Image search anomaly analysis method and device, electronic equipment and storage medium

Abstract

The invention relates to the technical field of computers, and provides an image search anomaly analysis method and device, electronic equipment and a storage medium. The method comprises the steps of obtaining a search term, a first description text and a second description text of a target image; comparing the search word with the first description text and the second description text to obtain a search comparison result; if the search comparison result indicates that the first description text does not contain the search word and the second description text contains the search word, determining that the processing parameter corresponding to the first description text is abnormal; and if the search comparison result indicates that the first description text contains the search word and the second description text does not contain the search word, determining that the processing parameter corresponding to the second description text is abnormal. According to the method, analysis factors are enriched, the analysis strength is enhanced, and the reliability of anomaly analysis is enhanced, so that the accuracy of image search anomaly analysis is improved.

Description

Technical Field

[0001] The present disclosure relates to the field of computer technologies, and in particular, to an image search anomaly analysis method, apparatus, electronic device, and storage medium. Background Art

[0002] With the development of computer technologies and image processing technologies, image search applications are used in many application scenarios. Among them, there are abnormal situations in the results of image search. At present, for the analysis method of image search anomalies, generally, the accuracy of the model, etc. is analyzed. This method has problems such as single analysis factors and small analysis intensity, resulting in low accuracy of image search anomaly analysis. Summary of the Invention

[0003] In view of this, the purpose of the present disclosure is to provide an image search anomaly analysis method, apparatus, electronic device, and storage medium. By comparing the search term, the first description text, and the second description text of the target image with the reference documents, the processing parameters with anomalies in the processing parameters corresponding to the first description text and the second description text are determined, enriching the analysis factors, enhancing the analysis intensity, enhancing the reliability of anomaly analysis, and thus improving the accuracy of image search anomaly analysis.

[0004] In a first aspect, an embodiment of the present disclosure provides an image search anomaly analysis method, and the image search anomaly analysis method includes:

[0005] Obtain the search term, the first description text, and the second description text of the target image, where the target image is an image with search anomalies, and the first description text and the second description text are used to describe specific information in the target image;

[0006] Compare the search term with the first description text and the second description text to obtain a search comparison result;

[0007] If the search comparison result indicates that the search term is not included in the first description text and the search term is included in the second description text, determine that the processing parameters corresponding to the first description text are abnormal;

[0008] If the search comparison result indicates that the search term is included in the first description text and the search term is not included in the second description text, determine that the processing parameters corresponding to the second description text are abnormal.

[0009] In a second aspect, an embodiment of the present disclosure provides an image search anomaly analysis apparatus, and the image search anomaly analysis apparatus includes:

[0010] A first acquisition module, configured to acquire a search term, a first description text, and a second description text of a target image, where the target image is an image with a search anomaly, and the first description text and the second description text are used to describe specific information in the target image;

[0011] A first comparison module, configured to compare the search term with the first description text and the second description text to obtain a search comparison result;

[0012] A first determination module, configured to determine that there is an anomaly in the processing parameter corresponding to the first description text if the search comparison result indicates that the first description text does not contain the search term and the second description text contains the search term;

[0013] A second determination module, configured to determine that there is an anomaly in the processing parameter corresponding to the second description text if the search comparison result indicates that the first description text contains the search term and the second description text does not contain the search term.

[0014] In a third aspect, an embodiment of the present disclosure provides an electronic device, including a processor and a memory, where the memory stores machine-executable instructions that can be executed by the processor, and the processor executes the machine-executable instructions to implement the above image search anomaly analysis method.

[0015] In a fourth aspect, an embodiment of the present disclosure provides a computer-readable storage medium, where the computer-readable storage medium stores computer-executable instructions, and when the computer-executable instructions are called and executed by a processor, the computer-executable instructions cause the processor to implement the above image search anomaly analysis method.

[0016] The embodiments of the present disclosure bring the following beneficial effects:

[0017] The above-mentioned method, device, electronic device and storage medium for analyzing image search anomalies obtain a search term, a first description text and a second description text of a target image, where the target image is an image with search anomalies, and the first description text and the second description text are used to describe specific information in the target image; compare the search term with the first description text and the second description text to obtain a search comparison result; if the search comparison result indicates that the first description text does not contain the search term and the second description text contains the search term, determine that the processing parameter corresponding to the first description text is abnormal; if the search comparison result indicates that the first description text contains the search term and the second description text does not contain the search term, determine that the processing parameter corresponding to the second description text is abnormal. In this method, by comparing the search term, the first description text and the second description text of the target image to determine the abnormal processing parameter among the processing parameters corresponding to the first description text and the second description text, the analysis factors are enriched, the analysis strength is enhanced, the reliability of the anomaly analysis is enhanced, and thus the accuracy of the image search anomaly analysis is improved.

