Lost and found method and device, storage medium and computer equipment

Abstract

According to the lost property finding method and device, the storage medium and the computer equipment provided by the invention, when the lost property query instruction is received, blind search of information is avoided by determining the lost property information and accurately positioning the matching data range, and the search efficiency is improved. Meanwhile, the lost article information and the feature vectors in the matching data range are matched in multiple dimensions such as article categories, time and space, appearance features and semantic description, the similarity between the lost article and the lost and found information can be considered more comprehensively, and therefore the matching hit rate is increased. Besides, the feature vectors in the lost property database are divided into public, semi-hidden and completely-hidden information according to user types for multi-level display, so that the requirement of information sharing is met, personal privacy is effectively protected, and the problem of privacy disclosure caused by public display of the details of the lost property is avoided. Finally, the feature vectors with the comprehensive matching scores higher than the preset threshold value are pushed to the query terminal, and the accuracy and reliability of information pushing are ensured.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and in particular to a lost and found method, apparatus, storage medium and computer equipment. Background Technology

[0002] In traditional scenarios involving lost and found notices, information dissemination primarily relies on online social groups, WeChat Moments, and other social media platforms, or on simple lost and found registration systems. These methods constitute the main technological means for handling such matters, and their core model is one-way, decentralized information dissemination and passive waiting, lacking effective aggregation and proactive matching capabilities.

[0003] Currently, some intelligent lost and found platforms with basic aggregation and matching functions have emerged. However, the lack of standardized item descriptions results in a low matching rate during searches. Furthermore, publicly displaying lost item details easily leads to the leakage of personal privacy. Overall, existing technologies have significant shortcomings in terms of matching accuracy and privacy protection. Summary of the Invention

[0004] The purpose of this application is to address at least one of the aforementioned technical deficiencies, particularly the low matching rate in existing technologies due to the lack of standardized item descriptions. Furthermore, publicly displaying lost item details easily leads to the leakage of personal privacy. These technical deficiencies result in significant shortcomings in the existing technologies regarding matching accuracy and privacy protection.

[0005] Firstly, this application provides a method for finding lost and found items, the method comprising:

[0006] When a lost item query command is received, the lost item information corresponding to the lost item query command is determined;

[0007] The matching data range is determined in the preset lost and found database based on the lost item information; when each feature vector in the lost and found database is displayed on the lost and found platform, it is divided into public information, semi-hidden information or completely hidden information for multi-level display according to the type of user it is targeting.

[0008] In terms of item category, time and space, appearance features, and semantic description, the lost item information is matched with each feature vector in the matching data range in multiple dimensions to obtain a comprehensive matching score between the lost item information and each feature vector.

[0009] The feature vector with a comprehensive matching score of higher than a preset threshold within the matching data range is pushed to the terminal device that initiated the lost item query command.

[0010] In one embodiment, the update process of the lost and found database includes:

[0011] When a lost item retrieval registration instruction is received, an image of the lost item is acquired, and a preset image recognition model is used to perform multi-level feature extraction on the lost item image to obtain multi-level category information and lost item feature information;

[0012] Based on the multi-level category information and the lost item feature information, a feature vector of the lost item image is generated, and the feature vector is written into the lost item database to update the lost item database.

[0013] In one embodiment, determining the lost item information corresponding to the lost item query instruction includes:

[0014] Parse the lost and found query command to obtain the item information uploaded by the user;

[0015] Category information, spatiotemporal information, appearance information, and semantic information are extracted from the item information, and lost item information corresponding to the lost item query instruction is generated based on the category information, the spatiotemporal information, the appearance information, and the semantic information.

[0016] In one embodiment, the step of performing multi-dimensional matching between the lost item information and each feature vector within the matching data range to obtain a comprehensive matching score between the lost item information and each feature vector includes:

[0017] The lost item category is determined based on the lost item information, and the dimensional weights of the item category dimension, time and space dimension, appearance feature dimension and semantic description dimension are determined based on the lost item category.

[0018] For each feature vector within the matching data range, calculate the matching score between the lost item information and the feature vector in each dimension, and then sum the matching scores in each dimension according to the dimension weights corresponding to each dimension to obtain the comprehensive matching score between the lost item information and the feature vector.

[0019] In one embodiment, pushing the feature vector whose comprehensive matching score with the lost item information within the matching data range is higher than a preset threshold to the terminal device that initiated the lost item query command includes:

[0020] Identify the terminal device that initiated the lost item search command;

[0021] Obtain target feature vectors whose comprehensive matching score with the lost item information is higher than a preset threshold within the matching data range, and identify the public information and semi-hidden information in each target feature vector according to preset public rules;

[0022] The public and semi-hidden information of each target feature vector is pushed to the terminal device.

