Methods, devices, equipment, and storage media for detecting order screenshot tampering

Abstract

This application relates to the field of logistics order verification technology, specifically to a method, apparatus, device, and storage medium for detecting order screenshot tampering. The method includes: extracting the area containing key information from the order screenshot corresponding to an e-commerce return order to obtain at least one key information image, wherein the key information is related to the normal flow of the return business; extracting features based on the key information image to obtain key information features; comparing the key information features with historical image features to determine whether the order screenshot has been tampered with, wherein the historical image features are the features corresponding to the key information in historical order screenshots that have not been tampered with. This application can improve the accuracy of identifying whether the order screenshot corresponding to an e-commerce return order has been tampered with.

Description

Technical Field

[0001] This application relates to the field of logistics order verification technology, specifically to a method, apparatus, equipment, and storage medium for detecting order screenshot tampering. Background Technology

[0002] With the continuous development of the e-commerce industry, online shopping has become an important way for consumers to shop. At the same time, e-commerce returns are also increasing. This expansion poses significant challenges to both e-commerce platforms and the logistics industry. Faced with the huge volume of returns, some logistics companies have implemented restrictions to ensure smooth operations. Specifically, when consumers submit e-commerce return orders, they must upload a screenshot of the order containing key information, and consumers are not allowed to modify this information. This ensures that the logistics service can extract the correct address and other key information from the e-commerce return order to complete the return process.

[0003] However, some unscrupulous consumers still tamper with key information in e-commerce return orders to obtain illegal benefits such as price differences. Currently, existing technology can detect duplicates using the similarity of full-image features. However, tampered order screenshots differ from ordinary tampered template images because, within the same e-commerce platform, not only are tampered order screenshots very similar to normal order screenshots, but different normal order screenshots also show high similarity. Therefore, accurately identifying whether the order screenshots corresponding to e-commerce return orders have been tampered with has become a pressing problem that needs to be solved. Summary of the Invention

[0004] Based on the defects and shortcomings of the prior art, this application proposes a method, apparatus, device and storage medium for detecting order screenshot tampering, which can improve the accuracy of identification when identifying whether the order screenshot corresponding to an e-commerce return order has been tampered with.

[0005] According to a first aspect of the embodiments of this application, a method for detecting order screenshot tampering is provided, comprising: extracting the area containing key information from an order screenshot corresponding to an e-commerce return order to obtain at least one key information image, wherein the key information is information related to the normal flow of the return business; performing feature extraction based on the key information image to obtain key information features; comparing the key information features with historical image features to determine whether the order screenshot has been tampered with, wherein the historical image features are features corresponding to the key information in historical order screenshots that have not been tampered with.

[0006] According to the order screenshot tampering detection method provided in the first aspect of the embodiments of this application, the key information image includes a product information image and a merchant information image; the step of extracting the area where the key information is located in the order screenshot corresponding to the e-commerce return order to obtain at least one key information image includes: extracting the returned product information in a first area in the order screenshot corresponding to the e-commerce return order to obtain the product information image, and extracting the returned merchant information in a second area to obtain the merchant information image.

[0007] According to the order screenshot tampering detection method provided in the first aspect of the embodiments of this application, the step of extracting features based on the key information image to obtain key information features includes: extracting features from the product information image to obtain product information features; concatenating the product information image and the merchant information image to obtain a combined information image; and extracting features from the combined information image to obtain combined information features; wherein, the key information features include the product information features and the combined information features.

[0008] According to the order screenshot tampering detection method provided in the first aspect of the embodiments of this application, the step of comparing the key information features with historical image features to determine whether the order screenshot has been tampered with includes: obtaining the product similarity between the product information features and each historical product feature, and determining the maximum product similarity among the product similarities, wherein the historical product features are the features corresponding to the returned product information in historical order screenshots that have not been tampered with in the past; determining whether the maximum product similarity is less than a first similarity threshold, if yes, then determining that the order screenshot has not been tampered with, if no, then determining whether the order screenshot has been tampered with based on the combined information features.

[0009] According to the order screenshot tampering detection method provided in the first aspect of the embodiments of this application, the step of determining whether the order screenshot has been tampered with based on the combined information features includes: obtaining the combined similarity between the combined information features and each historical combined feature, and determining the maximum combined similarity among the combined similarities, wherein the historical combined features are the features corresponding to the returned goods information and the returned merchant information in historical order screenshots that have not been tampered with in the past; determining whether the maximum combined similarity is greater than a second similarity threshold, if yes, then determining that the order screenshot has not been tampered with, if no, then determining that the order screenshot has been tampered with.

