Article name recognition method, computer device, readable storage medium and program product

Abstract

The application relates to an article name recognition method, a computer device, a computer readable storage medium and a computer program product, which are applied to the technical field of big data, and the method comprises the following steps: acquiring an article name of an article to be recognized; extracting name description information of the article name, wherein the name description information is used for representing a name structure of the article name; classifying the article name according to the name description information of the article name, so as to obtain a name category of the article name; extracting name keywords corresponding to the article name according to a keyword extraction mode corresponding to the name category; and recognizing an article category of the article to be recognized according to the name keywords, so as to obtain an article category recognition result, wherein the article category recognition result is used for representing the article category of the article to be recognized under a preset classification granularity. The method can improve the article name recognition efficiency.

Description

Technical Field

[0001] This application relates to the field of big data technology, and in particular to a method for identifying item names, computer equipment, computer-readable storage media, and computer program products. Background Technology

[0002] With the development of technology and the increasing maturity of big data technology, people's demand for goods transportation is also growing, and the types of goods being transported are becoming more diverse. To ensure transportation quality, it is necessary to determine the category of the item based on its name, thereby determining the corresponding shipping standards and accepting the item according to those standards. Currently, related technologies often directly determine the item category based on the item name provided by the sender. However, item names often contain a lot of misleading information and cannot accurately reflect the item category. Therefore, it is necessary to identify the item name to ensure it accurately reflects the item category. Summary of the Invention

[0003] Therefore, it is necessary to provide a method, apparatus, computer equipment, computer-readable storage medium, and computer program product that can accurately reflect the category of an item in the item name recognition result, in order to address the above-mentioned technical problems.

[0004] Firstly, this application provides a method for identifying the name of an item, including:

[0005] Obtain the name of the item to be identified;

[0006] Extract the name description information of the item name, which is used to characterize the name structure of the item name;

[0007] Based on the name description information of the item name, the item name is classified to obtain the name category of the item name;

[0008] Extract the name keywords corresponding to the item name according to the keyword extraction method corresponding to the name category;

[0009] Based on the name keywords, the item category of the item to be identified is identified to obtain the item category identification result. The item category identification result is used to characterize the item category of the item to be identified at a preset classification granularity.

[0010] In one embodiment, the name description information includes a name description structure and a field length; classifying the item name according to the name description information to obtain the item name category includes: if the name description structure of the item name is a pure character structure, then the item name category is determined as a first category; if the name description structure of the item name is not a pure character structure, and the field length of the item name is greater than a preset length threshold, then the item name category is determined as a second category; if the name description structure of the item name is not a pure character structure, and the field length of the item name is not greater than a preset length threshold, then the item name category is determined as a third category.

[0011] In one embodiment, the step of extracting the name keywords corresponding to the item name according to the keyword extraction method corresponding to the name category includes: if the name category is the first category, then searching in a preset knowledge base according to the item name to obtain search results; and extracting the name keywords corresponding to the item name from the search results according to the semantic information of the search results.

[0012] In one embodiment, extracting the name keywords corresponding to the item name from the search results based on the semantic information of the search results includes: filtering category keywords used to characterize the item category from the search results based on the semantic information of the search results; and removing effect keywords used to describe the use effect of the item from the category keywords based on the semantic information of the search results and the position of the category keywords in the search results, thereby obtaining the name keywords corresponding to the item name.

[0013] In one embodiment, the step of extracting the name keywords corresponding to the item name according to the keyword extraction method corresponding to the name category further includes: if the name category is the second category, then extracting the name keywords corresponding to the item name from the item name according to the semantic information of the item name; if the name category is the third category, then extracting the name keywords corresponding to the item name from the item name.

[0014] In one embodiment, the step of identifying the item category of the item to be identified based on the name keywords to obtain an item category identification result includes: acquiring an item image corresponding to the item to be identified; identifying the item packaging information corresponding to the item to be identified from the item image; extracting packaging features from the item packaging information; extracting name features from the name keywords; generating target features based on the packaging features and the name features; and identifying the item category of the item to be identified based on the target features to obtain an item category identification result.

[0015] In one embodiment, generating target features based on the packaging features and the name features includes: weighting the packaging features according to a preset first weight to obtain packaging-weighted features; weighting the name features according to a preset second weight to obtain name-weighted features; and fusing the packaging-weighted features and the name-weighted features to obtain target features.

[0016] Secondly, this application also provides an item name recognition device, comprising:

[0017] The acquisition module is used to obtain the name of the item to be identified;

[0018] The first extraction module is used to extract the name description information of the item name, wherein the name description information is used to characterize the name structure of the item name;

[0019] The classification module is used to classify the item name according to the name description information of the item name to obtain the name category of the item name;

[0020] The second extraction module is used to extract the name keywords corresponding to the item name according to the keyword extraction method corresponding to the name category;

[0021] The identification module is used to identify the item category of the item to be identified based on the name keywords, and obtain the item category identification result. The item category identification result is used to characterize the item category of the item to be identified at a preset classification granularity.

