Commodity information identification method and device, program product, electronic equipment and storage medium
By obtaining product information and product certificate pictures entered by users, using optical character recognition and text reasoning models to extract the image content into structured text, and comparing the information entered by users with the extracted structured text, the problem of low audit efficiency in the existing technology is solved, and faster and more accurate product information review is achieved.
Patent Information
- Application Number
- CN202510362140.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-03-26
- Publication Date
- 2025-06-17
AI Technical Summary
In the prior art, the review efficiency of product information is low and relies on manual identification, resulting in inefficiency.
By obtaining product information and product certificate pictures entered by the user, the image content is extracted into structured text using optical character recognition and text reasoning models, and the information entered by the user is compared with the extracted structured text to determine the correctness of product information.
It improves the efficiency of product information review, and compares content through computer programs, which is faster and more accurate than manual review.
Smart Images
Figure CN120164228A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of image recognition technology, and more specifically, to a commodity information recognition method and device, a program product, an electronic device, and a storage medium. Background Art
[0002] On the Internet e-commerce platform, merchants will submit product information to auditors for review to determine the accuracy and compliance of product content. Among them, the fabric, item number, brand, price, etc. of clothing products are key information of the products. Merchants usually fill in these contents and attach the product certificate; auditors are responsible for reviewing whether the fabric, item number, brand, price and other information filled in by the merchant are consistent with the relevant content on the certificate picture. If not, the correct information needs to be fed back to the merchant in text form. However, the traditional audit method is highly dependent on manual recognition, so the audit efficiency is low. Summary of the invention
[0003] The purpose of the embodiments of the present application is to provide a commodity information identification method and device, program product, electronic device, and storage medium to solve the technical problem of low audit efficiency in the prior art.
[0004] In a first aspect, an embodiment of the present application provides a method for identifying product information, comprising: obtaining first text content and a product certificate image, wherein the first text content includes product information input by a user; determining a corresponding second text content based on the product certificate image, wherein the second text content is a structured text; and comparing the first text content with the second text content to determine whether the product information is correct.
[0005] In the above scheme, the content of the product certificate image is structured and extracted to obtain the second text content of the structured text; the first text content corresponding to the product information input by the user is compared with the second text content to determine whether the product information is correct. Among them, since the second text content of the structured text is extracted based on the product certificate image, it is convenient to compare the content. At the same time, since a computer program is used for content comparison, the efficiency of product information review can be improved compared with the manual review method in the prior art.
[0006] In an alternative embodiment, if the product information includes fabric information, comparing the first text content with the second text content includes: standardizing the first fabric content and splicing the second fabric content, where the first fabric content is the content in the first text content corresponding to the fabric information, and the second fabric is the content in the second text content corresponding to the fabric information; determining whether the spliced second fabric content is included in the processed first fabric content, where the fact that the spliced second fabric content is not included in the processed first fabric content indicates that the product information is incorrect. In the above solution, inclusive detection is performed on each of the multiple strings obtained by splicing the second fabric content with the string obtained by standardizing the first fabric content. If any one of the multiple strings is included in the string obtained by standardizing the first fabric content, it is considered that the fabric passes the detection; if all the multiple strings can pass the detection in this way, it is considered that the fabric information input by the user is correct, otherwise it is considered that the fabric information input by the user is incorrect.
[0007] In an alternative embodiment, if the product information includes goods information, comparing the first text content with the second text content includes: converting the first goods content and the second goods content into lowercase letters, where the first goods content is the content in the first text content corresponding to the goods information, and the second goods content is the content in the second text content corresponding to the goods information; determining whether the converted first goods content is the same as the converted second goods content, where the fact that the converted first goods content is not the same as the converted second goods content indicates that the product information is incorrect. In the above solution, the first goods content is uniformly converted into lowercase letters, and the second goods content is uniformly converted into lowercase letters, and then it is determined whether the strings are the same. If they are the same, it is considered that the goods information input by the user is correct, otherwise it is considered that the goods information input by the user is incorrect.
[0008] In an alternative embodiment, if the product information includes price information, comparing the first text content with the second text content includes: determining whether the first price content is equal to the second price content, where the first price content is the content in the first text content corresponding to the price information, and the second price content is the content in the second text content corresponding to the price information, and the fact that the first price content is not equal to the second price content indicates that the product information is incorrect. In the above solution, a numerical comparison is made between the first price content and the second price content. If they are equal, it is considered that the price information input by the user is correct, otherwise it is considered that the price information input by the user is incorrect.
[0009] In an alternative embodiment, determining the corresponding second text content according to the product certificate picture includes: using optical character recognition to extract the text in the product certificate picture to obtain the corresponding third text content; inputting the third text content into a pre-trained text reasoning model to obtain the second text content output by the text reasoning model. In the above solution, by integrating optical character recognition technology and a text reasoning model, the second text content is extracted from the product certificate picture. Among them, optical character recognition technology can make the text content extracted from the product certificate picture more accurate; through the structured extraction ability of the text reasoning model, the text content obtained from the product certificate picture is more structured and easier to be parsed and used by computer programs.