[0018] Other features and advantages of the present disclosure will be described in the following specification, and in part, will be obvious from the specification, or will be understood by implementing the present disclosure. The objectives and other advantages of the present disclosure are achieved and obtained by the structures particularly pointed out in the specification, the claims and the drawings.

[0019] To make the above objectives, features and advantages of the present disclosure more obvious and understandable, the following specifically enumerates preferred embodiments and, in conjunction with the accompanying drawings, the detailed description is as follows. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] To more clearly illustrate the specific embodiments of the present disclosure or the technical solutions in the prior art, the following will briefly introduce the drawings required for use in the description of the specific embodiments or the prior art. Obviously, the following drawings are some embodiments of the present disclosure. For those skilled in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0021] Figure 1 It is a schematic diagram of an embodiment of the method for analyzing image search anomalies provided by an embodiment of the present disclosure;

[0022] Figure 2 It is a schematic diagram of another embodiment of the method for analyzing image search anomalies provided by an embodiment of the present disclosure;

[0023] Figure 3 It is a schematic diagram of an image search anomaly analysis device provided by an embodiment of the present disclosure;

[0024] Figure 4 Schematic diagram of an electronic device provided by an embodiment of the present disclosure. Specific implementation manners

[0025] To make the objectives, technical solutions, and advantages of this embodiment clearer, the technical solutions of the present disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are some but not all of the embodiments of the present disclosure. Based on the embodiments in the present disclosure, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the scope of protection of the present disclosure.

[0026] This embodiment provides an image search anomaly analysis method, device, electronic device, and storage medium. It can be applied to image search anomaly analysis scenarios in any field, especially in image search anomaly analysis scenarios in the field of autonomous driving.

[0027] In an image search anomaly analysis method in one embodiment of the present disclosure, it can run on a terminal device or a server. Among them, the terminal device can be a local terminal device. When the image search anomaly analysis method runs on the server, the method can be implemented and executed based on a cloud interaction system, where the cloud interaction system includes a server and a client device.

[0028] For ease of understanding, the specific process of this embodiment will be described below. Please refer to Figure 1 In an embodiment of the image search anomaly analysis method in this embodiment, the following steps are included:

[0029] Step 101: Obtain a search term, a first description text, and a second description text of a target image, where the target image is an image with a search anomaly, and the first description text and the second description text are used to describe specific information in the target image;

[0030] Among them, by way of example rather than limitation, the first description text and the second description text may be texts without a front-back logical relationship. For example, the first description text and the second description text are texts obtained by different tools or algorithms respectively, where the tool or algorithm is a tool or algorithm for processing an image and generating text. There is a front-back logical relationship between the first description text and the second description text. For example, the second description text is the text after optimizing the first description text. By way of example rather than limitation, the specific information may be feature information corresponding to scene elements corresponding to the target image, or may also be feature information corresponding to a preset requirement for generating text.

[0031] Step 102: Compare the search term with the first description text and the second description text to obtain a search comparison result;

[0032] Among them, by way of example and not limitation, the search term can be compared with the first description text and the search term can be compared with the second description text to obtain a search comparison result. Among them, when making the comparison, multiple comparison algorithms can be combined for multi-level analysis or the comparison can be performed through a model with a multi-layer structure combination to improve the accuracy of the comparison, thereby improving the accuracy of anomaly analysis. For example, it can be: Based on a preset stemming algorithm, perform stemming processing on the search term, the first description text, and the second description text respectively to obtain an initial word, a first initial text, and a second initial text; Based on a preset lemmatization algorithm, perform stemming processing on the initial word, the first initial text, and the second initial text respectively to obtain a target word, a first target text, and a second target text; Based on a preset matching algorithm, match the target word with the first target text, and based on the preset matching algorithm, match the target word with the second target text to obtain a search comparison result.

[0033] In one implementation, after comparing the search term with the first description text and the second description text to obtain a search comparison result, it can also be: If the search comparison result indicates that the search term is not included in the first description text and the search term is not included in the second description text, then determine that the processing parameter corresponding to the first description text is abnormal; Obtain the first similarity between the search term and the first description text, the second similarity between the search term and the second description text, and the difference information between the first description text and the second description text; If the first similarity is greater than the second similarity and the difference information indicates that there is a difference between the second description text and the first description text, then determine that the processing parameter corresponding to the second description text is abnormal.

