Image testing method and device, equipment and storage medium

Abstract

The invention provides an image test method and device, equipment and a storage medium, and the method comprises the steps: obtaining test description information of a target interface in a game interface test scene, obtaining an interface picture of the target interface in a current to-be-tested version, generating task indication information corresponding to the game interface test scene, and inputting the test description information, the task indication information and the interface picture into a large language model, and executing a task corresponding to the test description information for the interface picture by the large language model based on the task indication information to obtain a test result of the target interface. According to the method and the device, a standard picture does not need to be used as a judgment reference, the check points in the interface picture can be checked only based on the test description information, and compared with a manual judgment mode, the efficiency is higher.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to an image testing method, device, equipment and storage medium. Background Art

[0002] Software version iteration will not only affect the operation of software logic, but also the UI interface. In some game applications, many interfaces in the game will reuse the underlying general interface. If the general interface changes due to version iteration, not only the general interface needs to be regression tested, but also all interfaces within the scope of the general interface need to be tested.

[0003] In the prior art, the first method of regression testing for resources and UI interfaces is to use a combination of manual and command methods to perform normal manual inspections on general interfaces, and to use commands to run the modified art resources in the game and manually confirm them. The second method is to use command screenshots of game resources, and perform image recognition comparison between the screenshots after version iteration and the original screenshots, so as to determine the problems in the images after version iteration.

[0004] However, the image recognition and comparison method still has false positives when the parameters are not adjusted properly, and the false positive images still need manual review. The manual review method is costly, inefficient and prone to errors. Summary of the invention

[0005] The purpose of the present application is to provide an image testing method, device, equipment and storage medium to address the deficiencies in the above-mentioned prior art, so as to solve the problems in the prior art that image testing relies on standard images, has false alarms, and manual inspection is costly, inefficient and prone to errors.

[0006] To achieve the above objectives, the technical solutions adopted in this application are as follows:

[0007] In a first aspect, the present application provides an image testing method, the method comprising:

[0008] Obtain test description information of a target interface in a game interface test scenario, and obtain an interface image of the target interface in the current version to be tested, wherein the test description information is used to describe checkpoints for the target interface and preset appearance features corresponding to the checkpoints;

[0009] Generate task indication information corresponding to the game interface test scenario, the task indication information including: model capability indication information and model output indication information;

[0010] The test description information, the task instruction information and the interface image are input into a large language model, and the large language model executes the task corresponding to the test description information for the interface image based on the task instruction information to obtain a test result of the target interface.

[0011] In a second aspect, the present application provides an image testing device, the device comprising:

[0012] An acquisition module, used to acquire test description information of a target interface in a game interface test scenario, and to acquire an interface image of the target interface in the current version to be tested, wherein the test description information is used to describe checkpoints for the target interface and preset appearance features corresponding to the checkpoints;

[0013] A generating module, used to generate task indication information corresponding to the game interface test scenario, wherein the task indication information includes: model capability indication information and model output indication information;

[0014] The test module is used to input the test description information, the task instruction information and the interface image into a large language model, and the large language model executes the task corresponding to the test description information for the interface image based on the task instruction information to obtain the test result of the target interface.

[0015] In a third aspect, an embodiment of the present application further provides an electronic device, comprising: a processor, a storage medium and a bus, wherein the storage medium stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor communicates with the storage medium through the bus, and the processor executes the machine-readable instructions to perform the steps of an image testing method as described in any one of the first aspects.

[0016] In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of an image testing method as described in any one of the first aspects are executed.

[0017] The beneficial effects of the present application are: obtaining the test description information of the target interface under the game interface test scenario, and obtaining the interface image of the target interface under the current version to be tested, describing the checkpoints for the target interface and the preset appearance features corresponding to the checkpoints through the test description information, generating the task indication information corresponding to the game interface test scenario, the task indication information includes: model capability indication information and model output indication information, inputting the test description information, task indication information and interface image into the large language model, and the large language model performs the task corresponding to the test description information for the interface image based on the task indication information, and obtains the test result of the target interface. Compared with the prior art, the present application does not have to rely on standard images as a judgment basis, and only needs to check the checkpoints in the interface image based on the test description information, and compared with the manual judgment method, the judgment through the large language model has higher efficiency. In addition, the test description information in the present application can be reused after being written once, and even if the version is updated again, regression testing can be quickly carried out on the updated interface image directly based on the test description information.