[0010] In an alternative embodiment, after comparing the first text content with the second text content to determine whether the product information is correct, the method further includes: generating a modification suggestion for the product information according to the comparison result. In the above solution, after determining whether the product information is correct, a modification suggestion for the product information can be further generated according to the comparison result, so that the user can quickly modify the product information based on the above modification suggestion.
[0011] In a second aspect, an embodiment of the present application provides a product information recognition device, including: an acquisition module, configured to acquire first text content and a product certificate picture, where the first text content includes product information input by a user; a determination module, configured to determine corresponding second text content according to the product certificate picture, where the second text content is structured text; a comparison module, configured to compare the first text content with the second text content to determine whether the product information is correct.
[0012] In the above solution, the content of the product certificate picture is extracted in a structured manner to obtain the second text content of structured text; by comparing the first text content corresponding to the product information input by the user with the above second text content, it is determined whether the product information is correct. Among them, since the second text content of structured text is extracted based on the product certificate picture, it is convenient for content comparison. At the same time, since computer programs are used for content comparison, compared with the manual review method in the prior art, the review efficiency of product information can be improved.
[0013] In an alternative embodiment, if the product information includes fabric information, the comparison module is specifically configured to: perform standardization processing on the first fabric content, and splice the second fabric content, where the first fabric content is the content corresponding to the fabric information in the first text content, and the second fabric is the content corresponding to the fabric information in the second text content; determine whether the spliced second fabric content is included in the processed first fabric content, where the fact that the spliced second fabric content is not included in the processed first fabric content indicates that the product information is incorrect. In the above solution, the multiple strings obtained by splicing the second fabric content are respectively subjected to an inclusion test with the string obtained by standardizing the first fabric content. If any one of the multiple strings is included in the string obtained by standardizing the first fabric content, it is considered that the fabric test passes; if all the multiple strings can pass the test in this way, it is considered that the fabric information input by the user is correct, otherwise it is considered that the fabric information input by the user is incorrect.
[0014] In an alternative embodiment, if the product information includes goods information, the comparison module is specifically configured to: convert the first goods content and the second goods content into lowercase letters, where the first goods content is the content corresponding to the goods information in the first text content, and the second goods content is the content corresponding to the goods information in the second text content; determine whether the converted first goods content is the same as the converted second goods content, where the fact that the converted first goods content is not the same as the converted second goods content indicates that the product information is incorrect. In the above solution, the first goods content is uniformly converted into lowercase letters, the second goods content is uniformly converted into lowercase letters, and then the strings are compared to see if they are the same. If they are the same, it is considered that the goods information input by the user is correct, otherwise it is considered that the goods information input by the user is incorrect.
[0015] In an alternative embodiment, if the product information includes price information, the comparison module is specifically configured to: determine whether the first price content is equal to the second price content, where the first price content is the content corresponding to the price information in the first text content, and the second price content is the content corresponding to the price information in the second text content, and the fact that the first price content is not equal to the second price content indicates that the product information is incorrect. In the above solution, the first price content and the second price content are compared numerically. If they are equal, it is considered that the price information input by the user is correct, otherwise it is considered that the price information input by the user is incorrect.
[0016] In an alternative embodiment, the determining module is specifically configured to: extract the text in the product certificate picture by using optical character recognition to obtain corresponding third text content; input the third text content into a pre-trained text inference model to obtain the second text content output by the text inference model. In the above solution, by integrating optical character recognition technology and a text inference model, the second text content is extracted from the product certificate picture. Among them, optical character recognition technology can make the text content extracted from the product certificate picture more accurate; through the structured extraction ability of the text inference model, the text content obtained from the product certificate picture is more structured and easier to be parsed and used by computer programs.
[0017] In an alternative embodiment, the product information recognition device further includes: a generating module, configured to generate a modification suggestion for the product information according to the comparison result. In the above solution, after determining whether the product information is correct, a modification suggestion for the product information can be further generated according to the comparison result, so that the user can quickly modify the product information based on the above modification suggestion.
[0018] In a third aspect, an embodiment of the present application provides a computer program product, including computer program instructions, which, when read and run by a processor, execute the product information recognition method as described in the first aspect.
[0019] In a fourth aspect, an embodiment of the present application provides an electronic device, including: a processor, a memory, and a bus; the processor and the memory communicate with each other through the bus; the memory stores computer program instructions executable by the processor, and the processor can execute the product information recognition method as described in the first aspect by invoking the computer program instructions.