[0034] Among them, by way of example and not limitation, the difference information may include but is not limited to content differences and structural differences. Among them, content differences such as, for example, new words, deleted words, and word expressions, and structural differences such as, for example, information hierarchy, sentence structure, information presentation method, and language style.

[0035] Among them, by way of example and not limitation, if the first similarity is less than or equal to the second similarity and the difference information indicates that there is a difference between the second description text and the first description text, then determine that the processing parameter corresponding to the first description text is abnormal.

[0036] By determining whether the processing parameter corresponding to the second description text is abnormal according to the search comparison result, the similarities corresponding to the search term, the first description text, and the second description text respectively, and the difference information between the first description text and the second description text, multi-angle and multi-level analysis is realized, enriching the analysis factors, enhancing the analysis strength, enhancing the reliability of anomaly analysis, and greatly improving the accuracy and adaptability of image search anomaly analysis.

[0037] Step 103, if the search comparison result indicates that the first description text does not contain the search term and the second description text contains the search term, it is determined that there is an abnormality in the processing parameter corresponding to the first description text;

[0038] Among them, by way of example and not limitation, the fact that the first description text does not contain the search term can be understood as: the first description text does not contain words with a similarity greater than the first preset value to the search term. The fact that the second description text contains the search term can be understood as: the second description text contains words with a similarity greater than the second preset value to the search term. By way of example and not limitation, the processing parameter can be understood as: a tool or rule or strategy for processing the target image and generating text. For example, a model for generating a description text of an image, a prompt word for prompting the model to generate a description text, and a cleaning rule for secondary processing of the description text generated by the model.

[0039] Among them, by way of example and not limitation, the processing parameter corresponding to the first description text with an abnormality includes at least one parameter. If there are more than one processing parameter corresponding to the first description text with an abnormality, further analysis needs to be carried out on each parameter.

[0040] Step 104, if the search comparison result indicates that the first description text contains the search term and the second description text does not contain the search term, it is determined that there is an abnormality in the processing parameter corresponding to the second description text.

[0041] Among them, by way of example and not limitation, the processing parameter corresponding to the second description text with an abnormality includes at least one parameter. If there are more than one processing parameter corresponding to the second description text with an abnormality, further analysis needs to be carried out on each parameter. Each parameter can be analyzed one by one, or the parameters can be combined for analysis.

[0042] The above image search anomaly analysis method compares the search term, the first description text, and the second description text of the target image to determine the processing parameter with an abnormality among the processing parameters corresponding to the first description text and the second description text, enriching the analysis factors, enhancing the analysis strength, enhancing the reliability of the anomaly analysis, and thus improving the accuracy of the image search anomaly analysis.

[0043] Please refer to Figure 2 , another embodiment of the method for determining the image processing parameter in this embodiment includes:

[0044] Step 201, generate the first description text of the target image based on the preset first processing parameter and second processing parameter;

[0045] Among them, by way of example and not limitation, determine the category of the target image, where the category of the target image is the scene category in the field corresponding to the target image; obtain the first processing parameter, the second processing parameter, and the third processing parameter corresponding to the category of the target image from the preset parameter matching library; identify the target image through the first processing parameter and perform text conversion to obtain the initial text, and adjust the initial text based on the second processing parameter to obtain the first description text of the target image, or process the target image through the first processing parameter based on the second processing parameter to obtain the first description text.

[0046] Step 202, optimize the first description text based on the preset third processing parameter to obtain the second description text; wherein, the first processing parameter, the second processing parameter, and the third processing parameter are parameters of different types;

[0047] Among them, by way of example and not limitation, the third processing parameter can be a cleaning rule for the quality, consistency, and accuracy of the data in the first description text. The third processing parameter includes, but is not limited to, rules for duplicate removal, standardization, semantic expansion, grammar and sentence pattern optimization, and sentiment analysis and adjustment. The third processing parameter is a cleaning rule that has been iteratively adjusted and optimized in advance according to test results and user feedback; the third processing parameter includes at least one parameter. For example, the third processing parameter includes a cleaning rule and an audit policy, where the audit policy is used to audit the processing results of the cleaning rule.