[0018] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, preferred embodiments are specifically cited below and described in detail with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0020] Figure 1 A flowchart of an image test provided by an embodiment of the present application is shown;

[0021] Figure 2 An example diagram of an interface picture provided by an embodiment of the present application is shown;

[0022] Figure 3 A flowchart of testing an interface image based on test description information provided by an embodiment of the present application is shown;

[0023] Figure 4 An example diagram of a display interface for an image test provided in an embodiment of the present application is shown;

[0024] Figure 5 A flowchart of obtaining a standard image provided by an embodiment of the present application is shown;

[0025] Figure 6A flowchart of testing an interface image based on test description information and a standard image provided by an embodiment of the present application is shown;

[0026] Figure 7 An example diagram of a display interface of another image test provided in an embodiment of the present application is shown;

[0027] Figure 8 A flowchart of testing a standard image based on test description information provided by an embodiment of the present application is shown;

[0028] Fig. 9 A schematic diagram of the structure of an image testing device provided in an embodiment of the present application is shown;

[0029] Fig.10 A schematic structural diagram of an electronic device provided in an embodiment of the present application is shown. DETAILED DESCRIPTION

[0030] To make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. The components of the embodiments of the present application generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application claimed for protection, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present application.

[0031] It should be noted that the term "comprising" will be used in the embodiments of the present application to indicate the existence of the features declared thereafter, but does not exclude the addition of other features.

[0032] In one embodiment of the present disclosure, the image testing method can be run on a local terminal device or a server. When the image testing method is run on a server, the method can be implemented and executed based on a cloud interaction system, wherein the cloud interaction system includes a server and a client device.

[0033] In an optional embodiment, various cloud applications can be run under the cloud interaction system, such as cloud games. Taking cloud games as an example, cloud games refer to a game mode based on cloud computing. In the operation mode of cloud games, the operating body of the game program and the main body of the game screen presentation are separated. The storage and operation of the image test method are completed on the cloud game server. The role of the client device is used for receiving and sending data and presenting the game screen. For example, the client device can be a display device with data transmission function close to the user side, such as a mobile terminal, a TV, a computer, a handheld computer, etc.; but the cloud game server in the cloud is used for information processing. When playing the game, the player operates the client device to send an operation instruction to the cloud game server. The cloud game server runs the game according to the operation instruction, encodes and compresses the game screen and other data, and returns it to the client device through the network. Finally, the client device decodes and outputs the game screen.

[0034] In an optional embodiment, taking a game as an example, a local terminal device stores a game program and is used to present a game screen. The local terminal device is used to interact with the player through a graphical user interface, that is, the game program is downloaded and installed by an electronic device and run conventionally. The local terminal device may provide the graphical user interface to the player in a variety of ways, for example, it may be rendered and displayed on a display screen of the terminal, or provided to the player through a holographic projection. For example, the local terminal device may include a display screen and a processor, the display screen is used to present a graphical user interface, the graphical user interface includes a game screen, and the processor is used to run the game, generate a graphical user interface, and control the display of the graphical user interface on the display screen.

[0035] In a possible implementation, an embodiment of the present invention provides an image testing method, which provides a graphical user interface through a terminal device, wherein the terminal device can be the local terminal device mentioned above, or can be a client device in the cloud interaction system mentioned above.

[0036] Regression testing of the interface means checking the application interface after the iteration of the application version to find errors in the application interface. Taking the regression testing of the game interface as an example, after the game version is iterated, the art resources and UI interface in the game interface need to be checked to determine whether the art resources and UI interface are modified correctly according to the expected modification method.

[0037] In game applications, many interfaces in the game will reuse the underlying common interface. Once the common interface is changed, it is necessary to test all the affected interfaces. The traditional way is to compare the game interface before the version update with the game interface after the version update to determine the location of the changed art resources or UI interface.

[0038] However, this method requires taking screenshots of the interface before the version update as a comparison benchmark, and there is a certain degree of false alarm rate. False alarm images still need to be manually reviewed again, so it still requires labor costs, and there are also problems with omissions and errors in manual review.

[0039] Based on this, the present application proposes an image testing method, which adds test description information to the interface image to be tested before the game version is iterated, and inputs the test description information and task instruction information into the large language model, and the large language model performs the test task on the interface image based on the task instruction information, so that the checkpoints in the interface image can be checked based on the test description information to obtain the test results of the target interface. The present application does not rely on standard images before the version update, has a higher accuracy rate, and can reduce labor costs and shorten the time of image regression testing.