[0022] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0023] Obtain the name of the item to be identified;

[0024] Extract the name description information of the item name, which is used to characterize the name structure of the item name;

[0025] Based on the name description information of the item name, the item name is classified to obtain the name category of the item name;

[0026] Extract the name keywords corresponding to the item name according to the keyword extraction method corresponding to the name category;

[0027] Based on the name keywords, the item category of the item to be identified is identified to obtain the item category identification result. The item category identification result is used to characterize the item category of the item to be identified at a preset classification granularity.

[0028] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the following steps:

[0029] Obtain the name of the item to be identified;

[0030] Extract the name description information of the item name, which is used to characterize the name structure of the item name;

[0031] Based on the name description information of the item name, the item name is classified to obtain the name category of the item name;

[0032] Extract the name keywords corresponding to the item name according to the keyword extraction method corresponding to the name category;

[0033] Based on the name keywords, the item category of the item to be identified is identified to obtain the item category identification result. The item category identification result is used to characterize the item category of the item to be identified at a preset classification granularity.

[0034] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, performs the following steps:

[0035] Obtain the name of the item to be identified;

[0036] Extract the name description information of the item name, which is used to characterize the name structure of the item name;

[0037] Based on the name description information of the item name, the item name is classified to obtain the name category of the item name;

[0038] Extract the name keywords corresponding to the item name according to the keyword extraction method corresponding to the name category;

[0039] Based on the name keywords, the item category of the item to be identified is identified to obtain the item category identification result. The item category identification result is used to characterize the item category of the item to be identified at a preset classification granularity.

[0040] The aforementioned item name recognition method, apparatus, computer equipment, computer-readable storage medium, and computer program product acquire the item name of the item to be identified; extract name description information of the item name, which is used to characterize the name structure of the item name; classify the item name according to the name description information to obtain the name category of the item name; extract name keywords corresponding to the item name according to the keyword extraction method corresponding to the name category; and identify the item category of the item to be identified according to the name keywords to obtain the item category recognition result, which is used to characterize the item category of the item to be identified at a preset classification granularity. The item name is classified according to the name description information, and name keywords corresponding to the item name are extracted using the keyword extraction method corresponding to the name category. Considering the diversity of item name expression forms, the item name is divided into different categories according to the name structure for personalized keyword extraction, ensuring that the extracted name keywords can characterize the true item category of the item to be identified. Therefore, the accuracy of item name recognition is improved. Attached Figure Description

[0041] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0042] Figure 1 This is a diagram illustrating the application environment of the item name recognition method in one embodiment;

[0043] Figure 2 This is a flowchart illustrating an item name recognition method in one embodiment;

[0044] Figure 3 This is a flowchart illustrating the steps of extracting name keywords corresponding to product names based on the keyword extraction method corresponding to name categories in one embodiment.

[0045] Figure 4 This is a flowchart illustrating the steps of identifying the item category of an item based on name keywords in one embodiment, and obtaining the item category identification result.

[0046] Figure 5This is a structural block diagram of an item name recognition device in one embodiment;

[0047] Figure 6 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0048] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0049] It should be noted that the information (including but not limited to item names, item images, etc.) and data (including but not limited to data used for analysis, stored data, and displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the acquisition, transmission, storage, use, and processing of the relevant data comply with the relevant provisions of national laws and regulations. Users can refuse or easily refuse content pushed to them (e.g., item name recognition results, etc.). In the embodiments of this application, certain existing solutions in the industry, such as software, components, and models, may be mentioned. These should be considered exemplary, and their purpose is merely to illustrate the feasibility of implementing the technical solution of this application, but does not mean that the applicant has already used or necessarily used such a solution.

[0050] In the express delivery and logistics sector, it is necessary to select the appropriate shipping standard based on the type of transported goods, and then accept and send items according to that standard. Transported goods are typically packaged in cardboard boxes or bags, making it difficult to identify their type from their appearance. Current technologies often rely on the semantic information of the item name provided by the sender to determine the category. However, item names frequently contain a lot of distracting information and cannot accurately reflect the true type of item, leading to incorrect acceptance and dispatch. For example, an item name may contain multiple subjects, descriptive words, and modifiers, making it easy to extract semantic information indicating the wrong item type; conversely, an item name containing only characters representing the item model may not provide semantic information about the item type. Therefore, it is necessary to identify item names to accurately reflect the item category.