[0023] In one embodiment, the multi-level display of information based on the type of user being targeted—whether it is public, semi-hidden, or completely hidden—includes:

[0024] Determine the types of users who view information on the lost and found platform, including non-matching users, matching users, and owners of lost items;

[0025] When the user type is a non-matching user, extract and display the public information in each feature vector of the lost and found database;

[0026] When the user type is a matching user, extract and display the public and semi-hidden information in each feature vector of the lost and found database;

[0027] When the user type is the owner of the lost item, the public information, semi-hidden information, and fully hidden information in each feature vector of the lost and found database are extracted and displayed.

[0028] In one embodiment, after pushing feature vectors whose comprehensive matching score with the lost item information within the matching data range is higher than a preset threshold to the terminal device that initiated the lost item query command, the method further includes:

[0029] If the terminal device initiates an identity verification command, then the lost item feature vector corresponding to the identity verification command within the matching data range is determined, and the verification information uploaded by the terminal device is obtained;

[0030] The lost item feature vector is compared with the verification information. If the comparison is successful, the terminal device is identified as the owner of the lost item.

[0031] Secondly, this application provides a lost and found device, the device comprising:

[0032] The instruction receiving module is used to determine the lost item information corresponding to the lost item query instruction when a lost item query instruction is received;

[0033] The range determination module is used to determine the range of matching data in a preset lost and found database based on the lost item information; when each feature vector in the lost and found database is displayed on the lost and found platform, it is divided into public information, semi-hidden information or completely hidden information for multi-level display according to the type of user it is targeting;

[0034] The multi-dimensional matching module is used to perform multi-dimensional matching between the lost item information and each feature vector in the matching data range in terms of item category dimension, time and space dimension, appearance feature dimension and semantic description dimension, so as to obtain the comprehensive matching score between the lost item information and each feature vector.

[0035] The lost item push module is used to push feature vectors whose comprehensive matching score with the lost item information within the matching data range is higher than a preset threshold to the terminal device that initiates the lost item query command.

[0036] Thirdly, this application provides a storage medium storing computer-readable instructions that, when executed by one or more processors, cause the one or more processors to perform the steps of the lost and found method as described in any of the above embodiments.

[0037] Fourthly, this application provides a computer device, including: one or more processors, and a memory;

[0038] The memory stores computer-readable instructions, which, when executed by the one or more processors, perform the steps of the lost and found method as described in any of the above embodiments.

[0039] As can be seen from the above technical solutions, the embodiments of this application have the following advantages:

[0040] The lost and found method, apparatus, storage medium, and computer equipment provided in this application, upon receiving a lost and found query instruction, determine the lost item information and accurately locate the matching data range, avoiding blind searching and improving search efficiency. Simultaneously, matching the lost item information with feature vectors within the matching data range across multiple dimensions, including item category, time and space, appearance features, and semantic description, allows for a more comprehensive consideration of the similarity between the lost item and the found information, thereby improving the matching hit rate. Furthermore, the feature vectors in the lost item database are divided into public, semi-hidden, and fully hidden information based on user type for multi-level display, satisfying the need for information sharing while effectively protecting personal privacy and avoiding privacy leaks caused by publicly displaying lost item details. Finally, feature vectors with a comprehensive matching score higher than a preset threshold are pushed to the query terminal, ensuring the accuracy and reliability of the pushed information and achieving the dual goals of improving matching accuracy and protecting user privacy. Attached Figure Description

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

[0042] Figure 1 A flowchart illustrating a lost and found method provided in an embodiment of this application;

[0043] Figure 2 This is a schematic diagram of the structure of a lost and found device provided in an embodiment of this application;

[0044] Figure 3 This is an internal structural diagram of a computer device provided in an embodiment of this application. Detailed Implementation

[0045] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0046] In one embodiment, this application provides a method for finding lost items, and the following embodiments illustrate the application of this method to a lost and found platform. It is understood that the lost and found platform is a software application, which may be a website application, an app, or a mini-program, etc., and this application does not impose specific limitations on it.

[0047] like Figure 1 As shown, this application provides a lost and found method, the method comprising:

[0048] S101: When a lost item query instruction is received, determine the lost item information corresponding to the lost item query instruction.

[0049] Lost item information refers to a set of data describing the characteristics and related circumstances of lost items, including but not limited to item category, physical attributes, time of loss, location of loss, and special identifiers.

[0050] In this step, after a user loses an item, they can upload the item information to the lost and found platform to initiate a lost and found search request. When the lost and found platform receives the lost and found search request, it parses the request to determine the lost item information corresponding to the search request.

[0051] S102: Determine the range of matching data in the preset lost and found database based on the lost item information.

[0052] The lost and found database stores feature vectors representing lost item information. The matching data range refers to the range of data in the lost and found database that matches the spatiotemporal information of the lost items. When each feature vector in the lost and found database is displayed on the lost and found platform, it is categorized into public information, semi-hidden information, or completely hidden information for multi-level display, depending on the target user type.