[0040] Next, combine Figure 1 The image testing method of the present application is described as follows: Figure 1 As shown, the method includes:

[0041] S101. Obtain test description information of a target interface in a game interface test scenario, and obtain an interface image of the target interface in a current version to be tested, wherein the test description information is used to describe checkpoints for the target interface and preset appearance features corresponding to the checkpoints.

[0042] The game interface may include multiple checkpoints. Before the game version is iterated, the user can enter at least one description information for each checkpoint. After receiving all the description information of the user for the current interface, the device uses all the description information as the test description information of the current game interface. The checkpoint can be a virtual object included in the game interface, or a preset position on the UI interface.

[0043] The preset appearance features of the checkpoints may be appearance features of the checkpoints that do not change with the game version update, including the color, shape, light and shadow effects of the checkpoints, and the positional relationship, perspective effect, color relationship, etc. of the checkpoints with other checkpoints. Exemplarily, the user may determine the checkpoints in the target interface of the initial game version, and add appearance description text to each checkpoint, and use each checkpoint and the appearance description text of each checkpoint as the test description information of the target interface.

[0044] In one possible implementation, after the art resources in the game scene are rendered in the game, the user can take screenshots of the game scene at a preset perspective and add test description information to the checkpoints in each screenshot. In another implementation, the user can also directly add test description information for virtual objects in the game scene. If the target interface at different perspectives includes the virtual object, the test description information of the virtual object is used as the test description information of the target interface.

[0045] The test description information is used to describe the art resources in the game interface and the appearance characteristics of the checkpoints in the UI interface. For example, Figure 2 It is a screenshot of the game scene at a preset perspective. The game interface includes vehicle A and tree B. The test description information may be description information of the checkpoints in vehicle A and tree B, such as "the body of vehicle A is purple, the windows of vehicle A are closed and the interior of the vehicle cannot be seen through the windows, the color of the rearview mirror of vehicle A is consistent with the color of the body, and the vehicle has a projection in the sun".

[0046] The test description information can be described in natural language, which can be Chinese, English or other languages, and this application does not limit this. As an example, when ChatGPT is used as the large language model, the test description information can be in markdown format.

[0047] After the game version iteration is completed, a screenshot can be taken for the current version to be tested based on a preset perspective to obtain multiple interface images of the target interface under the current version to be tested. The interface images are the images that need to be tested, and each interface image corresponds to a test description information. In a possible implementation method, the same identifier can be added to the interface image and the test description information corresponding to the interface image.

[0048] S102. Generate task indication information corresponding to the game interface test scenario, where the task indication information includes: model capability indication information and model output indication information.

[0049] The model capability indication information may indicate the task description that the model needs to perform, the input data of the model, etc. For example, the model capability indication information may be "set the large language model as an expert in image recognition and image detection, the input data includes interface pictures and test description information, and the task goal is to check the checkpoints in the interface picture based on the test description information and find out the existing errors."

[0050] The model output indication information may indicate the language of the model output result, the steps of the model execution task, and the form of the model output result. Exemplarily, the model output indication information may indicate "the input test description information is in Chinese, and the output result is also in Chinese", and indicate that the steps of the model execution task are "check the checkpoints specified in the test description information, check each checkpoint without interference, and if the checkpoint inspection result cannot be clearly determined, it needs to be noted in the returned result", and indicate that the form of the output result is limited to "if the checkpoint inspection result is normal, reply 0, if there is an abnormal checkpoint, list the abnormal content".

[0051] The task instruction information is used to test the game interface, and the task instruction information of all game interfaces can be the same. As a possible implementation, the user can encode the features that need to be tested at the checkpoint to generate the task instruction information. In another implementation, the user can also input a natural description language. After receiving the natural description language input by the user, the device can extract and encapsulate the natural description language input by the user to generate the task instruction information.

[0052] S103: input the test description information, task instruction information and interface image into the large language model, and the large language model executes the task corresponding to the test description information on the interface image based on the task instruction information to obtain the test result of the target interface.

[0053] Among them, the large language model can be a natural language processing neural network model based on the Generative Pre-trained Transformer architecture.

[0054] In the first implementation manner, the large language model may be configured in advance based on the task instruction information, and after the configuration is completed, the interface images and the test description information corresponding to the interface images are sequentially input into the large language model.

[0055] In the second implementation, the task instruction information, each interface image, and the test description information corresponding to each interface image can be pre-packaged into a test item, and after obtaining the test items of all interface images, all the test items are sequentially input into the large language model. Specifically, the task instruction information and the test description information can be packaged based on the prompt technology to obtain the prompt word of the interface image, and the prompt word and the interface image are used as a test item.

