一些实施例中，上述执行主体可以从文本识别信息集合301中筛选出满足第一预设条件组的文本识别信息作为候选文本信息，得到候选文本信息集合。其中，上述第一预设条件组可以包括但不限于以下至少一项：文本识别信息对应的标识框的面积大于等于1cm 2，文本识别信息对应的标识框的中心坐标值为正数。上述标识框可以用于框定上述展示图片中的文本信息。上述标识框的中心坐标值是在图像坐标系下的坐标值。上述图像坐标系可以是以图像左上角为原点，以与上述图像的宽边平行的线为横轴，以与上述图像的长边平行的线为纵轴建立的坐标系。
步骤203，响应于确定候选文本信息集合中存在包括预设的价格标识关键字信息集合中任一价格标识关键字信息的候选文本信息，基于候选文本信息集合、价格标识关键字信息集合和预设的价格相关标识符信息集合，生成第一候选文本信息、第二候选文本信息集合和第三候选文本信息。
在一些实施例中，上述执行主体响应于确定候选文本信息集合中存在包括预设的价格标识关键字信息集合中任一价格标识关键字信息的候选文本信息，基于候选文本信息集合、价格标识关键字信息集合和预设的价格相关标识符信息集合，生成第一候选文本信息、第二候选文本信息集合和第三候选文本信息。其中，上述价格标识关键字信息集合可以包括但不限于以下至少一项：“到手价”，“店庆价”，“新人价”，“开学价”，“优惠价”，“活动价”，“券后价”，“疯抢价”，“抢购价”，“促销价”，“钜惠价”，“惊喜价”，“狂欢价”，“限时价”。上述价格相关标识符信息集合可以包括但不限于以下至少一项：“元”，“￥”，“$”。可以包括以下步骤：
第一步，通过BF(Brute Force，暴力)算法从上述候选文本信息集合中筛选出包含上述价格表示关键字信息集合中任一价格标识关键字的候选文本信息作为第一候选文本信息。
第二步，通过Boyer-Moore字符串搜索算法从上述候选文本信息集合中筛选出包含文本数值的候选文本信息作为第二候选文本信息，得到第二候选文本信息集合。
第三步，通过Sunday算法从上述候选文本信息集合中筛选出包含上述价格相关标识符信息集合中任一价格相关标识符信息的候选文本信息作为第三候选文本信息。
步骤204，基于第一候选文本信息、第二候选文本信息集合和第三候选文本信息，生成第一距离值集合和第二距离值集合。
在一些实施例中，上述执行主体基于第一候选文本信息、第二候选文本信息集合和第三候选文本信息，生成第一距离值集合和第二距离值集合，可以包括以下步骤：
第一步，通过两点间距离公式确定上述第一候选文本信息对应的标识框的左上角的坐标值与上述第二候选文本信息集合中每个第二候选文本信息对应的标识框的左上角的坐标值之间的距离值以生成第一距离值，得到第一距离值集合。其中，距离值可以是欧式距离等等。
第二步，通过两点间距离公式确定上述第三候选文本信息对应的标识框的左上角的坐标值与上述第二候选文本信息集合中每个第二候选文本信息对应的标识框的左上角的坐标值之间的距离值以生成第二距离值，得到第二距离值集合。
步骤205，基于第二候选文本信息集合、第一距离值集合和第二距离值集合，生成候选价格信息。
在一些实施例中，上述执行主体基于第二候选文本信息集合、第一距离值集合和第二距离值集合，生成候选价格信息。可以是从上述第二候选文本信息集合中筛选出对应的第一距离值和第二距离值分别为上述第一距离值集合和上述第二距离值集合中最大值的第二候选文本信息作为候选价格信息。
本公开的上述各个实施例具有如下有益效果：通过本公开的一些实施例的基于图像识别的价格信息处理方法生成的价格信息的准确度有所提高。具体来说，造成生成的价格信息不够精确的原因在于：采用爬虫程序爬取的价格信息往往与新品预售等场景下的实际价格不 符。基于此，本公开的一些实施例的基于图像识别的价格信息处理方法通过获取物品展示图像中的文本识别信息集合。然后，根据预设条件筛选出符合预设条件的文本识别信息。在实际情况中，价格信息的显示位置一般与价格标识关键字信息或价格相关标识符信息的显示位置之间的距离较近，通过距离进行筛选以得到物品在对应的新品预售等场景下的价格信息。由于在新品预售等场景下，物品对应的价格信息往往会预先更新在物品展示图像中，而未同步地在显示页面中进行更新。因此，此种方法通过物品展示图像中的文本识别信息集合确定物品的价格信息，能够较好的表征新品预售等场景下的物品对应的价格。从而，提高了生成的价格信息的准确度。