[0053] In this step, after the lost item information is confirmed, the lost and found platform can extract the spatiotemporal information from the lost item information, including the time and location of loss. Using this spatiotemporal information, data matching this information is filtered through the lost item database, and the set of filtered data is used as the matching data range.

[0054] Specifically, when determining the matching data range in the lost and found database based on spatiotemporal information, the loss time in the spatiotemporal information can be expanded to obtain a query time range, and an area within a certain radius centered on the loss location in the spatiotemporal information can be delineated as the query location range. Then, data matching both the query time range and the query location range can be determined from the lost and found database, and a matching data range is formed based on the retrieved data.

[0055] S103: In terms of item category, time and space, appearance features and semantic description, the lost item information is matched with each feature vector in the matching data range in multiple dimensions to obtain the comprehensive matching score between the lost item information and each feature vector.

[0056] Among them, the feature vector refers to structured data composed of multi-dimensional attributes of lost item information.

[0057] In this step, the lost item information is first transformed into a vector, resulting in a lost item vector. Then, the lost item vector is matched against various feature vectors within the matching data range across the dimensions of item category, time and space, appearance features, and semantic description. The overall matching score between the lost item vector and each feature vector is then calculated. This overall matching score is used to quantify the similarity between the lost item information and the feature vector across each dimension.

[0058] Specifically, when performing vector transformation on lost item information, for item categories, it can be converted into a numerical representation of a category label; for time and space, the timestamp of the time of loss and the geographic coordinates of the location of loss can be extracted; for appearance features, features such as color and shape can be converted into numerical values ​​through preset quantization rules; for semantic description, natural language processing techniques, such as word embedding or text vectorization methods, can be used to convert text descriptions into vector form. The final result is the lost item vector.

[0059] Furthermore, when performing multi-dimensional matching between the lost item vector and various feature vectors within the matching data range, taking a specific feature vector as an example, for the item category dimension, classification algorithms can be used to calculate the similarity between the lost item vector and the item category represented by that feature vector; for the time-space dimension, the difference in timestamps and the distance of geographical coordinates can be calculated; for the appearance feature dimension, Euclidean distance or other similarity metrics can be used to calculate the matching degree of features such as color and shape; for the semantic description dimension, methods such as cosine similarity can be used to calculate the similarity between text vectors. Finally, the similarity scores from all dimensions are combined to obtain a comprehensive matching score.

[0060] S104: Push the feature vector with a comprehensive matching score of higher than the preset threshold with the lost item information within the matching data range to the terminal device that initiated the lost item query command.

[0061] The preset threshold is an empirical value that can be adjusted according to actual needs; it can generally be set to 0.8.

[0062] In this step, the overall matching score of each feature vector calculated previously is compared with a preset threshold. If the overall matching score of a feature vector is higher than the preset threshold, then this feature vector is considered to be highly relevant to the lost item information, and the lost item represented by that feature vector can be pushed to the corresponding terminal device. Furthermore, the push process can be carried out via network communication protocols, sending these matching feature vectors to the terminal device that initiated the lost item query command.

[0063] For example, on a mobile application-based lost and found platform, a user submits a lost item search request via the mobile app. The platform performs multi-dimensional matching of the lost item information with feature vectors within the matching data range on the backend and calculates a comprehensive matching score. Then, based on a preset threshold (e.g., 0.8, indicating a matching degree of 80% or higher), the platform filters out all feature vectors with comprehensive matching scores higher than this threshold. The platform then pushes the lost item information represented by these feature vectors to the user's mobile app in list form via the network. The user can then view this highly matched lost item information on their phone, including a detailed description of the item, the location where it was lost, and the time it was found.

[0064] Understandably, if there are no feature vectors whose comprehensive matching scores exceed the preset threshold, the range of matching data will be redefined periodically, and multi-dimensional matching and subsequent processes will be performed again.

[0065] In the above embodiments, upon receiving a lost item query command, the system determines the lost item information and accurately locates the matching data range, avoiding blind searching and improving search efficiency. Simultaneously, matching the lost item information with feature vectors within the matching data range across multiple dimensions, including item category, time and space, appearance features, and semantic description, allows for a more comprehensive consideration of the similarity between the lost item and the found information, thereby improving the matching hit rate. Furthermore, the feature vectors in the lost item database are categorized into public, semi-hidden, and fully hidden information based on user type for multi-level display. This satisfies the need for information sharing while effectively protecting personal privacy, avoiding privacy leaks caused by publicly displaying lost item details. Finally, feature vectors with a comprehensive matching score exceeding a preset threshold are pushed to the query terminal, ensuring the accuracy and reliability of the pushed information and achieving the dual goals of improving matching accuracy while protecting user privacy.

[0066] In one embodiment, the process of updating the lost and found database includes:

[0067] S1: When a lost item retrieval registration instruction is received, an image of the lost item is acquired, and a preset image recognition model is used to extract multi-level features from the lost item image to obtain multi-level category information and lost item feature information.