[0056] The large language model can perform tasks according to the task instruction information, thereby checking the checkpoints in the interface image based on the features of the checkpoints described in the test description information, and determining whether the features of the checkpoints in the interface image are consistent with the features of the checkpoints in the test description information.

[0057] As an optional implementation, in order to further improve the accuracy of the large language model detection, the game interface before the version update can also be used as a standard image, and the standard image can also be used as input data for the large language model. At this time, it is also possible to further compare whether the interface image is consistent with the standard image, and whether the standard image is consistent with the test description information.

[0058] If the features of all checkpoints are consistent, the test result of the target interface can be determined to be normal. If there are inconsistent checkpoint features, the inconsistent checkpoints can be used as abnormal checkpoints, and the abnormal results of the abnormal checkpoints can be output.

[0059] For example, suppose the test description information is "Vehicle A is purple", and the large language model checks the vehicle A in the interface picture 1 and finds that the vehicle A is red. At this time, the large language model can output the abnormal result "The color of the vehicle A in the interface picture 1 is inconsistent, the description information is purple, and the vehicle A in the picture is red".

[0060] In the embodiment of the present application, the test description information of the target interface under the game interface test scenario is obtained, and the interface image of the target interface under the current version to be tested is obtained. The checkpoints for the target interface and the preset appearance features corresponding to the checkpoints are described by the test description information, and the task indication information corresponding to the game interface test scenario is generated. The task indication information includes: model capability indication information and model output indication information. The test description information, task indication information and interface image are input into the large language model, and the large language model performs the task corresponding to the test description information for the interface image based on the task indication information to obtain the test result of the target interface. Compared with the prior art, the present application does not have to rely on standard images as a judgment basis. It only needs to check the checkpoints in the interface image based on the test description information, and compared with the manual judgment method, the judgment through the large language model has higher efficiency. In addition, the test description information in the present application can be reused after being written once, and even if the version is updated again, regression testing can be quickly carried out on the updated interface image directly based on the test description information.

[0061] In the first implementation, detection can be performed based on the test description information and the interface image to obtain the detection result of the interface image. The model capability indication information is used to instruct the large language model to check the interface image according to the test description information; the model output indication information is used to indicate the output result of the large language model under different inspection results.

[0062] Optionally, the test description information may include description information of multiple checkpoints, and the model capability indication information may indicate the large language model to check each checkpoint in the interface image according to the description information of each checkpoint. The model output indication information may indicate the output result of the large language model when the check result is normal and when the check result is abnormal.

[0063] like Figure 3 As shown, the process of inputting the test description information, the task instruction information and the interface image into the large language model, and the large language model executing the task corresponding to the test description information for the interface image based on the task instruction information to obtain the test result of the target interface includes:

[0064] S301, inputting test description information, task instruction information and interface pictures into a large language model.

[0065] As a possible implementation method, the interface image, the test description information corresponding to the interface image, and the task instruction information can be used as a test item. For example, the prompt engineering technology is used to encapsulate the test description information and the task instruction information into a test case, and the test case and the interface image are stored in the database as a test item.

[0066] When inputting data, the items to be tested in the database can be asynchronously read in batches through pre-written scripts, and the items to be tested can be input into the large language model through the API interface.

[0067] S302: The large language model compares and checks the interface image and the test description information according to the model capability indication information to obtain a first inspection result.

[0068] Among them, the first inspection result can represent whether the characteristics of each inspection point in the interface image are consistent with the characteristics of each inspection point in the test description information.

[0069] The large language model can generate and execute image detection and recognition tasks according to the model capability indication information, thereby comparing and checking the checkpoint features in the interface image with the checkpoint features in the test description information to obtain the first inspection result of each checkpoint.

[0070] S303: The large language model determines a first output result corresponding to the first inspection result according to the model output indication information, and uses the first output result as a test result of the target interface.

[0071] The model output indication information can indicate the return results when the interface image is consistent with the checkpoints in the test description information and when they are inconsistent.

[0072] As a possible implementation, when the checkpoints in the interface image and the checkpoints in the test description information are inconsistent, the first output result may be a description of the checkpoints in the interface image and a description of the checkpoints in the test description information.

[0073] As another possible implementation method, if the description of the checkpoint in the test description information is unclear, the large language model may not be able to determine whether the checkpoint in the interface image is consistent with the checkpoint in the test description information. The first output result can indicate that the checkpoint check failed to prompt the user to modify the description information of the checkpoint in the test description information.