进一步参考图4，其示出了基于图像识别的价格信息处理方法的另一些实施例的流程400。该基于图像识别的价格信息处理方法的流程400，包括以下步骤：
步骤401，从物品展示图像中提取文本信息作为文本识别信息，得到文本识别信息集合。
在一些实施例中，上述执行主体可以通过RNN(Recurrent Neural Network，循环神经网络)模型，LSTM(Long Short-Term Memory，长短期记忆网络)模型从上述物品展示图像中识别并提取出文本信息以生成文本识别信息集合。其中，上述文本信息可以包括但不限于以下至少一项：文本内容信息，文本置信度数值，标识框尺寸信息。上述标识框尺寸信息可以包括但不限于以下至少一项：标识框中心坐标值，标识框长度值，标识框宽度值。
作为示例，上述文本识别信息集合可以是[[六档大风，0.29652，(80，80)，40，45]，[五叶大风量高效节能，0.99917，(240，80)，100，45]]。
步骤402，从文本识别信息集合中筛选出满足第一预设条件组的文本识别信息作为候选文本信息，得到候选文本信息集合。
在一些实施例中，上述执行主体可以从文本识别信息集合中筛选出满足第一预设条件组的文本识别信息作为候选文本信息，得到候选 文本信息集合。其中，上述第一预设条件组可以包括但不限于一下至少一项：文本识别信息包括的文本置信度数值大于等于预设置信度数值，文本识别信息包括的标识框长度值大于等于预设标识框长度值且标识框宽度值大于等于预设标识框宽度值。
作为示例，上述预设置信度数值可以是0.1。上述预设标识框长度值可以是15(像素)。上述预设标识框宽度值可以是15(像素)。
步骤403，从候选文本信息集合中筛选出文本内容信息中包含价格标识关键字信息集合中任一价格标识关键字的候选文本信息作为第一候选文本信息。
在一些实施例中，上述执行主体可以从候选文本信息集合中筛选出文本内容信息中包含价格标识关键字信息集合中任一价格标识关键字的候选文本信息作为第一候选文本信息。其中，上述价格标识关键字信息可以包括但不限于一下至少一项：“新品价”，“搭配价”，“体验价”，“正品价”，“上新价”，“划算价”，“品牌价”，“尊享价”。
作为示例，上述执行主体可以通过KMP(Knuth-Morris-Pratt，模式匹配)算法从上述候选文本信息集合中筛选出文本内容信息中包含价格标识关键字信息集合中任一价格标识关键字的候选文本信息作为第一候选文本信息。
步骤404，从候选文本信息集合中筛选出文本内容信息中包含数值的候选文本信息作为第二候选文本信息，得到第二候选文本信息集合。
在一些实施例中，上述执行主体可以通过各种方式从候选文本信息集合中筛选出文本内容信息中包含价格标识关键字信息集合中任一价格标识关键字的候选文本信息作为第一候选文本信息。
作为示例，上述执行主体可以通过Rabin–Karp算法从候选文本信息集合中筛选出文本内容信息中包含价格标识关键字信息集合中任一价格标识关键字的候选文本信息作为第一候选文本信息。
步骤405，从候选文本信息集合中筛选出文本内容信息中包含价格相关标识符信息集合中任一价格相关标识符信息的候选文本信息作为第三候选文本信息。
在一些实施例中，上述执行主体可以通过步骤202，步骤403，步骤和404中提及的任一算法从候选文本信息集合中筛选出文本内容信息中包含价格相关标识符信息集合中任一价格相关标识符信息的候选文本信息作为第三候选文本信息。其中，上述价格相关标识符信息集合可以包括以下至少一项：“元”，“$”，“￥”，“￡”。
步骤406，将第一候选文本信息包括的标识框中心坐标值与第二候选文本信息集合中每个第二候选文本信息包括的标识框中心坐标值的距离值确定为第一距离值，得到第一距离值集合。