[0068] S2: Based on multi-level category information and lost item feature information, generate feature vectors for lost item images and write the feature vectors into the lost item database to update the lost item database.

[0069] Image recognition models refer to algorithmic models used to automatically identify and extract multi-level visual features from images. Multi-level category information refers to category labels obtained by hierarchically classifying lost item images, such as "electronic products" → "mobile phones" → "iPhone 14 Pro". Lost item feature information refers to detailed feature data extracted from images, such as color, texture, shape, brand logos, wear marks, etc.

[0070] In this embodiment, upon receiving a lost and found registration instruction, an image of the lost item is first acquired. This can be done by the user uploading an image, taking a photo on-site, or acquiring an image from another device. Next, a preset image recognition model is used to extract multi-level features from the lost item image. The image recognition model automatically analyzes the image content, extracting multi-level category information and lost item feature information. For example, the model might first identify the image as an "item," then further identify it as an "electronic product," and finally determine it as a "mobile phone." Simultaneously, the model also extracts feature information such as the lost item's color (e.g., black) and shape (e.g., rectangle).

[0071] Specifically, the multi-level feature extraction process of the image recognition model is implemented through deep learning algorithms. The model inputs image data into a neural network, which processes it through multiple convolutional layers, pooling layers, and fully connected layers, ultimately outputting multi-level category information and lost item feature information. This information is then further processed to generate feature vectors for the lost item image.

[0072] Furthermore, the training process of the image recognition model in this application includes: acquiring a batch of image samples and their corresponding multi-level category information and lost item feature information; using the image samples as training data and the multi-level category information and lost item feature information as sample labels; iteratively training a pre-trained model; continuously adjusting network parameters through a backpropagation algorithm during training to improve the model's feature recognition ability; and determining the pre-trained model as the image recognition model when the training conditions are met. These training conditions can be set to reach a preset number of iterations or to a loss accuracy less than a preset threshold; this application does not impose specific limitations on this.

[0073] Understandably, by extracting multi-level category information and lost item feature information through image recognition models, the characteristics of lost items can be quantified and structured, and detailed features of lost items can be deeply obtained. This can improve the efficiency of the query on the one hand, and the accuracy of subsequent matching on the other.

[0074] In one embodiment, determining the lost item information corresponding to the lost item query instruction includes:

[0075] S1: Parse the lost and found query command to obtain the item information uploaded by the user.

[0076] S2: Extract category information, spatiotemporal information, appearance information, and semantic information from the item information, and generate lost item information corresponding to the lost item query command based on the category information, spatiotemporal information, appearance information, and semantic information.

[0077] Among these, category information refers to the multi-level classification information to which the lost item belongs. Spatiotemporal information refers to the temporal and spatial attributes related to the loss event. Appearance information refers to the external features of the lost item that can be directly perceived visually. Semantic information refers to the semantic description, including notes and text in images.

[0078] In this embodiment, upon receiving a lost item query command, the system first parses the item information uploaded by the user. Specifically, the user-provided item information can be structured to extract category information, spatiotemporal information, appearance information, and semantic information. For example, natural language processing techniques (such as word segmentation, part-of-speech tagging, and named entity recognition) can be used to parse the user-input item information. Furthermore, based on the extracted category information, spatiotemporal information, appearance information, and semantic information, lost item information corresponding to the lost item query command is generated.

[0079] Specifically, by parsing lost item query commands and extracting category, spatiotemporal, appearance, and semantic information, user input can be systematically decomposed, transforming unstructured natural language into standardized, multi-dimensional representations of lost item information. Generating structured lost item information based on this information not only ensures the completeness and accuracy of query elements but also provides a clear and computable data foundation for subsequent matching and retrieval.

[0080] In one embodiment, the lost item information is matched with each feature vector within the matching data range in a multi-dimensional manner to obtain a comprehensive matching score between the lost item information and each feature vector, including:

[0081] S1: Determine the category of lost items based on the lost item information, and determine the dimensional weights of the item category dimension, time and space dimension, appearance feature dimension and semantic description dimension based on the lost item category.

[0082] S2: For each feature vector within the matching data range, calculate the matching score between the lost item information and the feature vector in each dimension, and sum the matching scores in each dimension according to the corresponding dimension weights to obtain the comprehensive matching score between the lost item information and the feature vector.

[0083] Among them, the dimension weight is used to measure the contribution of the corresponding dimension in the calculation of the comprehensive matching score.

[0084] In this embodiment, category information is extracted from the lost item information to determine the category of the lost item. Next, based on the lost item category, the dimensional weights for the item category dimension, temporal-spatial dimension, appearance feature dimension, and semantic description dimension are determined. These weights can be preset based on experience or data analysis. For example, for electronic products, the dimensional weight corresponding to the category dimension might be set to 0.4, the dimensional weight corresponding to the temporal-spatial dimension to 0.3, the dimensional weight corresponding to the appearance feature dimension to 0.2, and the dimensional weight corresponding to the semantic description dimension to 0.1.