[0048] Generating the first description text of the target image through the first processing parameter and the second processing parameter, and optimizing the first description text through the third processing parameter enhances the emotional expression of the second description text, reduces redundant information, and improves the readability, fluency, accuracy, and consistency of the second description text.

[0049] Step 203, obtain the search term, the first description text, and the second description text of the target image, where the target image is an image with a search anomaly, and the first description text and the second description text are used to describe specific information in the target image;

[0050] Among them, by way of example and not limitation, the search term can be a keyword input by the user. For example, when obtaining the search term for the target image, it is possible to: in response to an input instruction, obtain the input information, and preprocess the input information (including but not limited to format conversion, semantic conversion, semantic expansion, and vocabulary splicing and combination) to obtain the search term; the search term can also be a retrieval term corresponding to the scene elements matched based on the keyword input by the user. For example, analyze the scene category of the keyword input by the user, and match the retrieval term corresponding to the scene elements of the scene category from the retrieval term library; the search term can also be a word formed by combining the main objects extracted from the target image. For example, based on a preset image processing algorithm (model), extract at least one main object (key element) from the target image, and combine the at least one main object to obtain the search term for the target image.

[0051] Step 204: Compare the search term with the first description text and the second description text to obtain a search comparison result.

[0052] Among them, comparing the search term with the first description text and the second description text to obtain a search comparison result can be: comparing the search term with the first description text and comparing the second description text to obtain a search comparison result. Among them, the search comparison result includes the comparison result corresponding to the first description text and the comparison result corresponding to the second description text.

[0053] By way of example and not limitation, when comparing the search term with the first description text, it is possible to: based on a first preset algorithm, calculate a first similarity between the search term and the first description text; based on the first preset algorithm, calculate a second similarity between the search term and the first description text; if the first similarity is greater than a second preset value and the second similarity is greater than a third preset value, it is determined that the first description text contains the search term; if the first similarity is less than or equal to the second preset value and / or the second similarity is less than or equal to the third preset value, it is determined that the first description text does not contain the search term. Among them, the execution process of comparing the second description text is similar to the execution process of comparing the search term with the first description text, and will not be elaborated here.

[0054] Step 205: If the search comparison result indicates that the first description text does not contain the search term and the second description text contains the search term, determine that the first processing parameter and / or the second processing parameter is abnormal.

[0055] Among them, by way of example and not limitation, if the first processing parameter is a model, the situations where the first processing parameter is abnormal include but are not limited to the following situations: First, the first processing parameter misidentifies the target object in the target image as another object. For example, misidentifies a bus as another object and describes it as another object in the first description text; Second, the information in the image is incorrect, but the information expression in the description text is correct.

[0056] In one implementation, based on a preset fourth processing parameter and a search term, a target image is recognized to obtain a recognition result, where the fourth processing parameter is at least one other processing parameter of the same type as the first processing parameter; an audit result of the target image based on the search term is obtained; if both the recognition result and the audit result indicate that the object corresponding to the search term is included in the target image, then the target similarity between the second processing parameter and the recognition result and the audit result is calculated; if the target similarity is greater than a preset similarity, it is determined that the first processing parameter is abnormal.

[0057] Among them, the fourth processing parameter is a model for processing an image and generating text. The fourth processing parameter can be, for example, a pre-trained model based on contrastive language-image pairs (Contrastive Language-Image Pre-training, CLIP) model; when the fourth processing parameter is more than one other processing parameter of the same type as the first processing parameter, first combine the processing parameters in the fourth processing parameter to obtain a target processing parameter to form a new network structure. Among them, the way of combining the processing parameters in the fourth processing parameter can be a series connection method (connecting multiple models in sequence to form an integrated network structure), a parallel connection method (executing different models simultaneously and combining the outputs of each model), an interaction method (combining different models in an interaction form to achieve mutual influence and information exchange between models, where there will be innovation in the model architecture and integration of modules), or a multi-scale fusion method (simultaneously capturing the details and global context information of the target image based on different scale feature information). Based on the target processing parameter and the search term, the target image is recognized to obtain a recognition result, where the recognition result indicates that the target object corresponding to the search term is included in the target image, or the target object corresponding to the search term is not included in the target image. This can greatly improve the accuracy and reliability of the recognition result.

[0058] Among them, by way of example rather than limitation, the audit result can be an audit result obtained by artificial intelligence audit or a result feedback by manual audit; the audit result is a result obtained by auditing the accuracy and reliability of the recognition result; the target similarity can be a weighted mean or an arithmetic mean of a third similarity (the similarity between the second processing parameter and the recognition result) and a fourth similarity (the target similarity between the second processing parameter and the audit result).