在一些实施例中，上述执行主体可以通过两点间距离公式确定上述第一候选文本信息包括的标识框中心坐标值与上述第二候选文本信息集合中每个第二候选文本信息包括的标识框中心坐标值的距离值以生成第一距离值，得到第一距离值集合。
步骤407，将第三候选文本信息包括的标识框中心坐标值与第二候选文本信息集合中每个第二候选文本信息包括的标识框中心坐标值的距离值确定为第二距离值，得到第二距离值集合。
在一些实施例中，上述执行主体可以通过两点间距离公式确定上述第三候选文本信息包括的标识框中心坐标值与上述第二候选文本信息集合中每个第二候选文本信息包括的标识框中心坐标值的距离值以生成第二距离值，得到第二距离值集合。
步骤408，从第二候选文本信息集合中筛选出对应的第一距离值满足第二预设条件的第二候选文本信息作为第一候选价格信息，得到第一候选价格信息集合。
在一些实施例中，上述执行主体可以从第二候选文本信息集合中筛选出对应的第一距离值满足第二预设条件的第二候选文本信息作为第一候选价格信息，得到第一候选价格信息集合。其中，上述第二预设条件也可是第一距离值小于等于上述第一距离值集合中各个第一距离值的均值。
步骤409，从第一候选价格信息集合中筛选出对应的第二距离值满足第三预设条件的第一候选价格信息作为候选价格信息。
在一些实施例中，上述执行主体可以从第一候选价格信息集合中 筛选出对应的第二距离值满足第三预设条件的第一候选价格信息作为候选价格信息。其中，上述第三预设条件可以是第二距离值等于上述第二距离值集合中的最小的第二距离值。
可选的，上述执行主体可以响应于确定候选文本信息集合中存在包括预设的价格相关关键字信息集合中任一价格相关关键字信息的候选文本信息，基于候选文本信息集合和价格相关关键字信息集合，生成第四候选文本信息。其中，上述价格相关关键字信息集合可以包括以下至少一项：“券”，“折”，“立减”，“满减”。
可选的，上述执行主体可以将第四候选文本信息包括的标识框中心坐标值与第二候选文本信息集合中每个第二候选文本信息包括的标识框中心坐标值的距离值确定为第三距离值，得到第三距离值集合。其中，可以通过两点间距离公式确定上述第四候选文本信息包括的标识框中心坐标值与上述第二候选文本信息集合中每个第二候选文本信息包括的标识框中心坐标之间的距离值。
可选的，上述执行主体可以从第二候选文本信息集合中筛选出对应的第三距离值满足第四预设条件的第二候选文本信息作为价格相关文本信息。其中，上述第四预设条件可以是第三距离值与上述第三距离值中的最小值相同。
可选的，上述执行主体可以响应于确定候选价格信息与页面价格信息的差值在预设范围内，将候选价格信息确定为待呈现价格信息。其中，上述预设范围可以是[0，100]。上述页面价格信息可以是通过爬虫程序从相应的物品的展示页面中爬取得到的。
可选的，上述执行主体可以基于上述待呈现价格信息和上述价格相关文本信息，生成呈现价格信息。
作为示例，上述待呈现价格信息可以是“399元”，上述价格相关文本信息可以是“满300元-40元”。因此，生成的呈现价格信息可以是“359元”。
作为又一示例，上述待呈现价格信息可以是“99元”，上述价格相关文本信息可以是“满200元-40元”。因此，生成的呈现价格信息可以是“99元”。
本公开的上述各个实施例具有如下有益效果：文本置信度数值较低的文本识别信息可以理解为文本信息识别的不够精准。同时，标识框长度值和标识框宽度值较小可以理解为标识框内部的文本信息所占的区域面积较小，实际情况中，此类文本信息一般为标点符号等，对于后续的价格信息的生成影响较小，因此通过上述第一预设条件组对获取到的文本识别信息进行初次筛选，过滤掉文本置信度数值、标识框长度值和标识框宽度值不满足条件的文本识别信息。减少了计算资源的消耗。此外，由于从物品展示图像中识别出的文本识别信息集合中包含价格标识关键字，数值或价格相关标识符信息中任一一项的文本识别信息对最后的价格生成影响较大。因此通过条件筛选得到第一候选文本信息、第二候选文本信息集合和第三候选文本信息。从而进一步的过滤掉对最后的价格生成影响较小信息，减少了计算资源的消耗。然后，在实际情况中，在物品展示图像中，在包含价格标识关键字信息的文本识别信息和包含价格相关标识符信息的文本识别信息附近的数值往往是价格信息。基于此，通过距离进行进一步的筛选，从而生成候选价格信息。此种方式相比于使用爬虫程序爬取页面中的价格信息作为价格信息。能够更好的适应新品预售等场景下的价格信息的识别和生成。从而，提高了生成的价格信息的准确度。