[0020] In a fifth aspect, an embodiment of the present application provides a computer-readable storage medium, which stores computer program instructions, and when the computer program instructions are run by a computer, the computer is enabled to execute the product information recognition method as described in the first aspect.
[0021] To make the above objects, features, and advantages of the present application more obvious and understandable, specific embodiments of the present application are hereinafter cited and described in detail in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS
[0022] To more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the accompanying drawings required for use in the embodiments of the present application. It should be understood that the following drawings only show certain embodiments of the present application and should not be regarded as limiting the scope. For those of ordinary skill in the art, without creative efforts, other related drawings can also be obtained based on these drawings.
[0023] Figure 1 Flowchart of a method for identifying commodity information provided by an embodiment of the present application;
[0024] Figure 2 Schematic diagram of a commodity certificate picture provided by an embodiment of the present application;
[0025] Figure 3 Structural block diagram of a commodity information identification device provided by an embodiment of the present application;
[0026] Figure 4 Structural block diagram of an electronic device provided by an embodiment of the present application. Detailed implementation manners
[0027] The following will describe the technical solutions in the embodiments of the present application in conjunction with the accompanying drawings in the embodiments of the present application.
[0028] Please refer to Figure 1 , Figure 1 which is a flowchart of a method for identifying commodity information provided by an embodiment of the present application. The method for identifying commodity information may specifically include the following steps:
[0029] Step S101: Obtain the first text content and the commodity certificate picture, where the first text content includes the commodity information input by the user.
[0030] Step S102: Determine the corresponding second text content according to the commodity certificate picture, where the second text content is structured text.
[0031] Step S103: Compare the first text content with the second text content to determine whether the commodity information is correct.
[0032] Specifically, in the above step S101, the first text content includes the commodity information input by the user. It should be noted that the present application does not specifically limit the specific implementation manners of the above commodity information, and those skilled in the art can make appropriate adjustments according to the actual situation. For example, the above commodity information may include fabric information, item number information, brand information, price information, etc.
[0033] In addition, the embodiments of the present application do not specifically limit the specific implementation manners of obtaining the first text content and the picture of the product certificate, and those skilled in the art can make appropriate adjustments according to the actual situation. For example, the first text content and the picture of the product certificate sent by an external device can be received; or, the first text content and the picture of the product certificate stored locally or in the cloud can be read; or, the first text content and the picture of the product certificate input by the user in real time can be read, etc.
[0034] As an implementation manner, the user can fill in the fabric information, item number information, brand information, and price information corresponding to the product on the interface. These information are all in text type, and the first text content after input is structured text, for example: structured text in JSON format. For example, the product information input by the user can be:
[0035] Fabric: 83.0% cotton, 14.0% nylon, 3.0% spandex
[0036] Item number: sr3027
[0037] Price: 99
[0038] Brand: SEMR Semir.
[0039] The first text content after submission is:
[0040] {
[0041] "Fabric": "83.0% cotton, 14.0% nylon, 3.0% spandex",
[0042] "Item number": "sr3027",
[0043] "Brand": "SEMR Semir",
[0044] "Price": 99
[0045] }
[0046] As another implementation manner, the user can upload a photo of the product certificate on the interface. Please refer to Figure 2 , Figure 2 which is a schematic diagram of a product certificate picture provided by the embodiments of the present application.
[0047] In the above step S102, the corresponding second text content can be determined according to the picture of the product certificate, and the above second text content is structured text. It should be noted that the specific implementation manner of determining the corresponding second text content according to the picture of the product certificate in the embodiments of the present application is not specifically limited, and those skilled in the art can make appropriate adjustments according to the actual situation. For example, the second text content can be extracted from the picture of the product certificate by using image recognition technology; or, the second text content can be extracted from the picture of the product certificate by using deep learning, etc.
[0048] For example, the second text content can be:
[0049] {
[0050] "fabric":
[0051] {"cotton":83},
[0052] {"nylon":14},
[0053] {"spandex":3}
[0054] ,
[0055] "article number":"SR3027",
[0056] "brand":"SEMIR Semir",
[0057] "price":99
[0058] }。
[0059] In the above step S103, the first text content can be compared with the second text content to determine whether the product information is correct. It can be understood that in the embodiments of the present application, all the content of the first text content and the second text content can be compared, or part of the content of the first text content and the second text content can be compared. The embodiments of the present application do not make specific limitations on this.
[0060] As an implementation manner, if part of the first text content is consistent with the second text content, it can be considered that the product information is correct; as another implementation manner, if all of the first text content is consistent with the second text content, it can be considered that the product information is correct; as still another implementation manner, if specific content in the first text content is consistent with the second text content, it is considered that the product information is correct.