[0051] The item name recognition method provided in this application can be applied to, for example... Figure 1In the application environment shown, terminal 102 communicates with server 104 via a network. A data storage system can store the data that server 104 needs to process. The data storage system can be integrated onto server 104 or placed on a cloud or other network server. Server 104 obtains the name of the item to be identified; extracts the name description information of the item name, which is used to characterize the name structure of the item name; classifies the item name according to the name description information to obtain the name category; extracts the name keywords corresponding to the item name according to the keyword extraction method corresponding to the name category; and identifies the item category of the item to be identified based on the name keywords to obtain the item category identification result, which is used to characterize the item category of the item to be identified at a preset classification granularity. Terminal 102 can be, but is not limited to, various personal computers, laptops, smartphones, tablets, IoT devices, and portable wearable devices. IoT devices can be smart speakers, smart TVs, smart air conditioners, smart in-vehicle devices, projection devices, etc. Portable wearable devices can be smartwatches, smart bracelets, head-mounted devices, etc. Headset devices can be virtual reality (VR) devices, augmented reality (AR) devices, smart glasses, etc. Server 104 can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing cloud computing services.

[0052] In one exemplary embodiment, such as Figure 2 As shown, an item name recognition method is provided, which can be applied to... Figure 1 Taking server 104 as an example, the explanation includes the following steps 202 to 206. Wherein:

[0053] Step 202: Obtain the item name of the item to be identified.

[0054] In step 202, the item to be identified is an item waiting to be identified by item name, and / or an item waiting to be transported. The item to be identified can be set by the user as needed, or it can be identified by responding to the item name identification instruction and determining the item waiting to be transported corresponding to the item name identification instruction as the item to be identified.

[0055] As one embodiment, step 202 includes: acquiring an image of the item to be identified, and extracting the name of the item from the image. The name can be obtained by performing object detection on the image, or by performing text recognition on the text in the image; it is not limited to these methods.

[0056] In another embodiment, step 202 includes: obtaining the item identifier of the item to be identified, and a preset identifier name file, wherein the preset identifier name file includes the correspondence between item identifiers and item names; querying the identifier name file based on the item identifier of the item to be identified to obtain the item name of the item to be identified. For example, the item identifier may include an item number, and the item identifier may be printed on the item packaging corresponding to the item to be identified, or it may be an identifier object placed on the item packaging corresponding to the item to be identified, such as an item identifier sticker, item identifier tag, etc.

[0057] Step 204: Extract the name description information of the item name. The name description information is used to characterize the name structure of the item name.

[0058] The name description information in step 204 includes at least one of the following: name description structure and field length.

[0059] As one embodiment, step 204 includes: extracting the field length of the product name to obtain name description information; or, extracting the name description structure of the product name to obtain name description information; or, extracting the field length and name description structure of the product name to obtain name description information.

[0060] As one embodiment, the name description structure of the extracted item name includes: extracting the description field type of the item name, wherein the description field type includes at least one of character type and text type, and the character type includes at least one of English character type and numeric character type; and generating the name description structure of the item name based on the description field type.

[0061] Furthermore, based on the description field type, a name description structure for the item name is generated, including: if the description field type includes a text type, then a non-pure character structure is determined as the name description structure for the item name; if the description field type includes both text and character types, then a non-pure character structure is determined as the name description structure for the item name; if the description field type only includes a character type, then a pure character structure is determined as the name description structure for the item name.

[0062] Step 206: Classify the item names according to the name description information to obtain the name categories of the item names.

[0063] As an example, step 206 includes: if the name description structure of the item name is a pure character structure, then the name category of the item name is determined as the first category; if the name description structure of the item name is not a pure character structure and the field length is greater than a preset length threshold, then the name category of the item name is determined as the second category; if the name description structure of the item name is not a pure character structure and the field length is not greater than the preset length threshold, then the name category of the item name is determined as the third category. The preset length threshold is set by the user as needed, or it can be an empirical value (e.g., a length that is prone to recognition errors when directly extracting keywords). The preset length threshold can be 10 characters long, or 8 characters long, or other lengths, and is not limited here.

[0064] Step 208: Extract the name keywords corresponding to the product name according to the keyword extraction method corresponding to the name category.

[0065] For example, step 208 includes: if the name category is a first category, performing an online search based on the item name to obtain search results; extracting name keywords corresponding to the item name from the search results; if the name category is a second or third category, extracting name keywords corresponding to the item name from the item name.

[0066] Step 210: Based on the name keywords, identify the item category of the item to be identified and obtain the item category identification result. The item category identification result is used to characterize the item category of the item to be identified at the preset classification granularity.

[0067] In step 210, the granularity of classification can be set by the user as needed, or it can be the same as the granularity of classification of the items in the standard shipping information. For example, the item category can be a category with a larger granularity, such as fruit or mobile phone, or a category with a relatively smaller granularity, such as banana or X brand mobile phone, or a category with a relatively smaller granularity, such as Fuji apple or X brand Y model mobile phone. There are no restrictions here. Then, the item name recognition result can include an item category label, and the item name recognition result can also include at least one of the item type label and the item brand label. The item name recognition result can also include at least one of the item type label, the item brand label, and the item model label. There are no restrictions here.