进一步参考图5，作为对上述各图所示方法的实现，本公开提供了一种基于图像识别的价格信息处理装置的一些实施例，这些装置实施例与图2所示的那些方法实施例相对应，该装置具体可以应用于各种电子设备中。
如图5所示，一些实施例的基于图像识别的价格信息处理装置500包括：提取单元501、筛选单元502、第一生成单元503、第二生成单元504和第三生成单元505。其中，提取单元501，被配置成从物品展示图像中提取文本信息作为文本识别信息，得到文本识别信息集合。筛选单元502，被配置成从文本识别信息集合中筛选出满足第一预设条件组的文本识别信息作为候选文本信息，得到候选文本信息集合。第一生成单元503，被配置成响应于确定上述候选文本信息集合中存 在包括预设的价格标识关键字信息集合中任一价格标识关键字信息的候选文本信息，基于上述候选文本信息集合、上述价格标识关键字信息集合和预设的价格相关标识符信息集合，生成第一候选文本信息、第二候选文本信息集合和第三候选文本信息。第二生成单元504，被配置成基于上述第一候选文本信息、上述第二候选文本信息集合和上述第三候选文本信息，生成第一距离值集合和第二距离值集合。第三生成单元505，被配置成基于上述第二候选文本信息集合、上述第一距离值集合和上述第二距离值集合，生成候选价格信息。
在一些实施例的可选的实现方式中，第一生成单元503进一步被配置成：从上述候选文本信息集合中筛选出文本内容信息中包含上述价格标识关键字信息集合中任一价格标识关键字的候选文本信息作为第一候选文本信息。
在一些实施例的可选的实现方式中，第一生成单元503进一步被配置成：从上述候选文本信息集合中筛选出文本内容信息中包含数值的候选文本信息作为第二候选文本信息，得到第二候选文本信息集合。
在一些实施例的可选的实现方式中，第一生成单元503进一步被配置成：从上述候选文本信息集合中筛选出文本内容信息中包含上述价格相关标识符信息集合中任一价格相关标识符信息的候选文本信息作为第三候选文本信息。
在一些实施例的可选的实现方式中，第二生成单元504进一步被配置成：将上述第一候选文本信息包括的标识框中心坐标值与上述第二候选文本信息集合中每个第二候选文本信息包括的标识框中心坐标值的距离值确定为第一距离值，得到第一距离值集合；将上述第三候选文本信息包括的标识框中心坐标值与上述第二候选文本信息集合中每个第二候选文本信息包括的标识框中心坐标值的距离值确定为第二距离值，得到第二距离值集合。
在一些实施例的可选的实现方式中，第三生成单元505进一步被配置成：从上述第二候选文本信息集合中筛选出对应的第一距离值满足第二预设条件的第二候选文本信息作为第一候选价格信息，得到第一候选价格信息集合；从上述第一候选价格信息集合中筛选出对应的 第二距离值满足第三预设条件的第一候选价格信息作为候选价格信息。
在一些实施例的可选实现方式中，装置500还包括：第四生成单元，被配置成响应于确定上述候选文本信息集合中存在包括预设的价格相关关键字信息集合中任一价格相关关键字信息的候选文本信息，基于上述候选文本信息集合，上述价格相关关键字信息集合，生成第四候选文本信息。
在一些实施例的可选的实现方式中，第四生成单元被进一步配置成：从上述候选文本信息集合中筛选出文本内容信息包含上述价格相关关键字信息集合中任一价格相关关键字信息的候选文本信息作为第四候选文本信息。
在一些实施例的可选的实现方式中，装置500还包括：将上述第四候选文本信息包括的标识框中心坐标值与上述第二候选文本信息集合中每个第二候选文本信息包括的标识框中心坐标值的距离值确定为第三距离值，得到第三距离值集合。