[0061] It should be noted that the embodiments of the present application do not specifically limit the specific implementation manners of comparing the first text content with the second text content, and those skilled in the art can make appropriate adjustments according to the actual situation. For example, the first text content and the second text content can be directly compared textually to determine whether they are consistent; or, it can be determined whether the second text content is included in the first text content, etc.
[0062] In the above solution, the content of the product certificate picture is structurally extracted to obtain the second text content of the structured text; by comparing the first text content corresponding to the product information input by the user with the above second text content, it is determined whether the product information is correct. Among them, since the second text content of the structured text is extracted based on the product certificate picture, it is convenient to perform content comparison. At the same time, since a computer program is used for content comparison, compared with the manual review method in the prior art, the review efficiency of the product information can be improved.
[0063] Further, on the basis of the above embodiments, if the product information includes fabric information, the above step S103 may specifically include the following steps:
[0064] Step 1), standardize the first fabric content and splice the second fabric content, where the first fabric content is the content corresponding to the fabric information in the first text content, and the second fabric is the content corresponding to the fabric information in the second text content.
[0065] Step 2), determine whether the spliced second fabric content is included in the processed first fabric content, where the fact that the spliced second fabric content is not included in the processed first fabric content indicates that the product information is incorrect.
[0066] Specifically, in the above step 1), the first fabric content refers to the content corresponding to the fabric information in the first text content, and the above first fabric content can be standardized. It should be noted that the embodiments of the present application do not specifically limit the specific implementation manners of the above standardization process, and those skilled in the art can make appropriate adjustments according to the actual situation. For example, the spaces in the first fabric content can be removed; or, the 0 in the first fabric content can be removed, etc.
[0067] As an implementation manner, the first fabric content is the fabric string input by the user, and the trailing 0 after the decimal point before the percentage sign in the fabric string can be removed to achieve standardization. For example: before standardization: 83.0% cotton 14.0% nylon 3.0% spandex; after standardization: 83% cotton 14% nylon 3% spandex.
[0068] The second fabric content refers to the content corresponding to the fabric information in the second text content, and the above-mentioned second fabric content can be spliced. As an implementation method, each fabric and content in the second fabric content can be spliced into a string containing a percentage, and each fabric can be spliced into two forms: "fabric content percentage" and "content percentage fabric". For example, {"cotton": 83} can be spliced into two strings: "cotton 83%" and "83% cotton".
[0069] As another implementation method, for the second fabric content, the trailing 0 after the decimal point before the percentage sign can also be removed.
[0070] In the above step 2), it can be determined whether the spliced second fabric content is included in the processed first fabric content. Among them, the fact that the spliced second fabric content is not included in the processed first fabric content indicates that the commodity information is incorrect.
[0071] As an implementation method, two strings generated from each fabric extracted according to the second fabric content can be respectively subjected to an inclusion test with the standardized string of the first fabric content. If any one of the two strings is included in the standardized string of the first fabric content, it is considered that the fabric passes the test; if each fabric and content extracted according to the second fabric content can pass the test in this way, it is considered that the fabric information input by the user is correct; otherwise, it is considered incorrect.
[0072] For example, it can be determined that "83% cotton" in "cotton 83%" or "83% cotton" is included in "83% cotton 14% nylon 3% spandex", so it can be considered that the fabric passes the test. And so on, the three fabrics of "cotton", "nylon", and "spandex" can all pass the test in this way, so the fabric information input by the user is correct.
[0073] In the above solution, multiple strings obtained by splicing the second fabric content are respectively subjected to an inclusion test with the standardized string of the first fabric content. If any one of the multiple strings is included in the standardized string of the first fabric content, it is considered that the fabric passes the test; if multiple strings can all pass the test in this way, it is considered that the fabric information input by the user is correct, otherwise it is considered that the fabric information input by the user is incorrect.
[0074] Furthermore, on the basis of the above embodiments, if the commodity information includes goods information, the above step S103 may specifically include the following steps:
[0075] Step 1), convert the first product content and the second product content into lowercase letters, where the first product content is the content corresponding to the product information in the first text content, and the second product content is the content corresponding to the product information in the second text content.
[0076] Step 2), determine whether the converted first product content is consistent with the converted second product content. If the converted first product content is inconsistent with the converted second product content, it indicates that the product information is incorrect.
[0077] Specifically, in the above Step 1), the first product content refers to the content corresponding to the product information in the first text content, and the English in the first product content can be converted into lowercase letters; the second product content refers to the content corresponding to the product information in the second text content, and the English in the second product content can be converted into lowercase letters. As an implementation manner, the above product information may include item number information and / or brand information.
[0078] In the above Step 2), it can be determined whether the converted first product content is consistent with the converted second product content. If the converted first product content is inconsistent with the converted second product content, it indicates that the product information is incorrect.