[0059] Determining whether the first processing parameter is abnormal through the audit result of the target image based on the search term and the recognition result obtained by recognizing the target image according to the fourth processing parameter enhances the robustness and improves the accuracy and reliability of the abnormal analysis of the processing parameter.

[0060] Step 206, if the search and comparison result indicates that the search term is included in the first description text and not included in the second description text, it is determined that the third processing parameter is abnormal;

[0061] Among them, the third processing parameter includes at least one parameter; when the number of parameters in the third processing parameter is more than one, further analysis needs to be performed on each parameter. Each parameter can be analyzed one by one, or the parameters can be combined for analysis.

[0062] Determining the specific processing parameter with an abnormality in the first description text or the second description text through the search and comparison result optimizes the processing flow, enhances the analysis strength, improves the reliability of the abnormality analysis, and thus improves the accuracy of the image search abnormality analysis.

[0063] Step 207, if the search and comparison result indicates that the search term is not included in the first description text and not included in the second description text, obtain the target recognition result, where the target recognition result is the result obtained by recognizing the target image based on other processing parameters and the search term, and the other processing parameters are parameters other than the processing parameters corresponding to the first description text and the second description text;

[0064] Among them, by way of example and not limitation, the target recognition result can be the result obtained by a computer or the result of manual recognition feedback; the target recognition result can be the result obtained by fusing the results obtained by recognizing the target image based on at least two other processing parameters respectively. For example, at least two other processing parameters include a first other parameter and a second other parameter. The target image is recognized based on the search term through the first other parameter to obtain a first recognition result, the target image is recognized based on the search term through the second other parameter to obtain a second recognition result, and the first recognition result and the second recognition result are fused based on a preset attention mechanism to obtain the target recognition result.

[0065] Step 208, compare and analyze the target recognition result with the first description text and the second description text respectively to obtain an analysis result;

[0066] Among them, by way of example and not limitation, it is possible to: calculate the similarity between the target recognition result and the first description text, and determine whether the similarity is greater than a first preset similarity threshold. If not, it is determined that the first description text does not include the target recognition result; calculate the similarity between the target recognition result and the second description text, and determine whether the similarity is greater than a second preset similarity threshold. If not, it is determined that the second description text does not include the target recognition result, so as to obtain an analysis result, where the analysis result is used to indicate whether the first description text includes the target recognition result and whether the second description text includes the target recognition result.

[0067] Step 209: Determine whether there are any abnormalities in the processing parameters corresponding to the first description text and / or the second description text according to the analysis results.

[0068] By performing anomaly analysis through the comparison result between the target recognition result obtained based on the parameters other than the processing parameters corresponding to the first description text and the second description text and the first description text and the second description text, the comprehensiveness of anomaly analysis is improved, and the accuracy of image search anomaly analysis is greatly enhanced.

[0069] In one implementation, if the analysis result indicates that the target recognition result is not included in the first description text and the target recognition result is not included in the second description text, it is determined that there is an abnormality in the processing parameters corresponding to the first description text, and there may be an abnormality in the processing parameters corresponding to the second description text.

[0070] Among them, by way of example rather than limitation, if the abnormal processing parameter corresponding to the first description text is a model, the situation where there is an abnormality in the processing parameter corresponding to the first description text may be: misidentifying other objects as the target object, that is, there is no target object in the image, but there is a target object in the description text. For example, there is no bus in the image, but the description text states that there is a bus.

[0071] By determining whether there are any abnormalities in the processing parameters corresponding to the first description text and / or the second description text according to the analysis result that the target recognition result is not included in the first description text and the target recognition result is not included in the second description text, the pertinence, efficiency, and accuracy of anomaly analysis are improved, and the stability of the image search anomaly analysis system is promoted.

[0072] The above image search anomaly analysis method compares the search terms, the first description text, and the second description text of the target image with the comparison file to determine the processing parameters with abnormalities among the processing parameters corresponding to the first description text and the second description text, enriching the analysis factors, enhancing the analysis strength, enhancing the reliability of anomaly analysis, and thus improving the accuracy of image search anomaly analysis.