[0074] For example, suppose that Figure 2 For vehicle A, the test description information includes "check whether the license plate of vehicle A is clear". The large language model cannot make a judgment on this description information, and the first output result can be "unable to determine whether the license plate of vehicle A is clear", thereby reminding the user to modify the test description information.

[0075] like Figure 4 As shown, it is a schematic diagram of an image detection interface provided by the present application. In the display interface, the user can select the item to be tested, which includes an interface image 401 and a test description information 402. After the selection is completed, the large language model is called through the API interface to obtain the test result shown in area 403. It should be understood that in the present application, multiple interface images can be tested at one time, and the test results of each interface image can be obtained.

[0076] In the second implementation, users can also use standard images as input data for the large language model. Before iterating the application version, you can first obtain standard images, such as Figure 5 As shown, specifically including:

[0077] S501 , obtaining an interface image of a target interface under a historical version to be tested, where the historical version to be tested is a version compiled before the current version to be tested.

[0078] S502: Use the interface image of the target interface in the historical version to be tested as the standard image.

[0079] The historical version to be tested may be any application version before the current version to be tested, such as the previous version of the current version, or the initial version after the application is developed. For example, if the current version is 2.0.1 and the previous version is 2.0.0, then a screenshot of the game interface of version 2.0.0 may be taken, specifically, the virtual objects in the game scene may be rendered, and then a snapshot of the game scene may be taken at a preset viewing angle to obtain interface images of each target interface.

[0080] Taking a car game as an example, the initial version of the resources can be rendered in the game, and after the game resources are dressed up through instructions, snapshots can be taken at a preset perspective for each game interface that needs to be tested, and the images can be saved as standard images.

[0081] In a possible implementation, taking 3D games as an example, multiple test points can be pre-set in the game scene, each test point has an identifier, and the position of the test point in each version is the same as the screenshot perspective. After taking a screenshot of each test point, the identifier of the test point can be added to the image, so that the standard image and the interface image of the same test point have the same identifier, and the standard image corresponding to the interface image can be quickly determined when the large language model is subsequently input, that is, the standard image and interface image under different versions of the same test point and the same perspective.

[0082] Optionally, the model capability indication information is used to instruct the large language model to check the interface image according to the test description information and the standard image; the model output indication information is used to indicate the output results of the large language model under different inspection results.

[0083] In the first implementation, the standard image may include multiple checkpoints, and the model capability indication information may instruct the large language model to check the checkpoints in the interface image according to the checkpoint features in the test description information and the checkpoints in the standard image. At this time, the model output indication information may indicate the output results of the large language model when the check result is normal and when the check result is abnormal.

[0084] In the second implementation, the model capability indication information may instruct the large language model to check each checkpoint in the interface image according to the description information of each checkpoint, and instruct the large language model to perform image comparison between the standard image and the interface image, thereby determining the difference between the standard image and the interface image. At this time, the model output indication information may indicate the output result when the large language model check result is normal and the check result is abnormal, and indicate the difference between the standard image and the interface image.

[0085] like Figure 6 As shown, the process of inputting the test description information, the task instruction information and the interface image into the large language model, and the large language model executing the task corresponding to the test description information for the interface image based on the task instruction information to obtain the test result of the target interface includes:

[0086] S601: Input test description information, task instruction information, standard pictures and interface pictures into a large language model.

[0087] Optionally, the test description information, task instruction information, standard pictures and interface pictures can be used as a test item. The test description information and task instruction information are encapsulated as a test case, and the test case, interface picture and the standard picture corresponding to the interface picture are stored in the database as a test item. When inputting data, the test items in the database can be asynchronously read in batches through a pre-written script, and the test items can be input into the large language model through the API interface.

[0088] S602: The large language model compares and checks the interface image and the test description information according to the model capability indication information to obtain a second inspection result.

[0089] The second inspection result can indicate whether the features of each checkpoint in the interface image are consistent with the features of each checkpoint in the test description information. The large language model can generate and perform image detection and recognition tasks according to the model capability indication information, thereby comparing and inspecting the features of the checkpoints in the interface image with the features of the checkpoints in the test description information to obtain the second inspection result of each checkpoint.

[0090] S603: The large language model compares and checks the interface image with the standard image according to the model capability indication information to obtain a third inspection result.

[0091] The third check result can indicate whether the standard image and the interface image are the same. The large language model can compare the interface image with the standard image to determine the location of the interface image that is different from the standard image.

[0092] S604: The large language model determines the second output result corresponding to the second inspection result and the third output result corresponding to the third inspection result according to the model output indication information, and uses the second output result and the third output result as the test results of the target interface.