[0010] According to the order screenshot tampering detection method provided in the first aspect of the embodiments of this application, the order screenshot is contained in the return request submitted by the consumer user in real time; after comparing the key information features with historical image features to determine whether the order screenshot has been tampered with, the method further includes: if it is determined that the order screenshot has not been tampered with, then allowing the return request; if it is determined that the order screenshot has been tampered with, then blocking the return request.

[0011] The method for detecting order screenshot tampering provided in the first aspect of the embodiments of this application further includes: obtaining the key information features based on the order screenshot, and obtaining the historical image features based on the historical order screenshot, using the same feature acquisition method.

[0012] According to a second aspect of the embodiments of this application, an order screenshot tampering detection device is provided, comprising: an image processing module, configured to extract the area containing key information from an order screenshot corresponding to an e-commerce return order to obtain at least one key information image, wherein the key information is information related to the normal flow of the return business; a feature processing module, configured to perform feature extraction based on the key information image to obtain key information features; and a judgment module, configured to compare the key information features with historical image features to determine whether the order screenshot has been tampered with, wherein the historical image features are features corresponding to the key information in historical order screenshots that have not been tampered with.

[0013] According to a third aspect of the embodiments of this application, an electronic device is provided, including: a memory and a processor; the memory is connected to the processor and is used to store a program; the processor is used to implement the order screenshot tampering detection method as described in the first aspect by running the program in the memory.

[0014] According to a fourth aspect of the embodiments of this application, a storage medium is provided, on which a computer program is stored, and when the computer program is run by a processor, it implements the order screenshot tampering detection method as described in the first aspect.

[0015] In this embodiment, the key information area is extracted from the order screenshot corresponding to an e-commerce return order to obtain at least one key information image. The key information is related to the normal flow of the return business. Feature extraction is performed based on the key information image to obtain key information features. The key information features are compared with historical image features to determine whether the order screenshot has been tampered with. The historical image features are the features corresponding to the key information in historical order screenshots that have not been tampered with. In the above solution, given the special nature of order screenshots, instead of directly judging whether the order screenshot has been tampered with based on the global features of the order screenshot, it selectively extracts key information images from the key information area and performs targeted feature extraction on some but more critical key information images. This avoids interference from irrelevant information in the feature comparison process, thereby improving the accuracy of identifying whether the order screenshot has been tampered with based on the comparison of key information features and historical image features. Attached Figure Description

[0016] 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 embodiments of this application. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0017] Figure 1 This is a flowchart illustrating a method for detecting order screenshot tampering, provided in an embodiment of this application.

[0018] Figure 2 This is an example screenshot of an order provided in an embodiment of this application.

[0019] Figure 3 This is one of the example images of order screenshot tampering provided in the embodiments of this application.

[0020] Figure 4 This is the second example of order screenshot tampering provided in this application embodiment.

[0021] Figure 5 This is an example image of product information provided in an embodiment of this application.

[0022] Figure 6 This is an example diagram of a combined information image provided in an embodiment of this application.

[0023] Figure 7 This is the third example of order screenshot tampering provided in this application embodiment.

[0024] Figure 8 This is an example of an unaltered order screenshot provided in an embodiment of this application.

[0025] Figure 9 A block diagram of an order screenshot tampering detection device provided in an embodiment of this application.

[0026] Figure 10 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0027] 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.

[0028] Exemplary methods This application addresses the problem in existing technologies that cannot accurately identify whether order screenshots have been tampered with, by providing a method for detecting order screenshot tampering. This method can be implemented as a software algorithm, which can run on any device with data processing capabilities, such as servers of e-commerce after-sales platforms, logistics after-sales platforms, or smart mobile devices. The scope of protection of this application is not limited to the type of device on which the software algorithm for detecting order screenshot tampering runs.

[0029] In one embodiment, such as Figure 1 As shown, the implementation process of the method for detecting order screenshot tampering includes the following steps: Step 101: Extract the area containing key information from the order screenshot corresponding to the e-commerce return order, and obtain at least one key information image. The key information is related to the normal flow of the return business.

[0030] In this embodiment, an e-commerce return order is a unique and structured order created online based on the consumer's return request. An order screenshot refers to a static image file generated by capturing the e-commerce return order interface displayed on an electronic device screen using a screenshot tool. Based on the order screenshot as static image evidence generated in digital transactions to fix and display the status of an e-commerce return order at a specific moment, further analysis is performed on the order screenshot of the e-commerce return order. For example... Figure 2 The order screenshot example shown has characteristics such as a single text color, a single background color, and a lack of natural transitions between colors. Figure 2(No colors are currently used; the different areas of lines and text are used as examples.) Additionally, various pieces of information in order screenshots are generally in relatively fixed positions, and the arrangement and expression of this information are also relatively fixed. For example, merchant information and product information are located in relatively fixed positions in the order screenshot. Figure 2 The screenshot shown is from the same e-commerce platform. The merchant information does not appear above the return method selection button. Similarly, the wording of information such as "Please select a return method" and "Have you encountered the following problems?" generally remains unchanged in screenshots from the same e-commerce platform. Therefore, the unique characteristics of order screenshots in terms of color, information layout, and information expression provide a reliable basis for extracting information from order screenshots.