[0085] Then, for each feature vector within the matching data range, the matching score between the lost item information and that feature vector is calculated across each dimension. For example, cosine similarity is used to calculate the matching score for the semantic description dimension, and Euclidean distance is used to calculate the matching score for the appearance feature dimension. Then, based on the dimension weights corresponding to each dimension, the matching scores across each dimension are weighted and summed. Assuming the matching score between the lost item information and a certain feature vector is 0.9 in the category dimension, 0.7 in the time-space dimension, 0.8 in the appearance feature dimension, and 0.6 in the semantic description dimension, the comprehensive matching score calculated based on the dimension weights in the example above is: 0.9 × 0.4 + 0.7 × 0.3 + 0.8 × 0.2 + 0.6 × 0.1 = 0.8.

[0086] It's understandable that dynamically assigning weights to different dimensions based on the lost item category ensures that more important features contribute more to the matching results, thus improving accuracy. In practical applications, this weighted matching method can adjust the matching strategy according to different lost item categories, making lost and found methods more flexible and efficient.

[0087] For example, the matching rules for each dimension can be set as follows:

[0088] (1) Item category matching (weight 40%): 1. Exactly the same: score 1, 2. Similar (e.g. "wallet" and "card holder"): score 0.8.

[0089] (2) Spatial and temporal range matching (weight 30%): 1. Time matching: if found within ±2 hours of the time of loss, score 1; 2. Location matching: within a radius of 500 meters, score 1, and the score decreases beyond that distance.

[0090] (3) Appearance feature matching (weight 20%): 1. Color matching: score 1, 2. Brand / model matching: score 1, 3. Special logo matching: score 1.

[0091] (4) Semantic description matching (weight 10%): Analyze the semantic similarity between the owner's description and the finder's description.

[0092] In one embodiment, a feature vector whose overall matching score with the lost item information within the matching data range is higher than a preset threshold is pushed to the terminal device that initiated the lost item query command, including:

[0093] S1: Determine the terminal device that initiated the lost item search command.

[0094] S2: Obtain target feature vectors whose comprehensive matching score with lost item information is higher than a preset threshold within the matching data range, and identify the public information and semi-hidden information in each target feature vector according to the preset public rules.

[0095] S3: Push the public and semi-hidden information of each target feature vector to the terminal device.

[0096] The preset public rules are used to distinguish between public information, semi-hidden information, and fully hidden information in the feature vectors. Public information is information that can be viewed by all users, while semi-hidden information is information that is only visible to the matching user.

[0097] In this embodiment, when pushing target feature vector information within the matching data range to the terminal device, the terminal device initiating the lost item query command can first be identified. This can be achieved through user login information, device identification (such as IP address, device ID), etc. Next, target feature vectors within the matching data range whose comprehensive matching score with the lost item information is higher than a preset threshold are obtained. Further, the public and semi-hidden information in each target feature vector are identified according to preset public rules, and finally, the public and semi-hidden information of each target feature vector is pushed to the terminal device.

[0098] For example, a preset public rule can be expressed as:

[0099] (1) Public information (visible to everyone): item category, color, approximate location, and time of finding.

[0100] (2) Semi-hidden information (visible only to the matched owner): brand, precise location, and photos of physical characteristics (sensitive information is blurred).

[0101] (3) Completely hidden information (visible after identity verification): list of items in wallet, high-definition photos, contact information.

[0102] It is understandable that by setting up public rules to distinguish between public and semi-hidden information, it can be ensured that users can obtain enough information to determine whether the lost item is related to them during the query process, without being exposed to too much sensitive information, thereby protecting the privacy of the owner and reducing the risk of information leakage.

[0103] In one embodiment, information is categorized into public, semi-hidden, or completely hidden information based on the type of user it is intended for, and displayed in multiple levels, including:

[0104] S1: Determine the type of user who is viewing information on the lost and found platform.

[0105] S2: When the user type is a non-matching user, extract and display the public information in each feature vector of the lost and found database; when the user type is a matching user, extract and display the public and semi-hidden information in each feature vector of the lost and found database; when the user type is the owner of the lost item, extract and display the public, semi-hidden, and fully hidden information in each feature vector of the lost and found database.

[0106] The user types include unmatched users, matched users, and lost user users. Users matched in multi-dimensional matching are considered matched users. After identity verification, matched users become lost user users, and the rest are unmatched users. Completely hidden information refers to information that is only visible to the lost user.

[0107] In this embodiment, after determining the user type, when the user type is a non-matching user, the platform extracts and displays the public information from each feature vector in the lost and found database. When the user type is a matching user, in addition to extracting the public information, the platform also extracts and displays semi-hidden information, such as partial contact information of the owner or more detailed item characteristics. When the user type is the owner, the platform extracts and displays public information, semi-hidden information, and completely hidden information, such as sensitive information like the owner's detailed contact information.