[0079] For example, after the item numbers in the first product content are unified into lowercase, it is "sr3027"; after the item numbers in the second product content are unified into lowercase, it is "sr3027", and the two contents are exactly the same. Therefore, the item number information input by the user is correct. Another example, after the brand information in the first product content is unified into lowercase, it is "semr semir"; after the brand information in the second product content is unified into lowercase, it is "semir semir", and the two contents are inconsistent. Therefore, the brand information input by the user is incorrect.
[0080] In the above solution, the first product content is uniformly converted into lowercase letters, the second product content is uniformly converted into lowercase letters, and then the strings are compared for consistency. If they are consistent, it is considered that the product information input by the user is correct; otherwise, it is considered that the product information input by the user is incorrect.
[0081] Further, on the basis of the above embodiment, if the product information includes price information, the above Step S103 may specifically include the following steps:
[0082] Determine whether the first price content is equal to the second price content, where the first price content is the content corresponding to the price information in the first text content, and the second price content is the content corresponding to the price information in the second text content. If the first price content is not equal to the second price content, it indicates that the product information is incorrect.
[0083] Specifically, it can be determined whether the numbers in the first price content and the second price content are equal. If they are equal, it is considered that the price information entered by the user is correct; otherwise, it is considered incorrect. For example, if the first price content is 99 and the second price content is 99, and the two numbers are equal, the price information entered by the user is correct.
[0084] In the above solution, the first price content and the second price content are compared numerically. If they are equal, it is considered that the price information entered by the user is correct; otherwise, it is considered that the price information entered by the user is incorrect.
[0085] Further, on the basis of the above embodiments, step S102 may specifically include the following steps:
[0086] Step 1), use optical character recognition to extract the text in the product certificate picture to obtain the corresponding third text content.
[0087] Step 2), input the third text content into a pre-trained text inference model to obtain the second text content output by the text inference model.
[0088] Specifically, in the above step 1), optical character recognition (OCR) can be used to extract the text in the product certificate picture to obtain the corresponding third text content. Taking the Figure 2 shown product certificate picture as an example, the third text content may be:
[0089] SR3027-1P
[0090] 22-24
[0091] Brand Name
[0092] Xxx
[0093] Product Name
[0094] Product Grade
[0095] Socks
[0096] Qualified Product
[0097] cm
[0098] Goods
[0099] No.
[0100] Xxx
[0101] Plantar Composition
[0102] 83.0% Cotton
[0103] 14.0% Nylon
[0104] 3.0% Spandex
[0105] (except jacquard thread)
[0106] Specification
[0107] Size
[0108] 24 - 26cm
[0109] Inspection
[0110] Verification
[0111] QC08 (Internal Code)
[0112] Execution Standard
[0113] FZ / T 73001 - 2016
[0114] Safety Class
[0115] GB18401 Class B
[0116] Manufacturer
[0117] Xxx
[0118] Address
[0119] Location
[0120] Xxx
[0121] Tel
[0122] Phone
[0123] Xxx
[0124] Production
[0125] Site
[0126] Washing Instructions
[0127] Certificate of Conformity
[0128] ￥99.00
[0129] In the above step 2), the third text content can be input into a pre - trained text reasoning model to obtain the second text content output by the text reasoning model, where the text reasoning model can be trained first.
[0130] As an implementation, the above text reasoning model can be implemented using Artificial Intelligence Generated Content (AIGC). For example, the following content can be used to train AIGC:
[0131] You are an expert in analyzing the text of clothing product certificates. I will provide you with an irregular text, and you need to extract the fabric, product number, brand name, and price contained in it.
[0132] Constraints:
[0133] 1. Fabric extraction rule: Separate the components and content in the fabric, and only extract the numbers for the content without %.
[0134] 2. Only extract the numbers for the price.
[0135] 3. The final result needs to be output in JSON format.
[0136] Output example:
[0137] {"Fabric": [{"Cotton": "90"}, {"Nylon": 14}], "Product number": "SR3027", "Brand": "SEMIR Semir", "Price": 99}.
[0138] In the above solution, by integrating optical character recognition technology and a text reasoning model, the second text content is extracted from the product certificate picture. Among them, optical character recognition technology can make the text content extracted from the product certificate picture more accurate; through the structured extraction ability of the text reasoning model, the text content obtained from the product certificate picture is more structured and easier to be parsed and used by computer programs.
[0139] Further, on the basis of the above embodiments, after the above step S103, the commodity information recognition method provided by the embodiments of the present application may specifically include the following steps:
[0140] Generate a modification suggestion for the commodity information according to the comparison result.
[0141] Specifically, according to the comparison result obtained in step S103, find the incorrect content items input by the user, and corresponding modification suggestions can be generated. For example, if it is found that the brand information input by the user is incorrect, the modification suggestion can be generated: [Brand] is inconsistent with the certificate recognition result, and it is recommended to be [SEMIR Semir].