[0031] In this embodiment, when unscrupulous consumers tamper with screenshots of e-commerce return orders for malicious purposes, they primarily target key information in the screenshots that affects the normal flow of the return process. For example, such as... Figure 3 As shown, the correct first merchant address information in the merchant information is altered to an incorrect second merchant address information; for example, such as... Figure 4 As shown, the correct first product image in the product information is altered to an incorrect second product image. However, fixed prompts in the order screenshot such as "Merchant Processing → Return Goods → Refund Completed," "Please Select Return Method," and "Have You Encountered the Following Problems?" as well as fixed functional buttons such as "View," "Platform Access," and "Cancel Application," are typically not altered. These less risky prompts and functional buttons can interfere with subsequent feature extraction and comparison processes in the order screenshot, resulting in extracted features that cannot accurately represent key information.

[0032] Based on the above, and taking into account the unique characteristics of order screenshots in terms of color, information layout, and information expression, and considering the importance of key information, the order screenshots are extracted to extract the areas containing key information, forming more targeted key information images with less distracting information.

[0033] In this embodiment, extracting the region containing key information to obtain at least one key information image includes: detecting the region containing key information in the order screenshot, extracting the region containing key information, and obtaining the key information image. Optionally, when detecting and extracting the region containing key information in the order screenshot, traditional image processing methods such as threshold segmentation based on grayscale value thresholds or edge detection methods based on grayscale values ​​can be used; alternatively, a classifier can be used to classify the features extracted by the sliding window; or the order screenshot can be input into a pre-trained deep learning detection model, and the deep learning detection model outputs the key information image. The deep learning detection model is trained on a neural network model using a pre-labeled detection training sample set. Optionally, the basic neural network architecture of the deep learning detection model can adopt the YOLO (You Only Look Once) framework, the Single Shot MultiBox Detector (SSD), or any other reasonable deep learning detection architecture. The specific method used to extract the key information area from the order screenshot can be flexibly determined based on various requirements such as actual situation, performance requirements, and specific needs. The scope of protection of this application is not limited to the specific implementation method of extracting the key information area from the order screenshot.

[0034] Step 102: Extract features based on the key information image to obtain key information features.

[0035] In this embodiment, after obtaining the key information image, feature extraction is performed on the key information image, which is more targeted and has less interference information, to obtain key information features that can more accurately represent the key information. This allows for easier feature comparison in the subsequent process, enabling the determination of whether the order screenshot has been tampered with, thus improving the efficiency of order screenshot processing. Optionally, when extracting key information features based on the key information image, feature capture can be performed based on features such as the color, texture, and shape of the image. Alternatively, the key information image can be input into a pre-trained deep learning extraction model, which outputs the key information features. The deep learning extraction model is trained on a neural network model using a pre-labeled feature extraction training sample set. Optionally, the basic neural network architecture of the deep learning extraction model can be a self-supervised visual model architecture such as Data Intensive Object Detection with Self-Supervision version 2 (DINOv2), a deep learning classification model architecture such as Residual Network (ResNet), or any other reasonable deep learning architecture. The specific methods used for key information feature extraction can be flexibly determined based on various requirements such as actual conditions, performance requirements, and specific needs. The scope of protection of this application is not limited to the specific implementation method of key information feature extraction.

[0036] Step 103: Compare the key information features with the historical image features to determine whether the order screenshot has been tampered with. The historical image features are the features corresponding to the key information in historical order screenshots that have not been tampered with in the past.

[0037] In this embodiment, historical order screenshots that have not been tampered with are collected. Based on the key information contained in these screenshots, features of the same category of key information are extracted from each tampered historical order screenshot, thereby obtaining the historical image features corresponding to each tampered historical order screenshot. Since the historical image features are extracted from tampered historical order screenshots, they serve as a tamper-free standard. By comparing the key information features with the historical image features, it is possible to accurately determine whether the order screenshot has been tampered with. Furthermore, based on the aforementioned implementation methods such as key information location detection, key information image extraction, and image data featureization, the accuracy of the key information features can be further improved, thereby enhancing the accuracy of identifying whether the order screenshot has been tampered with.