[0068] As an embodiment, step 210 includes at least one of the following: if the name keyword includes a brand keyword, then generate an item brand label based on the brand keyword; if the name keyword includes a model keyword, then generate an item model label based on the model keyword; if the name keyword includes a category keyword, then generate an item category label based on the category keyword.

[0069] In another embodiment, step 210 includes: obtaining a pre-trained item category recognition model, and mapping name keywords to item name recognition results through the item category recognition model.

[0070] Optionally, the training process of the item category recognition model includes: obtaining multiple first training samples, wherein each first training sample consists of training name keywords of the training item and the recognition result of the real item name; and iteratively training the item category recognition model to be trained using multiple first training samples to obtain the item category recognition model.

[0071] Optionally, after step 210, the method further includes: if the item name recognition result cannot correctly represent the item category, returning to step 302 and subsequent steps until the item name recognition result can correctly represent the item category.

[0072] This can be determined by whether the item name recognition result mapped by the item category recognition model is a name character (for example, if the model cannot recognize the item name when the process of extracting the target name is performed by the model, it will report an error, in which case it is determined that the extracted target name does not correctly represent the item category of the item to be identified), or by user feedback information.

[0073] As an embodiment, after step 210, the method further includes: obtaining a preset shipping standard table, wherein the shipping standard table includes at least one correspondence between shipping item category and shipping standard information, wherein the shipping item category is used to characterize the item category of the shipping item under a preset shipping classification granularity (refer to the preset classification granularity setting mentioned above), the shipping standard information includes at least one of shipping timeliness standard information and shipping access standard information, the shipping access standard information includes access shipping information or non-access shipping information, and the shipping timeliness standard information is used to characterize the shipping timeliness restriction of an item category when shipping; selecting the target shipping standard information that matches the shipping item category and item name recognition result from each shipping standard information in the shipping standard table, and generating the shipping recognition result of the item to be identified based on the target shipping standard information, wherein the shipping recognition result includes a result that meets the shipping standard or a result that does not meet the shipping standard.

[0074] Further, the process involves selecting target shipping standard information from each shipping standard information in the shipping standard table that matches the shipping item category and the item name recognition result. This includes: obtaining the similarity between the item name recognition result and each shipping item category in the shipping standard table; selecting target shipping item categories from each shipping item category in the shipping standard table whose similarity meets the preset similarity conditions; and determining the shipping standard information in the shipping standard table that corresponds to the target shipping item category as the target shipping standard information that matches the shipping item category and the item name recognition result.

[0075] The preset similarity condition can be greater than a preset similarity threshold, or it can be the maximum value among all similarities.

[0076] Based on the target shipping standard information, a shipping identification result for the item to be identified is generated, including: if the target shipping standard information includes information on prohibited shipping, then the result that does not meet the shipping standard is determined as the shipping identification result for the item to be identified; if the target shipping standard information includes information on permitted shipping, then the shipping information for the item to be identified is obtained, the shipping information including shipping location information, which includes at least one of shipping start location information, shipping end location information, and shipping distance information; according to the shipping information for the item to be identified, shipping timeliness information corresponding to the item to be identified is generated; if the shipping timeliness information meets the shipping timeliness standard information in the target shipping standard information, the result that meets the shipping standard is determined as the shipping identification result for the item to be identified; if the shipping timeliness information does not meet the shipping timeliness standard information in the target shipping standard information, the result that does not meet the shipping standard is determined as the shipping identification result for the item to be identified.

[0077] Thus, by using the above-mentioned item name recognition method, the accuracy of item name recognition can be improved, which can reduce unreasonable shipping behavior caused by inaccurate item category recognition leading to inaccurate subsequent shipping results (specifically, items that can actually be shipped being refused due to inaccurate item name recognition, or items that cannot actually be shipped being accepted due to inaccurate item name recognition), thereby improving the accuracy of item acceptance and shipping.

[0078] In the above-described item name recognition method, the following steps are taken: First, the item name to be identified is obtained. Then, the name description information of the item name is extracted, which is used to characterize the name structure of the item name. Based on the name description information, the item name is classified to obtain its name category. Next, name keywords corresponding to the name category are extracted using a keyword extraction method. Finally, the item category of the item to be identified is recognized based on the name keywords, resulting in an item category recognition result. This item category recognition result characterizes the item category of the item to be identified at a preset classification granularity. The item name is classified according to its name description information, and name keywords corresponding to the name category are extracted. Considering the diversity of item name expressions, the item name is divided into different categories based on its name structure for personalized keyword extraction. This ensures that the extracted name keywords can characterize the true item category of the item to be identified, thus improving the accuracy of item name recognition.

[0079] In one exemplary embodiment, such as Figure 3As shown, based on the keyword extraction method corresponding to the name category, the name keywords corresponding to the product name are extracted, including steps 302 to 304. Wherein:

[0080] Step 302: If the name category is the first category, then search in the preset knowledge base according to the item name to obtain the search results.