[0142] In the above solution, after determining whether the commodity information is correct, a modification suggestion for the commodity information can be further generated according to the comparison result, so that the user can quickly modify the commodity information based on the above modification suggestions.
[0143] Further, on the basis of the above embodiments, the embodiments of the present application may include: a commodity information entry module, a picture OCR recognition module, an AIGC text reasoning module, an information comparison module, and a modification suggestion generation module. The functions of each module are introduced below:
[0144] Commodity Information Entry Module: This module is mainly responsible for the input of commodity information. The commodity information includes fabric information, item number information, brand information, price information in text type, and the commodity certificate picture in picture type, and provides the function of submitting for review.
[0145] Picture OCR Recognition Module: This module extracts the text information on the commodity certificate picture through OCR technology according to the commodity certificate picture input by the user. Input the commodity certificate picture and output the recognized text content.
[0146] AIGC Text Reasoning Module: This module further processes the text information output by the picture OCR recognition module. By training AIGC, it corrects the text content with OCR recognition errors according to its own role and the input text content, combined with the knowledge of the large model, and structurally arranges the corrected text content. Finally, it outputs the structured text in JSON format that can be recognized by computer programs, where fabric, item number, brand, and price are the respective Keys of the JSON text, and the Values are the recognized fabric information, item number information, brand information, and price information.
[0147] Information Comparison Module: This module compares the fabric information, item number information, brand information, and price information of the commodity input by the user with the fabric, item number, brand, and price inferred by AIGC according to certain rules, and records the inconsistent content items as the output of this module.
[0148] Modification Suggestion Generation Module: This module takes the inconsistent content items output by the information comparison module and the structured text output by the AIGC text reasoning module as inputs, and generates modification suggestions for each inconsistent content. The modification suggestions include the correct fabric information, item number information, brand information, and price information output by the AIGC text reasoning module.
[0149] The commodity recognition method provided by the embodiment of this application has the following specific advantages compared with the existing review methods:
[0150] 1. The extraction of the content of the commodity certificate picture is more accurate and easier to use: By integrating the OCR picture recognition program and the AIGC text content reasoning program, the text content obtained from the commodity certificate picture is more accurate; through the AIGC text content structured extraction ability, the text content obtained from the commodity certificate picture is more structured and easier to be parsed and used by computer programs.
[0151] 2. More intelligent information comparison: In fabric comparison, the trailing 0 after the decimal point is removed; in the comparison of item numbers and brand names, the English letters are unified into lowercase. This method is more friendly for judging the information input by users and the information generated by AIGC, with stronger fault tolerance, so it appears more flexible and intelligent.
[0152] 3. More accurate audit results and modification suggestions: By integrating OCR and AIGC, the information on the product certificate obtained by this audit technology is more accurate, which further improves the accuracy of the audit results and the accuracy of the information extracted from the product certificate used in the modification suggestions. Moreover, this computer program audit will not have the "work fatigue" phenomenon of manual audit, and the accuracy rate will not decrease with the prolonging of the audit time. In addition, due to the characteristics of AIGC itself, continuously inputting more product certificate text data to AIGC, the text information extracted by AIGC will be more accurate, that is, the longer the audit time, the higher the accuracy rate.
[0153] 4. Faster audit speed: Automating all three steps in the audit rejection process is more efficient and accurate than the traditional audit method.
[0154] Please refer to Figure 3 , Figure 3 FIG.
[0155] In the above solution, the content of the product certificate picture is structured and extracted to obtain the second text content of the structured text; by comparing the first text content corresponding to the product information input by the user with the above second text content, it is determined whether the product information is correct. Among them, since the second text content of the structured text is extracted based on the product certificate picture, it is convenient for content comparison. At the same time, since a computer program is used for content comparison, compared with the manual audit method in the prior art, the audit efficiency of product information can be improved.
[0156] Further, based on the above embodiments, if the product information includes fabric information, the comparison module 303 is specifically configured to: perform standardization processing on the first fabric content, and splice the second fabric content, where the first fabric content is the content corresponding to the fabric information in the first text content, and the second fabric is the content corresponding to the fabric information in the second text content; determine whether the spliced second fabric content is included in the processed first fabric content, where the fact that the spliced second fabric content is not included in the processed first fabric content indicates that the product information is incorrect.
[0157] In the above solution, the multiple strings obtained by splicing the second fabric content are respectively subjected to inclusion detection with the string obtained by standardizing the first fabric content. If any one of the multiple strings is included in the string obtained by standardizing the first fabric content, it is considered that the fabric passes the detection; if all the multiple strings can pass the detection in this way, it is considered that the fabric information input by the user is correct, otherwise it is considered that the fabric information input by the user is incorrect.