[0108] Specifically, by identifying the types of users viewing information on the lost and found platform and displaying different levels of information based on these user types, the platform can meet the needs of different users and improve the efficiency of lost and found services while protecting privacy and information security. Furthermore, it can effectively reduce the risk of information leakage and protect the privacy of both the owners and the users.

[0109] In one embodiment, after pushing the feature vector with a comprehensive matching score of higher than a preset threshold with the lost item information within the matching data range to the terminal device that initiated the lost item query command, the lost item retrieval method further includes:

[0110] S1: If the terminal device initiates an identity verification command, determine the lost item feature vector corresponding to the identity verification command within the matching data range, and obtain the verification information uploaded by the terminal device.

[0111] S2: Compare the lost item feature vector with the verification information. If the comparison is successful, the terminal device is identified as the owner user.

[0112] Verification information refers to the information provided by a user to prove that they are the owner of the lost item.

[0113] In this embodiment, when a terminal device initiates an identity verification command, the lost item feature vector corresponding to the command within the matching data range can first be determined. Specifically, this can be done by recognizing the user-initiated identity verification command and finding the lost item feature vector related to the command within the matching data range. For example, a user might select a specific lost item record on the platform and initiate an identity verification command; the platform will determine the corresponding lost item feature vector based on the user's selection. Next, the verification information uploaded by the terminal device is obtained. By comparing the lost item feature vector with the verification information, if the comparison is successful, the user corresponding to the terminal device is identified as the owner of the lost item. Afterwards, contact can be made, and arrangements for collection can be made. This identity verification mechanism improves the security and accuracy of lost item collection, thereby preventing others from fraudulently claiming lost items.

[0114] For example, the verification information that the platform requires users to provide may include: "Please describe which bank cards are in your wallet?" (open-ended question), "Approximately how much cash is in your wallet?" (amount range), "What special markings are on your wallet?" (multiple choice).

[0115] In one embodiment, the lost and found method of this application can also access an electronic map to mark the location of lost items and provide a filtering function, allowing filtering by category, time, and location. Furthermore, when marking the location of lost items, a heat map of high-incidence areas for lost items can be generated to facilitate better management by the venue.

[0116] In one embodiment, the lost and found method of this application can also be shared with various different regional locations to achieve one-stop inquiry.

[0117] It should be understood that although the steps in the flowcharts of the embodiments described above 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 embodiments described above 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.

[0118] The lost and found device provided in the embodiments of this application is described below. The lost and found device described below can be referred to in correspondence with the lost and found method described above.

[0119] like Figure 2As shown, this application provides a lost and found device 200, the device comprising:

[0120] The instruction receiving module 201 is used to determine the lost item information corresponding to the lost item query instruction when it receives the lost item query instruction;

[0121] The scope determination module 202 is used to determine the range of matching data in the preset lost and found database based on the lost and found information; when each feature vector in the lost and found database is displayed on the lost and found platform, it is divided into public information, semi-hidden information or completely hidden information for multi-level display according to the type of user it is targeting.

[0122] The multidimensional matching module 203 is used to perform multidimensional matching between lost item information and various feature vectors within the matching data range in terms of item category dimension, time and space dimension, appearance feature dimension and semantic description dimension, so as to obtain the comprehensive matching score between lost item information and each feature vector.

[0123] The lost item push module 204 is used to push feature vectors with a comprehensive matching score of higher than a preset threshold with the lost item information within the matching data range to the terminal device that initiates the lost item query command.

[0124] In the above embodiments, upon receiving a lost item query command, the system determines the lost item information and accurately locates the matching data range, avoiding blind searching and improving search efficiency. Simultaneously, matching the lost item information with feature vectors within the matching data range across multiple dimensions, including item category, time and space, appearance features, and semantic description, allows for a more comprehensive consideration of the similarity between the lost item and the found information, thereby improving the matching hit rate. Furthermore, the feature vectors in the lost item database are categorized into public, semi-hidden, and fully hidden information based on user type for multi-level display. This satisfies the need for information sharing while effectively protecting personal privacy, avoiding privacy leaks caused by publicly displaying lost item details. Finally, feature vectors with a comprehensive matching score exceeding a preset threshold are pushed to the query terminal, ensuring the accuracy and reliability of the pushed information and achieving the dual goals of improving matching accuracy while protecting user privacy.

[0125] In one embodiment, the range determination module includes:

[0126] The feature extraction submodule is used to acquire the image of the lost item when a lost item retrieval registration instruction is received, and to perform multi-level feature extraction on the lost item image using a preset image recognition model to obtain multi-level category information and lost item feature information.

[0127] The vector writing submodule is used to generate feature vectors for lost item images based on multi-level category information and lost item feature information, and write the feature vectors into the lost item database to update the lost item database.