[0073] Corresponding to the above method embodiments, refer to Figure 3 the schematic diagram of an image search anomaly analysis device shown in

[0074] The first acquisition module 301 is configured to acquire the search term, the first description text, and the second description text of the target image, where the target image is an image with search anomalies, and the first description text and the second description text are used to describe specific information in the target image;

[0075] The first comparison module 302 is configured to compare the search term with the first description text and the second description text to obtain a search comparison result;

[0076] The first determination module 303 is configured to determine that there is an abnormality in the processing parameter corresponding to the first description text if the search comparison result indicates that the search term is not included in the first description text and the search term is included in the second description text;

[0077] The second determination module 304 is configured to determine that there is an abnormality in the processing parameter corresponding to the second description text if the search comparison result indicates that the search term is included in the first description text and the search term is not included in the second description text.

[0078] The above image search anomaly analysis device compares the target image with the search term, the first description text and the second description text to determine the abnormal processing parameter among the processing parameters corresponding to the first description text and the second description text, enriching the analysis factors, enhancing the analysis strength, enhancing the reliability of the anomaly analysis, and thus improving the accuracy of the image search anomaly analysis.

[0079] Optionally, the image search anomaly analysis device further includes:

[0080] The generation module 305 is configured to generate a first description text of the target image based on a preset first processing parameter and a second processing parameter;

[0081] The optimization module 306 is configured to optimize the first description text based on a preset third processing parameter to obtain a second description text; wherein, the first processing parameter, the second processing parameter and the third processing parameter are parameters of different types.

[0082] Optionally, the first determination module 303 can also be used for:

[0083] If the search comparison result indicates that the search term is not included in the first description text and the search term is included in the second description text, determine that the first processing parameter and / or the second processing parameter is abnormal;

[0084] The step of determining that there is an abnormality in the processing parameter corresponding to the second description text if the search comparison result indicates that the search term is included in the first description text and the search term is not included in the second description text includes:

[0085] If the search comparison result indicates that the search term is included in the first description text and the search term is not included in the second description text, determine that the third processing parameter is abnormal.

[0086] Optionally, the first determination module 303 can also be used for:

[0087] Based on a preset fourth processing parameter, identify the target image based on the search term to obtain an identification result, where the fourth processing parameter is at least one other processing parameter of the same type as the first processing parameter;

[0088] Obtain the review result of the target image based on the search term;

[0089] If both the recognition result and the review result indicate that the object corresponding to the search term is included in the target image, then calculate the target similarity between the second processing parameter and the recognition result and the review result;

[0090] If the target similarity is greater than the preset similarity, it is determined that the first processing parameter is abnormal.

[0091] Optionally, the image search anomaly analysis device further includes:

[0092] A second acquisition module 307, configured to obtain a target recognition result if the search comparison result indicates that the search term is not included in the first description text and not included in the second description text, where the target recognition result is the result obtained by recognizing the target image based on other processing parameters and the search term, and the other processing parameters are parameters other than the processing parameters corresponding to the first description text and the second description text;

[0093] A second comparison module 308, configured to respectively compare and analyze the target recognition result with the first description text and the second description text to obtain an analysis result;

[0094] A determination module 309, configured to determine whether the processing parameters corresponding to the first description text and / or the second description text are abnormal according to the analysis result.

[0095] Optionally, the determination module 309 may further be configured to:

[0096] If the analysis result indicates that the target recognition result is not included in the first description text and the target recognition result is not included in the second description text, it is determined that the processing parameter corresponding to the first description text is abnormal, and the processing parameter corresponding to the second description text may be abnormal.

[0097] Optionally, the image search anomaly analysis device further includes:

[0098] A third determination module 310, configured to determine that the processing parameter corresponding to the first description text is abnormal if the search comparison result indicates that the search term is not included in the first description text and not included in the second description text;

[0099] A third acquisition module 311, configured to obtain the first similarity between the search term and the first description text, the second similarity between the search term and the second description text, and the difference information between the first description text and the second description text;

[0100] A fourth determination module 312, configured to determine that there is an abnormality in the processing parameter corresponding to the second description text if the first similarity is greater than the second similarity and the difference information indicates that there is a difference between the second description text and the first description text.

[0101] This embodiment further provides an electronic device, including a processor and a memory. The memory stores machine-executable instructions that can be executed by the processor, and the processor executes the machine-executable instructions to implement the above-mentioned image search anomaly analysis method. The electronic device can be a server or a terminal device.

[0102] See Figure 4 As shown, the electronic device includes a processor 400 and a memory 401. The memory 401 stores machine-executable instructions that can be executed by the processor 400, and the processor 400 executes the machine-executable instructions to implement the above-mentioned image search anomaly analysis method.