[0093] When the checkpoints in the interface image and the test description information are inconsistent, the second output result may be a description of the checkpoints in the interface image and the description of the checkpoints in the test description information. The second output result may also be feedback on the abnormal description of the checkpoints in the test description information.

[0094] When there is a difference between the interface image and the standard image, the third output result may be a description of the difference. For example, assuming that vehicle A in the interface image is red and vehicle A in the standard image is white, the third output result may be "the vehicle colors in the interface image and the standard image are inconsistent, the vehicle color in the interface image is red, and the vehicle color in the standard image is white".

[0095] like Figure 7, which is another image detection interface schematic diagram provided by the present application. In the display interface, the user can select the items to be tested, which include an interface image 701, test description information 702 of the interface image, and a standard image 703 corresponding to the interface image. After the selection is completed, the large language model is called through the API interface to obtain the test results shown in area 704.

[0096] As a possible implementation method, the large language model can also check the standard image based on the test description information to find errors in the standard image. In this case, the model capability indication information is used to instruct the large language model to check the standard image according to the test description information, and the model output indication information is used to indicate the output results of the large language model under different inspection results. Figure 8 As shown, the method of the present application also includes:

[0097] S801. The large language model compares and checks the standard image and the test description information according to the model capability indication information to obtain a fourth inspection result.

[0098] The fourth inspection result can indicate whether the checkpoint features in the standard image are consistent with the description information of the checkpoint in the test description information. The large language model can create and execute the image detection and recognition task according to the model capability indication information, recognize the checkpoints in the standard image according to the test description information, and determine whether the checkpoints are consistent with the description information of the checkpoints included in the test description information, and obtain the fourth inspection result of each checkpoint.

[0099] S802: The large language model determines a fourth output result corresponding to the fourth inspection result according to the model output indication information, and uses the fourth output result as a test result of the target interface.

[0100] When the checkpoints in the standard image and the checkpoints in the test description information are inconsistent, the fourth output result may be a description of the checkpoints in the standard image and a description of the checkpoints in the test description information. The fourth output result may also be feedback on an abnormal description of the checkpoints in the test description information.

[0101] Optionally, the test description information includes test description information of multiple checkpoints, and the test description information of each checkpoint is used to describe the appearance characteristics of the checkpoint. If the output result indicates that there is an abnormality in the description of the checkpoint, for example, the description information of the checkpoint includes a degree adverb that the large language model cannot judge, the test description information can be corrected until the large language model can make an accurate judgment based on the test description information. The process of correcting the test description information includes:

[0102] If there are abnormal results in the test results of the target interface, the test description information of the checkpoint corresponding to the abnormal result is corrected to obtain the corrected test description information.

[0103] After outputting feedback of the checkpoint description anomaly to the user, the device may receive new description information for the checkpoint input by the user, generate a modification instruction, and modify the description information of the checkpoint with the anomaly to the new description information input by the user through the modification instruction.

[0104] It should be noted that multiple tests can be performed in the present application until the output result does not include feedback of abnormal test description information of the checkpoint.

[0105] Continue to refer to Figure 4 When the method of the present application is applied to an electronic device, a test configuration interface is provided through a graphical user interface of the electronic device, and the test configuration interface includes a first configuration area 401, a second configuration area 402 and a third configuration area 403.

[0106] The above steps of obtaining the interface image of the target interface under the current version to be tested include:

[0107] In response to a first selection instruction for the first configuration area, an interface image of a target interface under the current version to be tested indicated by the first selection instruction is obtained.

[0108] In an optional implementation, the first configuration area can be linked to a secondary interface, which displays all interface images that the user can select. After the user selects an interface image, the user returns to the secondary interface. Figure 4 In the test configuration interface shown, the interface picture selected by the user can be displayed in the first configuration area 401.

[0109] The above step of obtaining the test description information of the target interface in the game interface test scenario includes:

[0110] In response to the configuration instruction for the second configuration area, the test description information in the second configuration area is acquired.

[0111] As a possible implementation, the user can input the test description text in the second configuration area in accordance with the standard image, and the electronic device can encapsulate the test description text input by the user into test description information. The user can input multiple test description texts, each test description text is used to describe the appearance characteristics of a checkpoint. The user can input the test description text before testing the current version to be tested.

[0112] After executing the task corresponding to the test description information on the interface image and obtaining the test result of the target interface, the method of the present application also includes: displaying the test result in a third configuration area.