[0128] In one embodiment, the instruction receiving module includes:

[0129] The instruction parsing submodule is used to parse lost item query instructions in order to obtain item information uploaded by the user;

[0130] The information generation submodule is used to extract category information, spatiotemporal information, appearance information, and semantic information from item information, and generate lost item information corresponding to the lost item query command based on the category information, spatiotemporal information, appearance information, and semantic information.

[0131] In one embodiment, the multidimensional matching module includes:

[0132] The weight determination submodule is used to determine the category of lost items based on the lost item information, and to determine the dimensional weights of the item category dimension, time and space dimension, appearance feature dimension and semantic description dimension based on the lost item category;

[0133] The scoring calculation submodule is used to calculate the matching score between the lost item information and the feature vector in each dimension for each feature vector within the matching data range, and to perform a weighted summation of the matching scores in each dimension according to the corresponding dimension weights to obtain the comprehensive matching score between the lost item information and the feature vector.

[0134] In one embodiment, the lost item notification module includes:

[0135] The device identification submodule is used to identify the terminal device that initiates the lost and found inquiry command;

[0136] The information recognition submodule is used to obtain target feature vectors whose comprehensive matching score with lost item information is higher than a preset threshold within the matching data range, and to identify the public information and semi-hidden information in each target feature vector according to the preset public rules.

[0137] The information push submodule is used to push the public and semi-hidden information of each target feature vector to the terminal device.

[0138] In one embodiment, the range determination module includes:

[0139] The User Identification Submodule is used to determine the type of user who is viewing information on the lost and found platform. User types include non-matching users, matching users, and lost owner users.

[0140] The first display submodule is used to extract and display the public information in each feature vector of the lost and found database when the user type is a non-matching user.

[0141] The second display submodule is used to extract and display the public and semi-hidden information in each feature vector of the lost and found database when the user type is a matching user.

[0142] The third display submodule is used to extract and display public, semi-hidden, and fully hidden information from each feature vector in the lost and found database when the user type is the owner of the lost item.

[0143] In one embodiment, the lost and found device further includes:

[0144] The identity verification module is used to determine the lost item feature vector corresponding to the identity verification command within the matching data range and to obtain the verification information uploaded by the terminal device if the terminal device initiates an identity verification command.

[0145] The information comparison module is used to compare the lost item feature vector with the verification information. If the comparison is successful, the terminal device is identified as the owner of the lost item.

[0146] The division of modules in the above-described lost and found device is merely illustrative. In other embodiments, the lost and found device can be divided into different modules as needed to complete all or part of its functions. Each module in the above-described lost and found 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 in software form, so that the processor can call and execute the operations corresponding to each module.

[0147] In one embodiment, this application also provides a storage medium storing computer-readable instructions that, when executed by one or more processors, cause the one or more processors to perform the steps of the lost and found method as described in any of the above embodiments.

[0148] In one embodiment, this application also provides a computer device storing computer-readable instructions that, when executed by one or more processors, cause the one or more processors to perform the steps of the lost and found method as described in any of the above embodiments.

[0149] Indicatively, such as Figure 3 As shown, Figure 3 This is a schematic diagram of the internal structure of a computer device 300 provided in an embodiment of this application. The computer device 300 can be provided as a server. (Refer to...) Figure 3The computer device 300 includes a processing component 302, which further includes one or more processors, and memory resources represented by memory 301 for storing instructions, such as application programs, that can be executed by the processing component 302. The application programs stored in memory 301 may include one or more modules, each corresponding to a set of instructions. Furthermore, the processing component 302 is configured to execute instructions to perform the lost and found method of any of the above embodiments.

[0150] The computer device 300 may also include a power supply component 303 configured to perform power management of the computer device 300, a wired or wireless network interface 304 configured to connect the computer device 300 to a network, and an input / output (I / O) interface 305. The computer device 300 may operate on an operating system stored in memory 301, such as Windows Server™, Mac OS X™, Unix™, Linux™, Free BSD™, or similar.

[0151] Those skilled in the art will understand that Figure 3 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.

[0152] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising a…" does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element. In this document, the singular forms "a," "an," and "the" may also include the plural forms unless the context clearly indicates otherwise. It should also be understood that the terms “comprising / including” or “having” specify the presence of the stated features, wholes, steps, operations, components, parts or combinations thereof, but do not exclude the possibility of the presence or addition of one or more other features, wholes, steps, operations, components, parts or combinations thereof. Meanwhile, the term “and / or” as used in this specification includes any and all combinations of the associated listed items.

[0153] The various embodiments in this specification are described in a progressive manner. Each embodiment focuses on the differences from other embodiments. The various embodiments can be combined as needed, and the same or similar parts can be referred to each other.