[0158] Further, based on the above embodiments, if the product information includes goods information, the comparison module 303 is specifically configured to: convert the first goods content and the second goods content into lowercase letters, where the first goods content is the content corresponding to the goods information in the first text content, and the second goods content is the content corresponding to the goods information in the second text content; determine whether the converted first goods content is consistent with the converted second goods content, where the fact that the converted first goods content is not consistent with the converted second goods content indicates that the product information is incorrect.
[0159] In the above solution, the first goods content is uniformly converted into lowercase letters, and the second goods content is uniformly converted into lowercase letters, and then the strings are compared for consistency. If they are consistent, it is considered that the goods information input by the user is correct, otherwise it is considered that the goods information input by the user is incorrect.
[0160] Further, based on the above embodiments, if the product information includes price information, the comparison module 303 is specifically configured to: determine whether the first price content is equal to the second price content, where the first price content is the content corresponding to the price information in the first text content, and the second price content is the content corresponding to the price information in the second text content, and the fact that the first price content is not equal to the second price content indicates that the product information is incorrect.
[0161] In the above solution, a numerical comparison is made between the first price content and the second price content. If they are equal, it is considered that the price information input by the user is correct; otherwise, it is considered that the price information input by the user is incorrect.
[0162] Further, based on the above embodiment, the determining module 302 is specifically configured to: use optical character recognition to extract the text in the product certificate picture to obtain the corresponding third text content; input the third text content into a pre-trained text inference model to obtain the second text content output by the text inference model.
[0163] In the above solution, by integrating optical character recognition technology and a text inference model, the second text content is extracted from the product certificate picture. Among them, optical character recognition technology can make the text content extracted from the product certificate picture more accurate; through the structured extraction ability of the text inference model, the text content obtained from the product certificate picture is more structured and easier to be parsed and used by computer programs.
[0164] Further, based on the above embodiment, the product information recognition device 300 further includes: a generating module, configured to generate a modification suggestion for the product information according to the comparison result.
[0165] In the above solution, after determining whether the product information is correct, a modification suggestion for the product information can be further generated according to the comparison result, so that the user can quickly modify the product information based on the above modification suggestion.
[0166] Please refer to Figure 4 , Figure 4 , which is a structural block diagram of an electronic device provided by an embodiment of the present application. The electronic device 400 includes: at least one processor 401, at least one communication interface 402, at least one memory 403, and at least one communication bus 404. Among them, the communication bus 404 is used to realize the direct connection communication between these components, the communication interface 402 is used to communicate with other node devices by signaling or data, and the memory 403 stores machine-readable instructions executable by the processor 401. When the electronic device 400 runs, the processor 401 communicates with the memory 403 through the communication bus 404, and when the machine-readable instructions are called by the processor 401, the above product information recognition method is executed.
[0167] For example, the processor 401 in the embodiment of the present application can read a computer program from the memory 403 through the communication bus 404 and execute the computer program to implement the following method: obtaining first text content and a picture of a product certificate, where the first text content includes product information input by a user; determining corresponding second text content according to the picture of the product certificate, where the second text content is structured text; comparing the first text content with the second text content to determine whether the product information is correct.
[0168] Among them, the processor 401 includes one or more, which can be an integrated circuit chip with signal processing capabilities. The above-mentioned processor 401 can be a general-purpose processor, including a central processing unit (Central Processing Unit, abbreviated as CPU), a microcontroller unit (Micro Controller Unit, abbreviated as MCU), a network processor (Network Processor, abbreviated as NP), or other conventional processors; it can also be a dedicated processor, including a neural network processor (Neural-network Processing Unit, abbreviated as NPU), a graphics processing unit (Graphics Processing Unit, abbreviated as GPU), a digital signal processor (Digital Signal Processor, abbreviated as DSP), an application specific integrated circuit (Application Specific Integrated Circuits, abbreviated as ASIC), a field programmable gate array (Field Programmable Gate Array, abbreviated as FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components. And when there are multiple processor 401s, a part of them can be general-purpose processors and another part can be dedicated processors.
[0169] The memory 403 includes one or more, which can be, but is not limited to, a random access memory (Random Access Memory, abbreviated as RAM), a read only memory (Read Only Memory, abbreviated as ROM), a programmable read-only memory (Programmable Read-Only Memory, abbreviated as PROM), an erasable programmable read-only memory (Erasable Programmable Read-Only Memory, abbreviated as EPROM), an electrically erasable programmable read-only memory (Electric Erasable Programmable Read-Only Memory, abbreviated as EEPROM), etc.
[0170] It can be understood that Figure 4 the structure shown is only schematic, and the electronic device 400 may also include more or fewer components than those shown in Figure 4 or have a configuration different from that shown in Figure 4 . Figure 4 Each component shown in can be implemented by hardware, software, or a combination thereof. In the embodiments of the present application, the electronic device 400 may be, but is not limited to, physical devices such as desktop computers, laptop computers, smartphones, smart wearable devices, in-vehicle devices, etc., and may also be virtual devices such as virtual machines. Additionally, the electronic device 400 is not necessarily a single device, and may also be a combination of multiple devices, such as a server cluster, etc.