[0113] Optionally, the test configuration interface may also include a fourth configuration area, referring to Figure 7 The fourth configuration area 703 is used to respond to the user's selection instruction and select a standard picture corresponding to the current interface picture from the standard picture storage file. At this time, the electronic device can test the interface picture based on the test description information and the standard picture.

[0114] Optionally, after obtaining the test result of the target interface, the method of the present application further includes:

[0115] If the test result indicates that there is an error in the checkpoint of the target interface, problem verification indication information is generated according to the test result, the target interface and the standard image of the target interface, so as to correct the error in the checkpoint based on the problem verification information.

[0116] In the first implementation method, problem verification indication information can be generated based on the test results, the target interface and the standard image of the target interface, and the problem verification indication information can be sent to the developer. The problem verification indication information is used to indicate the interface image with problems in the test results, as well as the difference between the interface image and the standard image, and instruct the developer to verify the cause of the difference.

[0117] In the second implementation method, if the test results indicate that there is an error in the checkpoint, the checkpoint can be located, the code segment related to the checkpoint can be determined, and problem verification indication information can be generated based on the location information of the code segment and the test results. The problem verification indication information is sent to the developer so that the developer can modify the code segment at the relevant location based on the test results in the problem verification indication information.

[0118] The following is a further description of obtaining an interface image of a target interface under the current version to be tested. In an optional implementation, this step includes:

[0119] The virtual scene under the current version to be tested is frozen to obtain a frozen virtual scene, and according to the position information of each test point in the virtual scene, the interface picture of each test point in the frozen virtual scene under a preset viewing angle is collected.

[0120] It should be understood that the game scene may include dynamic game resources and animation effects. In order to ensure the accuracy of the game scene test, before obtaining the interface image and standard image, all resources in the game scene can be frozen to keep all resources in the game scene in an initialized state.

[0121] The test point may be a specific position pre-marked in the game scene, and the position information of the test point may be the coordinate position of the test point in the game scene. The viewing angle of each test point may be pre-set by the user and remain unchanged, and the game scene under this viewing angle is the target interface. When collecting interface images, the game scene may be screenshotted based on the preset viewing angle of each test point at the position coordinates of each test point to obtain an interface image.

[0122] The above is an explanation of an embodiment of obtaining interface images in a three-dimensional game scene. It should be noted that for some two-dimensional games, the game perspective and screen are fixed. At this time, each target interface can be directly screenshoted to obtain an interface image of each target interface.

[0123] Based on the same inventive concept, an image testing device corresponding to the image testing method is also provided in the embodiment of the present application. Since the principle of solving the problem by the device in the embodiment of the present application is similar to the above-mentioned image testing method in the embodiment of the present application, the implementation of the device can refer to the implementation of the method, and the repeated parts will not be repeated.

[0124] Fig. 9 A schematic diagram of the structure of an image testing device provided in an embodiment of the present application is shown. Fig. 9 The device includes an acquisition module 901, a generation module 902 and a testing module 903.

[0125] The acquisition module 901 is used to acquire the test description information of the target interface in the game interface test scenario, and acquire the interface image of the target interface in the current version to be tested, wherein the test description information is used to describe the checkpoints for the target interface and the preset appearance features corresponding to the checkpoints;

[0126] A generating module 902 is used to generate task indication information corresponding to the game interface test scenario, where the task indication information includes: model capability indication information and model output indication information;

[0127] The test module 903 is used to input the test description information, task instruction information and interface image into the large language model, and the large language model executes the task corresponding to the test description information on the interface image based on the task instruction information to obtain the test result of the target interface.

[0128] In a feasible implementation scheme, the model capability indication information is used to instruct the large language model to check the interface image according to the test description information; the model output indication information is used to indicate the output results of the large language model under different inspection results.

[0129] In a feasible implementation scheme, the testing module 903 is specifically used for:

[0130] Input the test description information, task instruction information and interface pictures into the large language model;

[0131] The large language model compares and checks the interface image and the test description information according to the model capability indication information to obtain a first inspection result;

[0132] The large language model determines a first output result corresponding to the first inspection result according to the model output indication information, and uses the first output result as the test result of the target interface.

[0133] In a feasible implementation scheme, the acquisition module 901 is further used for:

[0134] Get the interface image of the target interface under the historical version to be tested, where the historical version to be tested is the version compiled before the current version to be tested;

[0135] The interface image of the target interface in the historical version to be tested is used as the standard image.

[0136] In a feasible implementation scheme, the model capability indication information is used to instruct the large language model to check the interface image according to the test description information and the standard image; the model output indication information is used to indicate the output result of the large language model under different inspection results.