[0154] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for finding lost and found items, characterized in that, The method includes: When a lost item query command is received, the lost item information corresponding to the lost item query command is determined; The matching data range is determined in the preset lost and found database based on the lost item information; when each feature vector in the lost and found database is displayed on the lost and found platform, it is divided into public information, semi-hidden information or completely hidden information for multi-level display according to the type of user it is targeting. In terms of item category, time and space, appearance features, and semantic description, the lost item information is matched with each feature vector in the matching data range in multiple dimensions to obtain a comprehensive matching score between the lost item information and each feature vector. The feature vector with a comprehensive matching score of higher than a preset threshold within the matching data range is pushed to the terminal device that initiated the lost item query command.

2. The lost and found method according to claim 1, characterized in that, The update process of the lost and found database includes: When a lost item retrieval registration instruction is received, an image of the lost item is acquired, and a preset image recognition model is used to perform multi-level feature extraction on the lost item image to obtain multi-level category information and lost item feature information; Based on the multi-level category information and the lost item feature information, a feature vector of the lost item image is generated, and the feature vector is written into the lost item database to update the lost item database.

3. The lost and found method according to claim 1, characterized in that, The step of determining the lost item information corresponding to the lost item query instruction includes: Parse the lost and found query command to obtain the item information uploaded by the user; Category information, spatiotemporal information, appearance information, and semantic information are extracted from the item information, and lost item information corresponding to the lost item query instruction is generated based on the category information, the spatiotemporal information, the appearance information, and the semantic information.

4. The lost and found method according to claim 1, characterized in that, The step of performing multi-dimensional matching between the lost item information and each feature vector within the matching data range to obtain a comprehensive matching score between the lost item information and each feature vector includes: The lost item category is determined based on the lost item information, and the dimensional weights of the item category dimension, time and space dimension, appearance feature dimension and semantic description dimension are determined based on the lost item category. For each feature vector within the matching data range, calculate the matching score between the lost item information and the feature vector in each dimension, and then sum the matching scores in each dimension according to the dimension weights corresponding to each dimension to obtain the comprehensive matching score between the lost item information and the feature vector.

5. The lost and found method according to claim 1, characterized in that, The step of pushing the feature vector with a comprehensive matching score higher than a preset threshold with the lost item information within the matching data range to the terminal device that initiated the lost item query command includes: Identify the terminal device that initiated the lost item search command; Obtain target feature vectors whose comprehensive matching score with the lost item information is higher than a preset threshold within the matching data range, and identify the public information and semi-hidden information in each target feature vector according to preset public rules; The public and semi-hidden information of each target feature vector is pushed to the terminal device.

6. The lost and found method according to claim 1, characterized in that, The method of categorizing information into public, semi-hidden, or completely hidden categories based on the type of user being targeted, and displaying it in multiple levels, includes: Determine the types of users who view information on the lost and found platform, including non-matching users, matching users, and owners of lost items; When the user type is a non-matching user, extract and display the public information in each feature vector of the lost and found database; When the user type is a matching user, extract and display the public and semi-hidden information in each feature vector of the lost and found database; When the user type is the owner of the lost item, the public information, semi-hidden information, and fully hidden information in each feature vector of the lost and found database are extracted and displayed.

7. The lost and found method according to any one of claims 1 to 6, characterized in that, After pushing feature vectors whose comprehensive matching score with the lost item information within the matching data range is higher than a preset threshold to the terminal device that initiated the lost item query command, the method further includes: If the terminal device initiates an identity verification command, then the lost item feature vector corresponding to the identity verification command within the matching data range is determined, and the verification information uploaded by the terminal device is obtained; The lost item feature vector is compared with the verification information. If the comparison is successful, the terminal device is identified as the owner of the lost item.

8. A lost and found device, characterized in that, The device includes: The instruction receiving module is used to determine the lost item information corresponding to the lost item query instruction when a lost item query instruction is received; The range determination module is used to determine the range of matching data in a preset lost and found database based on the lost item information; when each feature vector in the lost and found database is displayed on the lost and found platform, it is divided into public information, semi-hidden information or completely hidden information for multi-level display according to the type of user it is targeting; The multi-dimensional matching module is used to perform multi-dimensional matching between the lost item information and each feature vector in the matching data range in terms of item category dimension, time and space dimension, appearance feature dimension and semantic description dimension, so as to obtain the comprehensive matching score between the lost item information and each feature vector. The lost item push module is used to push feature vectors whose comprehensive matching score with the lost item information within the matching data range is higher than a preset threshold to the terminal device that initiates the lost item query command.

9. A storage medium, characterized in that: The storage medium stores computer-readable instructions that, when executed by one or more processors, cause the one or more processors to perform the steps of the lost and found method as described in any one of claims 1 to 7.

10. A computer device, characterized in that, include: One or more processors, and memory; The memory stores computer-readable instructions that, when executed by the one or more processors, perform the steps of the lost and found method as described in any one of claims 1 to 7.

Images