[0038] In this embodiment, the key information features and historical image features can adopt any reasonable feature representation method such as feature vector or feature matrix. Specifically, the feature representation method of key information features and historical image features can be flexibly set and adjusted according to the feature extraction method (such as deep learning extraction model can output feature vector of a certain dimension), specific situation and needs.

[0039] In one embodiment, the key information images include product information images and merchant information images. In the order screenshot corresponding to an e-commerce return order, extracting the area containing key information to obtain at least one key information image includes: extracting the returned product information from a first area in the order screenshot corresponding to the e-commerce return order to obtain a product information image, and extracting the returned merchant information from a second area to obtain a merchant information image.

[0040] In this embodiment, key information includes returned product information and returning merchant information. Returned product information refers to the relevant information about the products to be returned in this return transaction, including but not limited to any one or more of the following: returned product images, return rules, quantity, type, price, and detailed description of the returned products. Returning merchant information refers to the relevant information about the merchant receiving the returned products in this return transaction, including but not limited to any one or more of the following: merchant address, merchant name, and merchant contact number. In most cases, for example... Figure 2 The order screenshot shown illustrates this: returned product information and returned merchant information are located in different areas. Therefore, the returned product information (located in the first area) and the returned merchant information (located in the second area) can be extracted separately. This allows for targeted extraction of multiple key information areas within the same order screenshot, enhancing the specificity and resistance to interference in key information extraction. In real-world business scenarios, returned product information and returned merchant information are the two types of key information most likely to be tampered with by malicious consumers. Furthermore, these two types of information have the greatest impact on the return process. Tampering with either returned product information or returned merchant information can severely disrupt the normal flow of the return process, potentially leading to failed returns, disputes, and other problems. Therefore, targeted extraction of the areas containing returned product information and returned merchant information can further improve the accuracy of identifying whether an order screenshot has been tampered with.

[0041] Of course, depending on the actual situation and needs, the area where the key information to be extracted is located can also be other areas. The key information image can also include images from other areas. For example, based on the area where the merchant address information is located, the area where the product image is located, etc., more granular information can be obtained, such as merchant address information images and product images. The area where the key information to be extracted is located can be flexibly set and adjusted.

[0042] In one embodiment, feature extraction based on a key information image is performed to obtain key information features, including: extracting features from a product information image to obtain product information features; stitching together the product information image and the merchant information image to obtain a combined information image; and extracting features from the combined information image to obtain combined information features; wherein the key information features include product information features and combined information features.

[0043] In this embodiment, due to the complexity of return transactions in real-world business scenarios—for example, when the same merchant sells the same product, different consumers may return the same product details in the order screenshots; or, when the same merchant sells different products, consumers may return different products to the same merchant—single product information features or merchant information features are insufficient to accurately determine whether an order screenshot has been maliciously tampered with. Therefore, based on the business logic of the return transaction, the key information features specifically employ product information features and combined information features. These multimodal, multi-perspective product information features and combined information features provide a reliable basis for cross-validation to accurately identify whether an order screenshot has been tampered with.

[0044] In this embodiment, product information features can be extracted from independent product information images, for example, such as... Figure 5 The exemplary product information image shown can be used to extract product information features from the product information image by directly employing any of the reasonable key information feature extraction implementation methods provided in the above embodiments. For combined information features, such as... Figure 6 The exemplary combined information image shown can be obtained by first stitching together the product information image and the merchant information image; then, any reasonable key information feature extraction implementation method provided in the above embodiments can be directly used to extract the combined information features from the combined information image.

[0045] In another specific implementation, feature extraction is performed based on key information images to obtain key information features, including: extracting features from product information images to obtain product information features; extracting features from merchant information images to obtain merchant information features; and combining product information features and merchant information features to obtain combined information features. Optionally, any reasonable implementation method for key information feature extraction provided in the above embodiments can be directly adopted to extract merchant information features from merchant information images. Optionally, when combining product information features and merchant information features, any reasonable combination method such as direct splicing or fusion based on mathematical operations can be used to obtain the final combined information features. However, splicing product information images and merchant information images before extracting features, compared to extracting product information features and merchant information features separately and then combining them, provides the combined information feature extraction process (especially when using a deep learning extraction model) with original visual information that preserves complete spatial relationships, thus having the opportunity to discover pixel-level tampering traces and consistency relationships, resulting in higher accuracy of the extracted combined information features.