[0081] In step 302, the preset knowledge base can be a knowledge base that includes all fields, a knowledge base that the user sets up according to the fields, or a knowledge base that matches the fields of each item to be identified.

[0082] For example, a search is performed in a preset knowledge base based on the item name to obtain search results, including: using the item name as an index to search the preset knowledge base to obtain search results.

[0083] Step 304: Extract the name keywords corresponding to the item name from the search results based on the semantic information of the search results.

[0084] For example, step 304 includes: filtering category keywords from the search results to characterize the item category based on the semantic information of the search results; and determining the name keywords corresponding to the item name based on the category keywords.

[0085] Furthermore, based on the semantic information of the search results, the process of filtering category keywords to represent item categories from the search results can be implemented by a name extraction model. This name extraction model can be an LLM (Large Language Model) model or other language models; no restrictions are placed here. The classification granularity of the item category name is less than or equal to the classification granularity of the item category recognition result.

[0086] It is understandable that the item category name obtained through semantic information recognition may contain descriptive terms for the item to be identified. Directly determining the item category name as the target name may lead to misclassification of the item category. For example, if the item name is chocolate-flavored ice cream, the identified item category name during semantic recognition will include both chocolate and ice cream. Directly determining the item category name as the target name would result in both chocolate and ice cream being identified as the item category, which is clearly inconsistent with the facts.

[0087] As one embodiment, determining the name keywords corresponding to the item name based on category keywords includes: removing effect keywords that describe the use effect of the item from the category keywords based on the semantic information of the search results and the position of the category keywords in the search results, thereby obtaining the name keywords corresponding to the item name.

[0088] Furthermore, based on the semantic information of the search results and the position of the category keywords in the search results, the effect keywords used to describe the use effect of the item are removed from the category keywords to obtain the name keywords corresponding to the item name. This includes: filtering effect keywords used to describe the use effect of the item from the category keywords based on the semantic information of the search results and the position of the category keywords in the search results; removing effect keywords from the category keywords to obtain the name keywords corresponding to the item name.

[0089] As one embodiment, based on the semantic information of the search results and the position of the category keywords in the search results, effect keywords for describing the use effect of the item are filtered from the category keywords, including: based on the semantic information of the search results and the position of the category keywords in the search results, effect keywords that are adjacent to other category keywords and are ranked higher in the search results are filtered from the category keywords.

[0090] In this way, keywords that are easily identified as category keywords are removed, reducing misjudgments of the item category and improving the accuracy of item name recognition.

[0091] As one embodiment, the effect keywords in the category keywords are removed to obtain the name keywords corresponding to the item name, including: removing the effect keywords in the category keywords to obtain at least one processing keyword; if there is a single processing keyword, the processing keyword is determined as the name keyword corresponding to the item name; if there are multiple processing keywords, the name keyword corresponding to the item name is determined according to the category keywords represented by the multiple processing keywords respectively.

[0092] Furthermore, based on the item categories represented by the multiple processing keywords, the name keywords corresponding to the item name are determined, including: if the item categories represented by the multiple processing keywords are the same, then the item category represented by the multiple processing keywords is determined as the name keyword corresponding to the item name. For example, the processing keywords include Happy Mid-Autumn Festival (1 / 1), Moon Ice Cream (2 / 1), Is it you, Moon? (0 / 1), The Secret of the Moon (1 / 1), Hide and Seek with the Moon (0 / 1), and The Taste of the Moon (1 / 1). These processing keywords also represent the item category of children's books, so children's books are determined as the name keyword corresponding to the item name. If the category keywords represented by the multiple processing keywords are not the same, then the item categories represented by the multiple processing keywords are determined as the name keywords corresponding to the item name. For example, the processing keywords include canvas tote bag, canvas backpack, and wireless power bank. These processing keywords represent the item categories of canvas bag and power bank, so canvas bag and power bank are determined as the name keywords corresponding to the item name.

[0093] Thus, after removing keywords that are easily identified as item categories, the processed keywords represent multiple category keywords. Based on the item categories represented by each of the processed keywords, the processed keywords are reduced as much as possible to obtain the name keywords corresponding to the item name. This allows the name keywords corresponding to the item name to concisely and clearly represent the true item category of the item to be identified.

[0094] Optionally, the method further includes: if the name category is the second category, then extracting the name keywords corresponding to the item name from the item name based on the semantic information of the item name; if the name category is the third category, then extracting the name keywords corresponding to the item name from the item name.

[0095] Optionally, the specific implementation of extracting name keywords corresponding to the item name from the item name based on the semantic information of the item name can refer to the specific implementation of extracting name keywords corresponding to the item name from the search results based on the semantic information of the search results in step 304 above, and is not limited here.

[0096] As one embodiment, extracting the name keywords corresponding to the item name from the item name includes: determining the item name as the name keywords corresponding to the item name.

[0097] Thus, when the name category is the third category, it means that the item name itself can concisely and clearly represent the true item category of the item to be identified. Therefore, directly determining the item name as the name keyword corresponding to the item name saves unnecessary processing and improves the extraction efficiency of name keywords.