在一些实施例的可选的实现方式中，装置500还包括：响应于确定上述候选价格信息与页面价格信息的差值在预设范围内，将上述候选价格信息确定为待呈现价格信息。
在一些实施例的可选的实现方式中，装置500还包括：基于上述待呈现价格信息和上述价格相关文本信息，生成呈现价格信息。
可以理解的是，该装置500中记载的诸单元与参考图2描述的方法中的各个步骤相对应。同时，考虑到物品对应的展示图片中可能包括满减信息。首先，通过第四生成单元识别出满减关键字。然后，通过规则过滤以及距离计算，进而确定满减数值。从而得到最终的价格信息，提高了价格信息生成的准确度。
下面参考图6，其示出了适于用来实现本公开的一些实施例的电子设备(如图1所示的计算设备101)600的结构示意图。图6示出的电子设备仅仅是一个示例，不应对本公开的实施例的功能和使用范围带来任何限制。
如图6所示，电子设备600可以包括处理装置(例如中央处理器、图形处理器等)601，其可以根据存储在只读存储器(ROM)602中的程序或者从存储装置608加载到随机访问存储器(RAM)603中的程序而执行各种适当的动作和处理。在RAM 603中，还存储有电子设备600操作所需的各种程序和数据。处理装置601、ROM 602以及RAM 603通过总线604彼此相连。输入/输出(I/O)接口605也连接至总线604。
通常，以下装置可以连接至I/O接口605：包括例如触摸屏、触摸板、键盘、鼠标、摄像头、麦克风、加速度计、陀螺仪等的输入装置606；包括例如液晶显示器(LCD)、扬声器、振动器等的输出装置607；包括例如磁带、硬盘等的存储装置608；以及通信装置609。通信装置609可以允许电子设备600与其他设备进行无线或有线通信以交换数据。虽然图6示出了具有各种装置的电子设备600，但是应理解的是，并不要求实施或具备所有示出的装置。可以替代地实施或具备更多或更少的装置。图6中示出的每个方框可以代表一个装置，也可以根据需要代表多个装置。
特别地，根据本公开的一些实施例，上文参考流程图描述的过程可以被实现为计算机软件程序。例如，本公开的一些实施例包括一种计算机程序产品，其包括承载在计算机可读介质上的计算机程序，该计算机程序包含用于执行流程图所示的方法的程序代码。在这样的一些实施例中，该计算机程序可以通过通信装置609从网络上被下载和安装，或者从存储装置608被安装，或者从ROM 602被安装。在该计算机程序被处理装置601执行时，执行本公开的一些实施例的方法中限定的上述功能。
需要说明的是，本公开的一些实施例中记载的计算机可读介质可以是计算机可读信号介质或者计算机可读存储介质或者是上述两者的任意组合。计算机可读存储介质例如可以是——但不限于——电、磁、光、电磁、红外线、或半导体的系统、装置或器件，或者任意以上的组合。计算机可读存储介质的更具体的例子可以包括但不限于：具有一个或多个导线的电连接、便携式计算机磁盘、硬盘、随机访问存储 器(RAM)、只读存储器(ROM)、可擦式可编程只读存储器(EPROM或闪存)、光纤、便携式紧凑磁盘只读存储器(CD-ROM)、光存储器件、磁存储器件、或者上述的任意合适的组合。在本公开的一些实施例中，计算机可读存储介质可以是任何包含或存储程序的有形介质，该程序可以被指令执行系统、装置或者器件使用或者与其结合使用。而在本公开的一些实施例中，计算机可读信号介质可以包括在基带中或者作为载波一部分传播的数据信号，其中承载了计算机可读的程序代码。这种传播的数据信号可以采用多种形式，包括但不限于电磁信号、光信号或上述的任意合适的组合。计算机可读信号介质还可以是计算机可读存储介质以外的任何计算机可读介质，该计算机可读信号介质可以发送、传播或者传输用于由指令执行系统、装置或者器件使用或者与其结合使用的程序。计算机可读介质上包含的程序代码可以用任何适当的介质传输，包括但不限于：电线、光缆、RF(射频)等等，或者上述的任意合适的组合。