[0046] In one embodiment, comparing key information features with historical image features to determine whether an order screenshot has been tampered with includes: obtaining the product similarity between product information features and various historical product features, and determining the maximum product similarity among the various product similarities, wherein the historical product features are the features corresponding to returned product information in historical order screenshots that have not been tampered with in the past; determining whether the maximum product similarity is less than a first similarity threshold, if so, determining that the order screenshot has not been tampered with, if not, determining whether the order screenshot has been tampered with based on the combined information features.

[0047] In this embodiment, based on key information features and historical image features, a multi-stage decision-making process is implemented to determine whether an order screenshot has been tampered with, thus better reflecting the actual return business logic. Specifically, firstly, based on the product similarity between the product information features and each historical product feature, the product information features and each historical product feature are compared separately. The historical product feature corresponding to the maximum product similarity is most similar to the current product information feature, so only the maximum product similarity needs to be judged. If the maximum product similarity is less than a first similarity threshold, it indicates that the returned product information in the order screenshot is original, meaning that there may be a situation such as the product being returned by a consumer for the first time, and it can be directly determined that the order screenshot has not been tampered with. If the maximum product similarity is not less than the first similarity threshold, further in-depth judgment based on combined information features is required.

[0048] In this embodiment, the product similarity between the product information feature and each historical product feature can be any reasonable similarity metric, such as the cosine similarity or Euclidean distance between the product information feature and each historical product feature. The first similarity threshold can be flexibly set and adjusted according to the actual situation and specific needs, and the scope of protection of this application is not limited by the specific value of the first similarity threshold.

[0049] In one embodiment, determining whether an order screenshot has been tampered with based on combined information features includes: obtaining the combined similarity between the combined information features and each historical combined feature, and determining the maximum combined similarity among the combined similarities, wherein the historical combined features are the features corresponding to the returned goods information and the returned merchant information in historical order screenshots that have not been tampered with in the past; determining whether the maximum combined similarity is greater than a second similarity threshold, if so, determining that the order screenshot has not been tampered with, if not, determining that the order screenshot has been tampered with.

[0050] In this embodiment, if the maximum product similarity is not less than the first similarity threshold, it indicates that the returned product corresponding to the order screenshot has a history of similar return transactions. Whether the order screenshot generated by the current return transaction has been tampered with requires further judgment based on the combined information features. Specifically, firstly, based on the combined information features and the combined similarity between each historical combined feature, the combined information features are compared with each historical combined feature. The historical combined feature corresponding to the maximum combined similarity is most similar to the current combined information feature, so only the maximum combined similarity needs to be judged. If the maximum combined similarity is greater than the second similarity threshold, it indicates that the returned product information in the order screenshot has extremely high similarity, and the returning merchant information also has extremely high similarity. That is, the order screenshot is for the same product from the same merchant, so it is determined that the order screenshot has not been tampered with. If the maximum combined similarity is not greater than the second similarity threshold, it indicates that the returned product information in the order screenshot has extremely high similarity, but the returning merchant information has low similarity. That is, the merchant's return information in the order screenshot is very likely to have been tampered with, so it is determined that the order screenshot has been tampered with.

[0051] In this embodiment, the product similarity between the combined information feature and each historical combined feature can be any reasonable similarity metric, such as the cosine similarity or Euclidean distance between the combined information feature and each historical combined feature. The second similarity threshold can be flexibly set and adjusted according to the actual situation and specific needs, and the scope of protection of this application is not limited by the specific value of the second similarity threshold.

[0052] In one embodiment, the order screenshot is included in a return request submitted in real time by the consumer. After comparing key information features with historical image features to determine whether the order screenshot has been tampered with, the process further includes: if it is determined that the order screenshot has not been tampered with, then allowing the return request; if it is determined that the order screenshot has been tampered with, then blocking the return request.

[0053] In this embodiment, the order screenshot Figure 1 Generally, return requests are submitted in real time by consumers. After determining whether the order screenshot has been tampered with, the return request can be processed in real time and automatically based on the specific circumstances of whether the order screenshot has been tampered with. That is, if it is determined that the order screenshot has not been tampered with, the return request is allowed; if it is determined that the order screenshot has been tampered with, the return request is blocked. This builds a real-time defense against return requests, reduces the cost of return operations, maintains the fairness of return operations, and improves the experience of normal consumers.

[0054] In one embodiment, based on the specific implementation process provided in the above embodiments, the key information features obtained from order screenshots are obtained using the same feature acquisition method as the historical image features obtained from historical order screenshots.