[0098] In this embodiment, the names of items belonging to the first category are searched in a preset knowledge base to obtain search results. Based on the semantic information of the search results, the name keywords corresponding to the item names are extracted from the search results. For the special case of item names belonging to the first category, that is, the name description structure of the item name is a pure character structure, the name keywords can be extracted from the search results by searching the item names and based on the semantic information of the search results. This ensures that the extracted name keywords can accurately represent the item category of the item to be identified, and improves the accuracy of name keyword extraction.

[0099] In one exemplary embodiment, such as Figure 4 As shown, based on the name keywords, the item category of the item to be identified is determined, and the item category identification result is obtained, including steps 402 to 410. Wherein:

[0100] Step 402: Obtain the image of the item to be identified, and identify the packaging information of the item to be identified from the image.

[0101] Among them, the image of the item can be obtained by image acquisition devices such as cameras and surveillance cameras, and the packaging information of the item includes at least one of the following: packaging shape information and packaging shape decoration information (e.g., packaging bag shape decoration information, packaging tape information, etc.). The packaging shape information includes at least one of the following: packaging shape and packaging size.

[0102] For example, identifying the packaging information of the item to be identified from the item image includes: the item category recognition model includes an image recognition layer, and the packaging information of the item to be identified is identified from the item image through the image recognition layer.

[0103] Step 404: Extract packaging features from the product packaging information.

[0104] The packaging features in step 404 include at least one of the packaging shape and packaging shape decoration features, and the packaging shape features include at least one of the packaging shape features and packaging size features.

[0105] For example, step 404 includes: the item category recognition model includes a feature extraction layer, through which packaging features are extracted from the item packaging information.

[0106] Step 406: Extract name features from name keywords.

[0107] For example, step 406 includes: the item category recognition model includes a feature extraction layer, through which name features are extracted from name keywords.

[0108] Step 408: Generate target features based on packaging features and name features.

[0109] As an example, step 408 includes: weighting the packaging features according to a preset first weight to obtain packaging weighted features; weighting the name features according to a preset second weight to obtain name weighted features; and fusing the packaging weighted features and the name weighted features to obtain target features.

[0110] Among them, the feature fusion methods include, but are not limited to, feature summation fusion and feature product fusion.

[0111] In this way, by using both the item packaging information and the target name as the basis for generating target features, the information channel of item packaging information is added. To a certain extent, this can improve the possibility that the target features can accurately represent the item category of the item to be identified, thus improving the accuracy of target feature extraction.

[0112] It is understandable that the following special cases may exist when packaging items, allowing packaging information to serve as one of the bases for identifying the item's category: When the item to be identified is a fan or other item with a unique shape, the packaging may be cylindrical. In this case, the shape of the packaging can be used as an auxiliary basis for identifying the item's category. When the packaging's decorative information includes the manufacturer's information, such as the manufacturer's brand logo on the packaging tape or packaging, this decorative information can be used as an auxiliary basis for identifying the item's category. When multiple keywords are extracted, and it is unclear which one correctly represents the item's category, for example, when the target names are "fan" and "air conditioner," which have significantly different sizes, the size of the packaging can be used as an auxiliary basis for identifying the item's category.

[0113] Furthermore, the method also includes: if the item packaging tape information is used to characterize the manufacturer information corresponding to the item to be identified, if the item packaging shape is an irregular shape, and / or if the target name has multiple item categories and the size difference between each item category is greater than a preset difference threshold, then the preset first weight is determined to be greater than the preset second weight.

[0114] Thus, in the case of the special circumstances of the product packaging mentioned above, the preset first weight of the packaging features extracted from the product packaging information is set relatively high, which further makes the product packaging information account for a higher proportion in the product category decision, thereby improving the accuracy of target feature extraction.

[0115] Step 410: Based on the target features, identify the item category of the item to be identified and obtain the item category identification result.

[0116] For example, step 410 includes: mapping the target features to the item name recognition result through the item category recognition model.

[0117] In this embodiment, an image of the item to be identified is obtained, and the packaging information of the item to be identified is obtained from the image. Packaging features are extracted from the packaging information. Name features are extracted from name keywords. Target features are generated based on the packaging features and name features. The item category of the item to be identified is identified based on the target features to obtain the item category identification result. By using packaging features and name features as the basis for generating target features, both the item packaging and the item name can provide a decision basis for item category identification. This process takes into account the influence of multiple factors on item category identification, thus improving the accuracy of item category identification.

[0118] As a detailed embodiment, the name of the item to be identified is obtained; the name description information of the item name is extracted, which is used to characterize the name structure of the item name; if the name description structure of the item name is a pure character structure, the name category of the item name is determined as the first category; if the name description structure of the item name is not a pure character structure, and the field length of the item name is greater than a preset length threshold, the name category of the item name is determined as the second category; if the name description structure of the item name is not a pure character structure, and the field length of the item name is not greater than the preset length threshold, the name category of the item name is determined as the third category.