[0103] Furthermore, Figure 4 The electronic device shown further includes a bus 402 and a communication interface 403. The processor 400, the communication interface 403, and the memory 401 are connected through the bus 402.

[0104] Among them, the memory 401 may include a high-speed random access memory (RAM, Random Access Memory), and may also include a non-volatile memory, such as at least one disk memory. Through at least one communication interface 403 (which can be wired or wireless), a communication connection is established between this system network element and at least one other network element, and the Internet, wide area network, local area network, metropolitan area network, etc. can be used. The bus 402 can be an ISA bus, a PCI bus, an EISA bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For the sake of representation, Figure 4 only a bidirectional arrow is used in

[0105] The processor 400 may be an integrated circuit chip with the ability to process signals. In the implementation process, each step of the above method can be completed by the integrated logic circuit of the hardware in the processor 400 or the instructions in the form of software. The above-mentioned processor 400 may be a general-purpose processor, including a central processing unit (CPU for short), a network processor (NP for short), etc.; it may also be a digital signal processor (DSP for short), an application specific integrated circuit (ASIC for short), a field-programmable gate array (FPGA for short), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components. It can implement or execute the various methods, steps and logic block diagrams disclosed in this embodiment. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. The steps of the method disclosed in combination with this embodiment can be directly embodied as being executed and completed by the hardware decoding processor, or executed and completed by a combination of the hardware and software modules in the decoding processor. The software module may be located in a mature storage medium in the art such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. This storage medium is located in the memory 401, and the processor 400 reads the information in the memory 47001 and combines its hardware to complete the steps of the image search anomaly analysis method.

[0106] This embodiment also provides a computer-readable storage medium, which stores computer-executable instructions. When the computer-executable instructions are called and executed by the processor, the computer-executable instructions cause the processor to implement the steps of the above image search anomaly analysis method.

[0107] The computer program product of the image search anomaly analysis method, device, electronic device and storage medium provided in this embodiment includes a computer-readable storage medium storing program code. The instructions included in the program code can be used to execute the method described in the previous method embodiment. For the specific implementation, reference can be made to the method embodiment, which will not be elaborated here.

[0108] Those skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working processes of the above-described systems and devices can refer to the corresponding processes in the foregoing method embodiments, which will not be elaborated here.

[0109] In addition, in the description of this embodiment, unless otherwise clearly specified and defined, the terms "installed", "connected", and "coupled" shall be understood in a broad sense. For example, it may be a fixed connection, a detachable connection, or an integral connection; it may be a mechanical connection or an electrical connection; it may be directly connected or indirectly connected through an intermediate medium, and it may be the communication inside two components. For those skilled in the art, the specific meanings of the above terms in this disclosure can be understood according to specific circumstances.

[0110] If the above-mentioned functions are implemented in the form of software function units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of this disclosure, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this disclosure. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical discs that can store program codes.

[0111] In the description of this disclosure, it should be noted that the orientation or positional relationship indicated by the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc. is based on the orientation or positional relationship shown in the drawings. It is only for the convenience of describing this disclosure and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus should not be construed as a limitation to this disclosure. In addition, the terms "first", "second", and "third" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0112] Finally, it should be noted that the above embodiments are only specific implementation manners of this disclosure, used to illustrate the technical solutions of this disclosure, and are not limitations thereto. The protection scope of this disclosure is not limited thereto. Although this disclosure has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: any person skilled in the art within the technical scope disclosed by this disclosure can still modify the technical solutions recorded in the foregoing embodiments, or can easily think of changes, or perform equivalent replacements for some of the technical features; and these modifications, changes, or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of this embodiment, and should all be covered by the protection scope of this disclosure. Therefore, the protection scope of this disclosure should be subject to the protection scope of the claims.

Claims

1. An image search anomaly analysis method, characterized in that: The method comprises: Acquire a search term, a first description text, and a second description text of a target image, wherein the target image is an image with a search anomaly, and the first description text and the second description text are used to describe specific information in the target image; Compare the search term with the first description text and the second description text to obtain a search comparison result; If the search comparison result indicates that the first description text does not contain the search term, and the second description text contains the search term, determining that there is an abnormality in the processing parameter corresponding to the first description text; If the search comparison result indicates that the first description text contains the search term and the second description text does not contain the search term, it is determined that there is an abnormality in the processing parameter corresponding to the second description text.