[0055] In this embodiment, order screenshots and historical order screenshots are processed using the same method. For example, when extracting the area containing key information from an order screenshot corresponding to an e-commerce return order, a deep learning detection model is used to detect and extract the key information area, obtaining the key information image. Then, a deep learning extraction model is used to extract key information features from the key information image. Similarly, for historical order screenshots, the same deep learning detection model is used to detect and extract the historical key information area, obtaining the historical key information image. Then, a deep learning extraction model is used to extract historical image features from the historical key information image. Using the same processing method for order screenshots and historical order screenshots improves the fairness and consistency of comparing key information features and historical image features, thereby further improving the accuracy of identifying whether order screenshots have been tampered with.

[0056] Optionally, the historical image features obtained from multiple tamper-free historical order screenshots can be pre-extracted and stored in any reasonable database such as the Facebook AI Similarity Search database or the Milvus feature database, thereby facilitating the subsequent reading and searching of historical image features and accelerating the efficiency and speed of identifying whether order screenshots have been tampered with.

[0057] In one specific embodiment, the order screenshot tampering detection method provided in the above embodiments can accurately identify whether the order screenshot has been tampered with, and thus proactively allow or block related return requests. For example, for... Figure 7 The example image shown illustrates order screenshot tampering. The left side is an untampered historical order screenshot, while the right side is an order screenshot corresponding to a real-time return request initiated by a malicious consumer with illegally altered merchant information. The returned product information in both is highly similar (both are the first product information), but the combined return information based on the merchant and product information has low similarity (the first product information in the historical order screenshot and the second product information in the current order screenshot). Therefore, based on the aforementioned order screenshot tampering detection method, it can be accurately determined that the order screenshot corresponding to the return request has been tampered with, thus blocking the return request. Regarding... Figure 8 The order screenshots shown are tamper-proof examples. The left side is an unaltered historical order screenshot, and the right side is an order screenshot corresponding to a real-time return request initiated by a normal consumer. The returned product information in both is highly similar (both are the first product information), and the combined return information based on the returning merchant information and returned product information is also highly similar (both are the first product information and the first merchant information). Therefore, based on the order screenshot tampering detection method described above, it can be accurately determined that the order screenshot corresponding to the return request has not been tampered with, and this return request is for the same product from the same merchant.

[0058] In this embodiment, the key information area is extracted from the order screenshot corresponding to an e-commerce return order to obtain at least one key information image. The key information is related to the normal flow of the return business. Feature extraction is performed based on the key information image to obtain key information features. The key information features are compared with historical image features to determine whether the order screenshot has been tampered with. The historical image features are the features corresponding to the key information in historical order screenshots that have not been tampered with. In the above solution, given the special nature of order screenshots, instead of directly judging whether the order screenshot has been tampered with based on the global features of the order screenshot, it selectively extracts key information images from the key information area and performs targeted feature extraction on some but more critical key information images. This avoids interference from irrelevant information in the feature comparison process, thereby improving the accuracy of identifying whether the order screenshot has been tampered with based on the comparison of key information features and historical image features.

[0059] Exemplary device Accordingly, embodiments of this application also provide a detection device for order screenshot tampering, such as... Figure 9 As shown, the device may include: Image processing module 901 is used to extract the area containing key information from the order screenshot corresponding to the e-commerce return order and obtain at least one key information image, wherein the key information is information related to the normal flow of the return business; Feature processing module 902 is used to extract features based on key information images to obtain key information features; The judgment module 903 is used to compare the key information features with the historical image features to determine whether the order screenshot has been tampered with. The historical image features are the features corresponding to the key information in historical order screenshots that have not been tampered with in the past.

[0060] In one embodiment, the key information images include product information images and merchant information images; Image processing module 901 is used to extract the returned product information from the first area of ​​the order screenshot corresponding to the e-commerce return order to obtain a product information image, and to extract the returned merchant information from the second area to obtain a merchant information image.

[0061] In one embodiment, the feature processing module 902 is used to extract features from the product information image to obtain product information features; to stitch the product information image and the merchant information image together to obtain a combined information image; and to extract features from the combined information image to obtain combined information features; wherein, the key information features include product information features and combined information features.

[0062] In one embodiment, the judgment module 903 is used to obtain the product similarity between the product information features and each historical product feature, and determine the maximum product similarity among the product similarities, wherein the historical product features are the features corresponding to the returned product information in the historical order screenshots that have not been tampered with in the past; it is judged whether the maximum product similarity is less than the first similarity threshold. If it is, it is determined that the order screenshot has not been tampered with. If not, it is judged whether the order screenshot has been tampered with based on the combined information features.

[0063] In one embodiment, the judgment module 903 is used to obtain the combination similarity between the combined information features and each historical combination feature, and to determine the maximum combination similarity among the various combination similarities, wherein the historical combination features are the features corresponding to the returned goods information and the returned merchant information in the historical order screenshots that have not been tampered with in the past; and to determine whether the maximum combination similarity is greater than the second similarity threshold. If it is, it is determined that the order screenshot has not been tampered with; if not, it is determined that the order screenshot has been tampered with.