在一些实施方式中，客户端、服务器可以利用诸如HTTP(HyperText Transfer Protocol，超文本传输协议)之类的任何当前已知或未来研发的网络协议进行通信，并且可以与任意形式或介质的数字数据通信(例如，通信网络)互连。通信网络的示例包括局域网(“LAN”)，广域网(“WAN”)，网际网(例如，互联网)以及端对端网络(例如，ad hoc端对端网络)，以及任何当前已知或未来研发的网络。
上述计算机可读介质可以是上述电子设备中所包含的；也可以是单独存在，而未装配入该电子设备中。上述计算机可读介质承载有一个或者多个程序，当上述一个或者多个程序被该电子设备执行时，使得该电子设备：从物品展示图像中提取文本信息作为文本识别信息，得到文本识别信息集合。从上述文本识别信息集合中筛选出满足第一预设条件组的文本识别信息作为候选文本信息，得到候选文本信息集合。响应于确定上述候选文本信息集合中存在包括预设的价格标识关键字信息集合中任一价格标识关键字信息的候选文本信息，基于上述候选文本信息集合、上述价格标识关键字信息集合和预设的价格相关 标识符信息集合，生成第一候选文本信息、第二候选文本信息集合和第三候选文本信息。基于上述第一候选文本信息、上述第二候选文本信息集合和上述第三候选文本信息，生成第一距离值集合和第二距离值集合。基于上述第二候选文本信息集合、上述第一距离值集合和上述第二距离值集合，生成候选价格信息。
可以以一种或多种程序设计语言或其组合来编写用于执行本公开的一些实施例的操作的计算机程序代码，上述程序设计语言包括面向对象的程序设计语言—诸如Java、Smalltalk、C++，还包括常规的过程式程序设计语言—诸如“C”语言或类似的程序设计语言。程序代码可以完全地在用户计算机上执行、部分地在用户计算机上执行、作为一个独立的软件包执行、部分在用户计算机上部分在远程计算机上执行、或者完全在远程计算机或服务器上执行。在涉及远程计算机的情形中，远程计算机可以通过任意种类的网络——包括局域网(LAN)或广域网(WAN)——连接到用户计算机，或者，可以连接到外部计算机(例如利用因特网服务提供商来通过因特网连接)。
附图中的流程图和框图，图示了按照本公开各种实施例的系统、方法和计算机程序产品的可能实现的体系架构、功能和操作。在这点上，流程图或框图中的每个方框可以代表一个模块、程序段、或代码的一部分，该模块、程序段、或代码的一部分包含一个或多个用于实现规定的逻辑功能的可执行指令。也应当注意，在有些作为替换的实现中，方框中所标注的功能也可以以不同于附图中所标注的顺序发生。例如，两个接连地表示的方框实际上可以基本并行地执行，它们有时也可以按相反的顺序执行，这依所涉及的功能而定。也要注意的是，框图和/或流程图中的每个方框、以及框图和/或流程图中的方框的组合，可以用执行规定的功能或操作的专用的基于硬件的系统来实现，或者可以用专用硬件与计算机指令的组合来实现。
描述于本公开的一些实施例中的单元可以通过软件的方式实现，也可以通过硬件的方式来实现。所描述的单元也可以设置在处理器中，例如，可以描述为：一种处理器包括提取单元、筛选单元、第一生成单元、第二生成单元和第三生成单元。其中，这些单元的名称在某种 情况下并不构成对该单元本身的限定，例如，筛选单元还可以被描述为“从文本识别信息集合中筛选出满足第一预设条件组的文本识别信息作为候选文本信息，得到候选文本信息集合的单元”。
本文中以上描述的功能可以至少部分地由一个或多个硬件逻辑部件来执行。例如，非限制性地，可以使用的示范类型的硬件逻辑部件包括：现场可编程门阵列(FPGA)、专用集成电路(ASIC)、专用标准产品(ASSP)、片上系统(SOC)、复杂可编程逻辑设备(CPLD)等等。
以上描述仅为本公开的一些较佳实施例以及对所运用技术原理的说明。本领域技术人员应当理解，本公开的实施例中所涉及的发明范围，并不限于上述技术特征的特定组合而成的技术方案，同时也应涵盖在不脱离上述发明构思的情况下，由上述技术特征或其等同特征进行任意组合而形成的其它技术方案。例如上述特征与本公开的实施例中公开的(但不限于)具有类似功能的技术特征进行互相替换而形成的技术方案。