2. The method according to claim 1, characterized in that Before the step of obtaining the search term, the first description text and the second description text of the target image, the method further includes: Generate a first description text of the target image based on the preset first processing parameter and the second processing parameter; Based on a preset third processing parameter, the first description text is optimized to obtain a second description text; wherein the first processing parameter, the second processing parameter and the third processing parameter are parameters of different types.

3. The method according to claim 2, characterized in that If the search comparison result indicates that the first description text does not contain the search term, and the second description text contains the search term, the step of determining that a processing parameter corresponding to the first description text is abnormal includes: If the search comparison result indicates that the first description text does not contain the search term, and the second description text contains the search term, determining that the first processing parameter and / or the second processing parameter is abnormal; If the search comparison result indicates that the first description text contains the search term and the second description text does not contain the search term, the step of determining that a processing parameter corresponding to the second description text is abnormal includes: If the search comparison result indicates that the first description text contains the search term and the second description text does not contain the search term, it is determined that there is an abnormality in the third processing parameter.

4. The method according to claim 3, characterized in that After the step of determining that the first processing parameter and / or the second processing parameter are abnormal if the search comparison result indicates that the first description text does not contain the search term and the second description text contains the search term, the method further includes: Recognize the target image based on the search word by using a preset fourth processing parameter to obtain a recognition result, wherein the fourth processing parameter is at least one other processing parameter of the same type as the first processing parameter; Obtaining a review result of the target image based on the search term; If both the recognition result and the audit result indicate that the target image contains an object corresponding to the search term, then calculating the target similarity between the second processing parameter and the recognition result and the audit result; If the target similarity is greater than a preset similarity, it is determined that the first processing parameter is abnormal.

5. The method according to claim 1, characterized in that After the step of comparing the search term with the first description text and the second description text to obtain the search comparison result, the method further includes: If the search comparison result indicates that the search word is not included in the first description text and the second description text does not include the search word, then obtaining a target recognition result, wherein the target recognition result is a result obtained by recognizing the target image based on other processing parameters and the search word, and the other processing parameters are parameters other than the processing parameters corresponding to the first description text and the second description text; Compare and analyze the target recognition result with the first description text and the second description text respectively to obtain an analysis result; It is determined whether there is an abnormality in processing parameters corresponding to the first description text and / or the second description text according to the analysis result.

6. The method according to claim 5, characterized in that The step of determining whether there is an abnormality in processing parameters corresponding to the first description text and / or the second description text according to the analysis result includes: If the analysis result indicates that the first description text does not contain the target recognition result and the second description text does not contain the target recognition result, it is determined that there is an abnormality in the processing parameters corresponding to the first description text and there may be an abnormality in the processing parameters corresponding to the second description text.

7. The method according to any one of claims 1 to 6, characterized in that: After the step of comparing the search term with the first description text and the second description text to obtain the search comparison result, the method further includes: If the search comparison result indicates that the first description text does not contain the search term, and the second description text does not contain the search term, determining that a processing parameter corresponding to the first description text is abnormal; Acquire a first similarity between the search term and the first description text, a second similarity between the search term and the second description text, and difference information between the first description text and the second description text; If the first similarity is greater than the second similarity, and the difference information indicates that there is a difference between the second description text and the first description text, it is determined that there is an abnormality in the processing parameter corresponding to the second description text.

8. An image search anomaly analysis device, characterized in that: The image search anomaly analysis device comprises: A first acquisition module is used to acquire a search term, a first description text, and a second description text of a target image, wherein the target image is an image with a search anomaly, and the first description text and the second description text are used to describe specific information in the target image; A first comparison module, used for comparing the search term with the first description text and the second description text to obtain a search comparison result; A first determination module, configured to determine that an abnormality exists in a processing parameter corresponding to the first description text if the search comparison result indicates that the first description text does not contain the search term and the second description text contains the search term; The second determination module is used to determine that there is an abnormality in the processing parameters corresponding to the second description text if the search comparison result indicates that the first description text contains the search term and the second description text does not contain the search term.

9. An electronic device, characterized in that: It comprises a processor and a memory, wherein the memory stores machine executable instructions that can be executed by the processor, and the processor executes the machine executable instructions to implement the image search anomaly analysis method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions. When the computer-executable instructions are called and executed by a processor, the computer-executable instructions prompt the processor to implement the image search anomaly analysis method according to any one of claims 1 to 7.

Images