[0171] The embodiments of the present application also provide a computer program product, including a computer program stored on a computer-readable storage medium. The computer program includes computer program instructions. When the computer program instructions are executed by a computer, the computer can execute the steps of the commodity information recognition method in the above embodiments, for example, including: Step S101: Obtain the first text content and the picture of the commodity certificate, where the first text content includes the commodity information input by the user. Step S102: Determine the corresponding second text content according to the picture of the commodity certificate, where the second text content is structured text. Step S103: Compare the first text content with the second text content to determine whether the commodity information is correct.
[0172] The embodiments of the present application also provide a computer-readable storage medium. The computer-readable storage medium stores computer program instructions. When the computer program instructions are run by a computer, the computer executes the commodity information recognition method described in the foregoing method embodiments.
[0173] In the embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of the units is only a logical function division, and there may be other division methods in actual implementation. Also, for example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point, the displayed or discussed coupling or direct coupling or communication connection between each other can be through some communication interfaces. The indirect coupling or communication connection of the devices or units can be electrical, mechanical, or other forms.
[0174] In addition, the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they can be located in one place, or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.
[0175] Furthermore, in each embodiment of the present application, each functional module can be integrated together to form an independent part, or each module can exist alone, or two or more modules can be integrated to form an independent part.
[0176] It should be noted that if a function is implemented in the form of a software functional module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in each embodiment of the present application. The aforementioned storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs that can store program codes.
[0177] In this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations.
[0178] The above are only the embodiments of the present application and are not used to limit the protection scope of the present application. For those skilled in the art, the present application can have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.
Claims
1. A commodity information identification method, characterized in that: include: Obtaining first text content and a product certificate image, wherein the first text content includes product information input by a user; Determine the corresponding second text content according to the product certificate image, wherein the second text content is structured text; The first text content is compared with the second text content to determine whether the product information is correct.
2. The commodity information identification method according to claim 1, characterized in that: If the product information includes fabric information, comparing the first text content with the second text content includes: Standardizing the first fabric content, and splicing the second fabric content, wherein the first fabric content is the content corresponding to the fabric information in the first text content, and the second fabric is the content corresponding to the fabric information in the second text content; It is determined whether the spliced second fabric content is included in the processed first fabric content, wherein the fact that the spliced second fabric content is not included in the processed first fabric content indicates that the product information is incorrect.
3. The commodity information identification method according to claim 1, characterized in that: If the commodity information includes product information, the comparing the first text content with the second text content includes: Convert the first product content and the second product content into lowercase letters, wherein the first product content is the content in the first text content corresponding to the product information, and the second product content is the content in the second text content corresponding to the product information; It is determined whether the converted first product content is consistent with the converted second product content, wherein the inconsistency between the converted first product content and the converted second product content indicates that the product information is incorrect.
4. The commodity information identification method according to claim 1, characterized in that: If the product information includes price information, comparing the first text content with the second text content includes: Determine whether the first price content is equal to the second price content, wherein the first price content is the content in the first text content corresponding to the price information, and the second price content is the content in the second text content corresponding to the price information. If the first price content is not equal to the second price content, it indicates that the product information is incorrect.
5. The commodity information identification method according to any one of claims 1 to 4, characterized in that: The determining of the corresponding second text content according to the product certificate image includes: Extracting the text in the product certificate image using optical character recognition to obtain corresponding third text content; The third text content is input into a pre-trained text reasoning model to obtain the second text content output by the text reasoning model.
6. The commodity information identification method according to any one of claims 1 to 4, characterized in that: After comparing the first text content with the second text content to determine whether the product information is correct, the method further includes: Generate modification suggestions for the product information based on the comparison results.
7. A commodity information identification device, characterized in that: include: An acquisition module, used to acquire first text content and a product certificate image, wherein the first text content includes product information input by a user; A determination module, configured to determine the corresponding second text content according to the product certificate image, wherein the second text content is a structured text; The comparison module is used to compare the first text content with the second text content to determine whether the product information is correct.
8. A computer program product, characterized in that The method comprises computer program instructions, and when the computer program instructions are read and executed by a processor, the method for identifying commodity information according to any one of claims 1 to 6 is executed.
9. An electronic device, characterized in that: include: processor, memory, and bus; The processor and the memory communicate with each other via the bus; The memory stores computer program instructions that can be executed by the processor, and the processor calls the computer program instructions to execute the commodity information identification method according to any one of claims 1 to 6.
10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer program instructions, and when the computer program instructions are executed by a computer, the computer is caused to execute the commodity information identification method according to any one of claims 1 to 6.