[0119] Furthermore, if the name category is the first category, a search is performed in the preset knowledge base based on the item name to obtain search results; based on the semantic information of the search results, category keywords used to represent the item category are filtered from the search results; based on the semantic information of the search results and the position of the category keywords in the search results, effect keywords used to describe the use effect of the item are removed from the category keywords to obtain the name keywords corresponding to the item name; if the name category is the second category, the name keywords corresponding to the item name are extracted from the item name based on the semantic information of the item name; if the name category is the third category, the name keywords corresponding to the item name are extracted from the item name.

[0120] Further, an image of the item to be identified is obtained, and the packaging information of the item to be identified is obtained from the image. Packaging features are extracted from the packaging information. Name features are extracted from the name keywords. The packaging features are weighted according to a preset first weight to obtain packaging weighted features. The name features are weighted according to a preset second weight to obtain name weighted features. The packaging weighted features and name weighted features are fused to obtain target features. Based on the target features, the item category of the item to be identified is identified to obtain the item category identification result.

[0121] Thus, by obtaining the name of the item to be identified; extracting the name description information of the item name, which is used to characterize the name structure of the item name; classifying the item name according to the name description information to obtain the name category of the item name; extracting the name keywords corresponding to the name category according to the keyword extraction method corresponding to the name category; and identifying the item category of the item to be identified according to the name keywords to obtain the item category identification result, which is used to characterize the item category of the item to be identified at the preset classification granularity. The item name is classified according to the name description information, and the name keywords corresponding to the item name are extracted according to the keyword extraction method corresponding to the name category. Considering the diversity of item name expression forms, the item name is divided into different categories according to the name structure for personalized keyword extraction, ensuring that the extracted name keywords can characterize the true item category of the item to be identified. Therefore, the accuracy of item name identification is improved.

[0122] Furthermore, for the special case of item names belonging to the first category, i.e., those with a purely character-based descriptive structure, a search can be performed on the item name. Based on the semantic information of the search results, name keywords can be extracted from the search results, ensuring that the extracted name keywords accurately represent the item category of the item to be identified, thus improving the accuracy of name keyword extraction. For item names belonging to the second category, the implementation of extracting name keywords from search results is referenced, ensuring that the extracted name keywords accurately represent the item category of the item to be identified. For item names belonging to the third category, it is indicated that the item name itself can concisely and clearly represent the true item category of the item to be identified. Therefore, the item name can be directly identified as the corresponding name keyword, saving unnecessary processing and improving the efficiency of name keyword extraction. In addition, packaging features and name features are used as the basis for generating target features, so that both the item packaging and the item name can provide a decision basis for item category identification. This allows the item category identification process to take into account the influence of multiple factors on item category identification, thus improving the accuracy of item category identification.

[0123] It should be understood that although the steps in the flowcharts of the above embodiments are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the above embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0124] Based on the same inventive concept, this application also provides an item name recognition device for implementing the item name recognition method described above. The solution provided by this device is similar to the solution described in the above method; therefore, the specific limitations in one or more item name recognition device embodiments provided below can be found in the limitations of the item name recognition method described above, and will not be repeated here.

[0125] In one exemplary embodiment, such as Figure 5 As shown, an item name recognition device 500 is provided, including: an acquisition module 502, a first extraction module 504, a classification module 506, a second extraction module 508, and a recognition module 510, wherein:

[0126] The acquisition module 502 is used to acquire the name of the item to be identified;

[0127] The first extraction module 504 is used to extract the name description information of the item name, and the name description information is used to characterize the name structure of the item name.

[0128] The classification module 506 is used to classify item names according to the name description information of the item name to obtain the name category of the item name;

[0129] The second extraction module 508 is used to extract the name keywords corresponding to the product name according to the keyword extraction method corresponding to the name category.

[0130] The identification module 510 is used to identify the category of the item to be identified based on the name keywords, and obtain the item category identification result. The item category identification result is used to characterize the item category of the item to be identified at a preset classification granularity.

[0131] In one embodiment, the name description information includes the name description structure and field length; the classification module 506 is further configured to determine the name category of the item name as a first category if the name description structure of the item name is a pure character structure; determine the name category of the item name as a second category if the name description structure of the item name is not a pure character structure and the field length of the item name is greater than a preset length threshold; and determine the name category of the item name as a third category if the name description structure of the item name is not a pure character structure and the field length of the item name is not greater than a preset length threshold.

[0132] In one embodiment, the second extraction module 508 is further configured to, if the name category is the first category, search in a preset knowledge base according to the item name to obtain search results; and extract name keywords corresponding to the item name from the search results according to the semantic information of the search results.

[0133] In one embodiment, the second extraction module 508 is further configured to filter category keywords that characterize the item category from the search results based on the semantic information of the search results; and to remove effect keywords that describe the use effect of the item from the category keywords based on the semantic information of the search results and the position of the category keywords in the search results, so as to obtain the name keywords corresponding to the item name.