Claims (17)

  1. 一种基于图像识别的价格信息处理方法，包括：
    从物品展示图像中提取文本信息作为文本识别信息，得到文本识别信息集合；
    从所述文本识别信息集合中筛选出满足第一预设条件组的文本识别信息作为候选文本信息，得到候选文本信息集合；
    响应于确定所述候选文本信息集合中存在包括预设的价格标识关键字信息集合中任一价格标识关键字信息的候选文本信息，基于所述候选文本信息集合、所述价格标识关键字信息集合和预设的价格相关标识符信息集合，生成第一候选文本信息、第二候选文本信息集合和第三候选文本信息；
    基于所述第一候选文本信息、所述第二候选文本信息集合和所述第三候选文本信息，生成第一距离值集合和第二距离值集合；
    基于所述第二候选文本信息集合、所述第一距离值集合和所述第二距离值集合，生成候选价格信息。
  2. 根据权利要求1所述的方法，其中，所述方法还包括：
    响应于确定所述候选文本信息集合中存在包括预设的价格相关关键字信息集合中任一价格相关关键字信息的候选文本信息，基于所述候选文本信息集合，所述价格相关关键字信息集合，生成第四候选文本信息。
  3. 根据权利要求1-2中任一所述的方法，其中，所述文本识别信息包括：文本内容信息，文本置信度数值，标识框尺寸信息；所述标识框尺寸信息包括：标识框中心坐标值，标识框长度值，标识框宽度值。
  4. 根据权利要求1-3中任一所述的方法，其中，所述第一预设条件组包括以下至少一项：
    文本识别信息包括的文本置信度数值大于等于预设置信度数值；文本识别信息包括的标识框长度值大于等于预设标识框长度值且标识框宽度值大于等于预设标识框宽度值。
  5. 根据权利要求1-3中任一所述的方法，其中，所述基于所述候选文本信息集合、所述价格标识关键字信息集合和预设的价格相关标识符信息集合，生成第一候选文本信息、第二候选文本信息集合和第三候选文本信息，包括：
    从所述候选文本信息集合中筛选出文本内容信息中包含所述价格标识关键字信息集合中任一价格标识关键字的候选文本信息作为第一候选文本信息。
  6. 根据权利要求1-3中任一所述的方法，其中，所述基于所述候选文本信息集合、所述价格标识关键字信息集合和预设的价格相关标识符信息集合，生成第一候选文本信息、第二候选文本信息集合和第三候选文本信息，包括：
    从所述候选文本信息集合中筛选出文本内容信息中包含数值的候选文本信息作为第二候选文本信息，得到第二候选文本信息集合。
  7. 根据权利要求1-3中任一所述的方法，其中，所述基于所述候选文本信息集合、所述价格标识关键字信息集合和预设的价格相关标识符信息集合，生成第一候选文本信息、第二候选文本信息集合和第三候选文本信息，包括：
    从所述候选文本信息集合中筛选出文本内容信息中包含所述价格相关标识符信息集合中任一价格相关标识符信息的候选文本信息作为第三候选文本信息。
  8. 根据权利要求1-3中任一所述的方法，其中，所述基于所述第一候选文本信息、所述第二候选文本信息集合和所述第三候选文本信息，生成第一距离值集合和第二距离值集合，包括：
    将所述第一候选文本信息包括的标识框中心坐标值与所述第二候选文本信息集合中每个第二候选文本信息包括的标识框中心坐标值的距离值确定为第一距离值，得到第一距离值集合；
    将所述第三候选文本信息包括的标识框中心坐标值与所述第二候选文本信息集合中每个第二候选文本信息包括的标识框中心坐标值的距离值确定为第二距离值，得到第二距离值集合。
  9. 根据权利要求1-3中任一所述的方法，其中，所述基于所述第二候选文本信息集合、所述第一距离值集合和所述第二距离值集合，生成候选价格信息，包括：
    从所述第二候选文本信息集合中筛选出对应的第一距离值满足第二预设条件的第二候选文本信息作为第一候选价格信息，得到第一候选价格信息集合；
    从所述第一候选价格信息集合中筛选出对应的第二距离值满足第三预设条件的第一候选价格信息作为候选价格信息。
  10. 根据权利要求1-9中任一所述的方法，其中，所述基于所述候选文本信息集合和所述价格相关关键字信息集合，生成第四候选文本信息，包括：
    从所述候选文本信息集合中筛选出文本内容信息包含所述价格相关关键字信息集合中任一价格相关关键字信息的候选文本信息作为第四候选文本信息。
  11. 根据权利要求3-10中任一所述的方法，其中，所述方法还包括：
    将所述第四候选文本信息包括的标识框中心坐标值与所述第二候选文本信息集合中每个第二候选文本信息包括的标识框中心坐标值的距离值确定为第三距离值，得到第三距离值集合。
  12. 根据权利要求11所述的方法，其中，所述方法还包括：
    从所述第二候选文本信息集合中筛选出对应的第三距离值满足第四预设条件的第二候选文本信息作为价格相关文本信息。
  13. 根据权利要求1-12中任一所述的方法，其中，所述方法还包括：
    响应于确定所述候选价格信息与页面价格信息的差值在预设范围内，将所述候选价格信息确定为待呈现价格信息。
  14. 根据权利要求13所述的方法，其中，所述方法还包括：
    基于所述待呈现价格信息和所述价格相关文本信息，生成呈现价格信息。
  15. 一种基于图像识别的价格信息处理装置，包括：
    提取单元，被配置成从物品展示图像中提取文本信息作为文本识别信息，得到文本识别信息集合；
    筛选单元，被配置成从所述文本识别信息集合中筛选出满足第一预设条件组的文本识别信息作为候选文本信息，得到候选文本信息集合；
    第一生成单元，被配置成响应于确定所述候选文本信息集合中存在包括预设的价格标识关键字信息集合中任一价格标识关键字信息的候选文本信息，基于所述候选文本信息集合、所述价格标识关键字信息集合和预设的价格相关标识符信息集合，生成第一候选文本信息、第二候选文本信息集合和第三候选文本信息；
    第二生成单元，被配置成基于所述第一候选文本信息、所述第二候选文本信息集合和所述第三候选文本信息，生成第一距离值集合和第二距离值集合；
    第三生成单元，被配置成基于所述第二候选文本信息集合、所述第一距离值集合和所述第二距离值集合，生成候选价格信息。
  16. 一种电子设备，包括：
    一个或多个处理器；
    存储装置，其上存储有一个或多个程序，
    当所述一个或多个程序被所述一个或多个处理器执行，使得所述一个或多个处理器实现如权利要求1-14中任一所述的方法。
  17. 一种计算机可读介质，其上存储有计算机程序，其中，所述程序被处理器执行时实现如权利要求1-14中任一所述的方法。
PCT/CN2021/125399 2020-11-10 2021-10-21 基于图像识别的价格信息处理方法、装置、设备和介质 WO2022100401A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202011246846.5A CN112308678A (zh) 2020-11-10 2020-11-10 基于图像识别的价格信息处理方法、装置、设备和介质
CN202011246846.5 2020-11-10