[0064] In one embodiment, the order screenshot is included in a return request submitted in real time by the consumer user; The order screenshot tampering detection device also includes a request processing module, which compares key information features with historical image features to determine whether the order screenshot has been tampered with. If it is determined that the order screenshot has not been tampered with, the return request is allowed; if it is determined that the order screenshot has been tampered with, the return request is blocked.

[0065] In one embodiment, obtaining key information features based on order screenshots and obtaining historical image features based on historical order screenshots both employ the same feature acquisition method.

[0066] The order screenshot tampering detection device provided in this embodiment belongs to the same application concept as the order screenshot tampering detection method provided in the above embodiments of this application. It can execute the order screenshot tampering detection method provided in any of the above embodiments of this application and has the corresponding functional modules and beneficial effects of the execution method. Technical details not described in detail in this embodiment can be found in the specific processing content of the order screenshot tampering detection method provided in the above embodiments of this application, and will not be repeated here.

[0067] Exemplary electronic devices This application also provides an electronic device, such as... Figure 10 As shown, the electronic device includes a memory 1000 and a processor 1001.

[0068] The memory 1000 is connected to the processor 1001 and is used to store programs.

[0069] The processor 1001 is used to implement the order screenshot tampering detection method in the above embodiment by running the program stored in the memory 1000.

[0070] Specifically, the aforementioned electronic device may also include: a communication interface 1002, an input device 1003, an output device 1004, and a bus 1005.

[0071] The processor 1001, memory 1000, communication interface 1002, input device 1003, and output device 1004 are interconnected via a bus. Among them: Bus 1005 may include a pathway for transmitting information between various components of a computer system.

[0072] The processor 1001 can be a general-purpose processor, such as a general-purpose central processing unit (CPU), a microprocessor, etc., or an application-specific integrated circuit (ASIC), or one or more integrated circuits used to control the execution of the program of the present invention. It can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), an off-the-shelf programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0073] Processor 1001 may include a main processor, as well as a baseband chip, modem, etc.

[0074] The memory 1000 stores a program that executes the technical solution of this invention, and may also store an operating system and other key business functions. Specifically, the program may include program code, which includes computer operation instructions. More specifically, the memory 1000 may include read-only memory (ROM), other types of static storage devices capable of storing static information and instructions, random access memory (RAM), other types of dynamic storage devices capable of storing information and instructions, disk storage, flash memory, etc.

[0075] Input device 1003 may include a device for receiving data and information input by the user, such as a keyboard, mouse, camera, scanner, light pen, voice input device, touch screen, pedometer or gravity sensor, etc.

[0076] Output device 1004 may include devices that allow information to be output to a user, such as a display screen, printer, speaker, etc.

[0077] The communication interface 1002 may include a device that uses any transceiver to communicate with other devices or communication networks, such as Ethernet, Radio Access Network (RAN), Wireless Local Area Network (WLAN), etc.

[0078] The processor 1001 executes the program stored in the memory 1000 and calls other devices, which can be used to implement the various steps of the order screenshot tampering detection method provided in the above embodiments of this application.

[0079] Exemplary computer program products and storage media In addition to the methods and devices described above, embodiments of this application may also be computer program products, which include computer program instructions that, when executed by a processor, cause the processor to perform the steps in the order screenshot tampering detection method described in the embodiments of this application.

[0080] The computer program product can be written in any combination of one or more programming languages ​​to perform the operations of the embodiments of this application. The programming languages ​​include object-oriented programming languages ​​such as Java and C++, as well as conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the user's computing device, partially on the user's computing device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0081] Furthermore, embodiments of this application may also be storage media storing a computer program, which is executed by a processor in the steps of the order screenshot tampering detection method described in the embodiments of this application.

[0082] For the foregoing method embodiments, in order to simplify the description, they are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, because according to this application, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0083] It should be noted that the various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For apparatus embodiments, since they are basically similar to method embodiments, the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments.

[0084] The steps in the methods of the various embodiments of this application can be adjusted, merged, or deleted in order according to actual needs, and the technical features described in each embodiment can be replaced or combined.

[0085] The modules and sub-modules in the devices and terminals provided in the various embodiments of this application can be merged, divided, and deleted according to actual needs.

[0086] It should be understood that the disclosed terminals, devices, and methods can be implemented in other ways, given the several embodiments provided in this application. For example, the terminal embodiments described above are merely illustrative. For instance, the division of modules or sub-modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple sub-modules or modules may be combined or integrated into another module, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or modules, and may be electrical, mechanical, or other forms.