[0134] In one embodiment, the second extraction module 508 is further configured to extract the name keywords corresponding to the item name from the item name based on the semantic information of the item name if the name category is a second category; and to extract the name keywords corresponding to the item name from the item name if the name category is a third category.

[0135] In one embodiment, the identification module 510 is further configured to acquire an image of the item to be identified, identify the packaging information of the item to be identified from the image, extract packaging features from the packaging information, extract name features from name keywords, generate target features based on the packaging features and name features, and identify the item category of the item to be identified based on the target features to obtain the item category identification result.

[0136] In one embodiment, the identification module 510 is further configured to weight the packaging features according to a preset first weight to obtain packaging weighted features; weight the name features according to a preset second weight to obtain name weighted features; and perform feature fusion on the packaging weighted features and the name weighted features to obtain target features.

[0137] Each module in the aforementioned item name recognition device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the operations corresponding to each module.

[0138] In one exemplary embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 6 As shown, the computer device includes a processor, memory, input / output interfaces, a communication interface, a display unit, and an input device. The processor, memory, and input / output interfaces are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interfaces. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The input / output interfaces are used for exchanging information between the processor and external devices. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, Near Field Communication (NFC), or other technologies. When executed by the processor, the computer program implements an item name recognition method. The display unit is used to form a visually visible image and can be a display screen, a projection device, or a virtual reality imaging device. The display screen can be an LCD screen or an e-ink screen. The input device of the computer device can be a touch layer covering the display screen, or buttons, trackballs, or touchpads set on the casing of the computer device, or external keyboards, touchpads, or mice, etc.

[0139] Those skilled in the art will understand that Figure 6 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0140] In one exemplary embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0141] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps in the above method embodiments.

[0142] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0143] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0144] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0145] The above embodiments are merely illustrative of several implementation methods of this application, and their descriptions are relatively specific and detailed. However, they should not be construed as limiting the scope of this application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for identifying the name of an item, characterized in that, The method includes: Obtain the name of the item to be identified; Extract the name description information of the item name, which is used to characterize the name structure of the item name; Based on the name description information of the item name, the item name is classified to obtain the name category of the item name; Extract the name keywords corresponding to the item name according to the keyword extraction method corresponding to the name category; Based on the name keywords, the item category of the item to be identified is identified to obtain the item category identification result. The item category identification result is used to characterize the item category of the item to be identified at a preset classification granularity.

2. The method according to claim 1, characterized in that, The name description information includes the name description structure and field length; the step of classifying the item name according to the name description information to obtain the name category of the item name includes: If the name description structure of the item name is a pure character structure, then the name category of the item name is determined as the first category; If the name description structure of the item name is not a pure character structure, and the field length of the item name is greater than a preset length threshold, then the name category of the item name is determined as the second category; If the name description structure of the item name is not a pure character structure, and the field length of the item name is not greater than a preset length threshold, then the name category of the item name is determined to be the third category.

3. The method according to claim 2, characterized in that, The step of extracting name keywords corresponding to the item name according to the keyword extraction method corresponding to the name category includes: If the name category is the first category, then a search is performed in the preset knowledge base based on the item name to obtain the search results; Based on the semantic information of the search results, extract the name keywords corresponding to the item name from the search results.

4. The method according to claim 3, characterized in that, The step of extracting the name keywords corresponding to the item name from the search results based on the semantic information of the search results includes: Based on the semantic information of the search results, category keywords used to characterize item categories are filtered from the search results; Based on the semantic information of the search results and the position of the category keywords in the search results, the effect keywords used to describe the use effect of the item are removed from the category keywords to obtain the name keywords corresponding to the item name.

5. The method according to claim 3, characterized in that, The step of extracting the name keywords corresponding to the item name according to the keyword extraction method corresponding to the name category further includes: If the name category is the second category, then the name keywords corresponding to the item name are extracted from the item name based on the semantic information of the item name; If the name category is the third category, then extract the name keywords corresponding to the item name from the item name.

6. The method according to any one of claims 1 to 5, characterized in that, The step of identifying the item category of the item to be identified based on the name keywords, and obtaining the item category identification result, includes: Obtain the image of the item to be identified, and identify the packaging information of the item to be identified from the image. Extract packaging features from the product packaging information; Extract name features from the name keywords; Based on the packaging features and the name features, generate target features; Based on the target features, the item category of the item to be identified is determined, and the item category identification result is obtained.

7. The method according to claim 6, characterized in that, The step of generating target features based on the packaging features and the name features includes: The packaging features are weighted according to a preset first weight to obtain the packaging weighted features; The name features are weighted according to a preset second weight to obtain the name weighted features; The target feature is obtained by fusing the packaging weighted feature and the name weighted feature.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 7.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 7.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 7.

Images