Publications (1)

Publication Number Publication Date
WO2022100401A1 true WO2022100401A1 (zh) 2022-05-19

Family

ID=74326652

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/125399 WO2022100401A1 (zh) 2020-11-10 2021-10-21 基于图像识别的价格信息处理方法、装置、设备和介质

Country Status (2)

Country Link
CN (1) CN112308678A (zh)
WO (1) WO2022100401A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112308678A (zh) * 2020-11-10 2021-02-02 北京沃东天骏信息技术有限公司 基于图像识别的价格信息处理方法、装置、设备和介质

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20140244392A1 (en) * 2013-02-27 2014-08-28 Xue Si Xing Digital Marketing Co., Ltd. Graphical Recognition Inventory Management and Marketing System
CN105912642A (zh) * 2016-04-08 2016-08-31 世纪禾光科技发展（北京）有限公司 产品价格数据采集方法及系统
CN110610166A (zh) * 2019-09-18 2019-12-24 北京猎户星空科技有限公司 文本区域检测模型训练方法、装置、电子设备和存储介质
CN112308678A (zh) * 2020-11-10 2021-02-02 北京沃东天骏信息技术有限公司 基于图像识别的价格信息处理方法、装置、设备和介质

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20140244392A1 (en) * 2013-02-27 2014-08-28 Xue Si Xing Digital Marketing Co., Ltd. Graphical Recognition Inventory Management and Marketing System
CN105912642A (zh) * 2016-04-08 2016-08-31 世纪禾光科技发展（北京）有限公司 产品价格数据采集方法及系统
CN110610166A (zh) * 2019-09-18 2019-12-24 北京猎户星空科技有限公司 文本区域检测模型训练方法、装置、电子设备和存储介质
CN112308678A (zh) * 2020-11-10 2021-02-02 北京沃东天骏信息技术有限公司 基于图像识别的价格信息处理方法、装置、设备和介质

Also Published As

Publication number Publication date
CN112308678A (zh) 2021-02-02

Similar Documents

Publication Publication Date Title
US11645316B2 (en) Question answering method and language model training method, apparatus, device, and storage medium
JP6397144B2 (ja) 画像からの事業発見
US20230394671A1 (en) Image segmentation method and apparatus, and device, and storage medium
US20220253631A1 (en) Image processing method, electronic device and storage medium
JP7394809B2 (ja) ビデオを処理するための方法、装置、電子機器、媒体及びコンピュータプログラム
WO2023077995A1 (zh) 信息提取方法、装置、设备、介质及产品
CN113420757B (zh) 文本审核方法、装置、电子设备和计算机可读介质
CN111522927A (zh) 基于知识图谱的实体查询方法和装置
CN112749300B (zh) 用于视频分类的方法、装置、设备、存储介质和程序产品
CN113468330B (zh) 信息获取方法、装置、设备及介质
CN112766284B (zh) 图像识别方法和装置、存储介质和电子设备
WO2022100401A1 (zh) 基于图像识别的价格信息处理方法、装置、设备和介质
CN113610034B (zh) 识别视频中人物实体的方法、装置、存储介质及电子设备
CN111160410A (zh) 一种物体检测方法和装置
US20230281696A1 (en) Method and apparatus for detecting false transaction order
WO2022017299A1 (zh) 一种文本检测方法、装置、电子设备及存储介质
US10341272B2 (en) Personality reply for digital content
CN112348615B (zh) 用于审核信息的方法和装置
CN113779276A (zh) 用于检测评论的方法和装置
WO2023130925A1 (zh) 字体识别方法、装置、可读介质及电子设备
CN115525781A (zh) 多模态虚假信息检测方法、装置和设备
CN113807056B (zh) 一种文档名称序号纠错方法、装置和设备
CN112035581A (zh) 基于模型的任务处理方法、装置、设备和介质
CN111932323B (zh) 物品信息界面显示方法、装置、设备和计算机可读介质
US20180293315A1 (en) Device for multiple condition search based on knowledge points

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: 21890928

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: 21890928

Country of ref document: EP

Kind code of ref document: A1