[0087] The modules or submodules described as separate components may or may not be physically separate. The components that constitute a module or submodule may or may not be physical modules or submodules; that is, they may be located in one place or distributed across multiple network modules or submodules. Some or all of the modules or submodules can be selected to achieve the purpose of this embodiment's solution, depending on actual needs.

[0088] Furthermore, the functional modules or sub-modules in the various embodiments of this application can be integrated into one processing module, or each module or sub-module can exist physically separately, or two or more modules or sub-modules can be integrated into one module. The integrated modules or sub-modules described above can be implemented in hardware or in the form of software functional modules or sub-modules.

[0089] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0090] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software unit executed by a processor, or a combination of both. The software unit can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0091] 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 one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0092] 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 detecting order screenshot tampering, characterized in that, include: In the order screenshot corresponding to the e-commerce return order, extract the area where the key information is located to obtain at least one key information image, wherein the key information is information related to the normal flow of the return business; Feature extraction is performed on the key information image to obtain key information features; The key information features are compared with historical image features to determine whether the order screenshot has been tampered with. The historical image features are the features corresponding to the key information in historical order screenshots that have not been tampered with.

2. The method for detecting order screenshot tampering according to claim 1, characterized in that, The key information images include product information images and merchant information images; The process involves extracting the area containing key information from the order screenshot corresponding to the e-commerce return order, obtaining at least one key information image, including: In the order screenshot corresponding to the e-commerce return order, the returned product information is extracted from the first area to obtain the product information image, and the returned merchant information is extracted from the second area to obtain the merchant information image.

3. The method for detecting order screenshot tampering according to claim 2, characterized in that, The step of extracting features from the key information image to obtain key information features includes: Feature extraction is performed on the product information image to obtain product information features; The product information image and the merchant information image are combined to obtain a composite information image; Feature extraction is performed on the combined information image to obtain the combined information features; The key information features include the product information features and the combined information features.

4. The method for detecting order screenshot tampering according to claim 3, characterized in that, The step of comparing the key information features with historical image features to determine whether the order screenshot has been tampered with includes: Obtain the product similarity between the product information features and each historical product feature, and determine the maximum product similarity among the product similarities, wherein the historical product features are the features corresponding to the returned product information in the historical order screenshots that have not been tampered with in history; If the maximum product similarity is less than the first similarity threshold, it is determined that the order screenshot has not been tampered with; otherwise, it is determined whether the order screenshot has been tampered with based on the combined information features.

5. The method for detecting order screenshot tampering according to claim 4, characterized in that, The step of determining whether the order screenshot has been tampered with based on the combined information features includes: Obtain the combination similarity between the combined information features and each historical combination feature, and determine the maximum combination similarity among the various combination similarities, wherein the historical combination features are the features corresponding to the returned goods information and the returned merchant information in the historical order screenshots that have not been tampered with in history; Determine whether the maximum combination similarity is greater than the second similarity threshold. If yes, it is determined that the order screenshot has not been tampered with; otherwise, it is determined that the order screenshot has been tampered with.

6. The method for detecting order screenshot tampering according to claim 1, characterized in that, The order screenshot includes the return request submitted by the consumer user in real time; After comparing the key information features with historical image features to determine whether the order screenshot has been tampered with, the method further includes: If it is determined that the order screenshot has not been tampered with, the return request will be approved; If it is determined that the order screenshot has been tampered with, the return request will be blocked.

7. The method for detecting order screenshot tampering according to any one of claims 1-6, characterized in that, Also includes: The key information features obtained from the order screenshot are obtained using the same feature acquisition method as the historical image features obtained from the historical order screenshot.

8. A detection device for order screenshot tampering, characterized in that, include: The image processing module is used to extract the area containing key information from the order screenshot corresponding to the e-commerce return order and obtain at least one key information image, wherein the key information is information related to the normal flow of the return business; The feature processing module is used to extract features based on the key information image to obtain key information features; The judgment module is used to compare the key information features with historical image features to determine whether the order screenshot has been tampered with, wherein the historical image features are the features corresponding to the key information in historical order screenshots that have not been tampered with in the past.

9. An electronic device, characterized in that, include: Memory and processor; The memory is connected to the processor and is used to store programs; The processor is used to implement the order screenshot tampering detection method as described in any one of claims 1-7 by running the program in the memory.

10. A storage medium, characterized in that, The storage medium stores a computer program, which, when executed by a processor, implements the method for detecting order screenshot tampering as described in any one of claims 1-7.

Images