[0137] In a feasible implementation scheme, the testing module 903 is specifically used for:

[0138] Input the test description information, task instruction information, standard pictures and interface pictures into the large language model;

[0139] The large language model compares and checks the interface image with the test description information according to the model capability indication information to obtain a second inspection result;

[0140] The large language model compares and checks the interface image with the standard image according to the model capability indication information to obtain a third inspection result;

[0141] The large language model determines the second output result corresponding to the second inspection result and the third output result corresponding to the third inspection result according to the model output indication information, and uses the second output result and the third output result as the test results of the target interface.

[0142] In a feasible implementation manner, the model capability indication information is used to instruct the large language model to check the standard image according to the test description information, and the model output indication information is used to indicate the output result of the large language model under different checking results.

[0143] In a feasible implementation scheme, the testing module 903 is further used to:

[0144] The large language model compares and checks the standard image and the test description information according to the model capability indication information to obtain a fourth inspection result;

[0145] The large language model determines a fourth output result corresponding to the fourth inspection result according to the model output indication information, and uses the fourth output result as the test result of the target interface.

[0146] In a feasible implementation scheme, the test description information includes test description information of multiple checkpoints;

[0147] The testing module 903 is also used for:

[0148] If there are abnormal results in the test results of the target interface, the test description information of the checkpoint corresponding to the abnormal result is corrected to obtain the corrected test description information.

[0149] In a feasible implementation, when the method is applied to an electronic device, a test configuration interface is provided through a graphical user interface of the electronic device, and the test configuration interface includes a first configuration area, a second configuration area, and a third configuration area;

[0150] The acquisition module 901 is specifically used for:

[0151] In response to a first selection instruction for the first configuration area, obtaining an interface image of a target interface under the current version to be tested indicated by the first selection instruction;

[0152] The acquisition module 901 is specifically used for:

[0153] In response to a configuration instruction for the second configuration area, acquiring test description information in the second configuration area;

[0154] The testing module 903 is specifically used for:

[0155] The test results are displayed in the third configuration area.

[0156] In a feasible implementation scheme, the testing module 903 is further specifically used for:

[0157] If the test result indicates that there is an error in the checkpoint of the target interface, problem verification indication information is generated according to the test result, the target interface and the standard image of the target interface, so as to correct the error in the checkpoint based on the problem verification information.

[0158] In a feasible implementation scheme, the acquisition module 901 is specifically used for:

[0159] The virtual scene under the current version to be tested is frozen to obtain a frozen virtual scene, and according to the position information of each test point in the virtual scene, the interface picture of each test point in the frozen virtual scene under a preset viewing angle is collected.

[0160] Compared with the prior art, this application does not need to rely on standard images as a judgment benchmark. It only needs to check the checkpoints in the interface image based on the test description information, and compared with manual judgment, judgment through a large language model is more efficient. In addition, the test description information in this application can be reused after being written once. Even if the version is updated again, regression testing can be quickly carried out on the updated interface image based on the test description information.

[0161] Fig.10 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application is shown, including: a processor 1001, a storage medium 1002 and a bus 1003, wherein the storage medium 1002 stores machine-readable instructions executable by the processor 1001. When the electronic device runs an image testing method in the embodiment, the processor 1001 communicates with the storage medium 1002 via the bus 1003, and the processor 1001 executes the machine-readable instructions. The processor 1001 performs the preamble of the method item to perform the following steps:

[0162] Obtain test description information of the target interface in the game interface test scenario, and obtain an interface image of the target interface in the current version to be tested, wherein the test description information is used to describe the checkpoints for the target interface and the preset appearance features corresponding to the checkpoints;

[0163] Generate task indication information corresponding to the game interface test scenario, the task indication information including: model capability indication information and model output indication information;

[0164] The test description information, task instruction information and interface image are input into the large language model, and the large language model executes the task corresponding to the test description information on the interface image based on the task instruction information to obtain the test result of the target interface.

[0165] In a feasible implementation scheme, the model capability indication information is used to instruct the large language model to check the interface image according to the test description information; the model output indication information is used to indicate the output results of the large language model under different inspection results.

[0166] In a feasible implementation scheme, the processor 1001 is specifically used to input the test description information, the task instruction information and the interface image into the large language model, and the large language model performs the task corresponding to the test description information on the interface image based on the task instruction information to obtain the test result of the target interface:

[0167] Input the test description information, task instruction information and interface pictures into the large language model;

[0168] The large language model compares and checks the interface image and the test description information according to the model capability indication information to obtain a first inspection result;

[0169] The large language model determines a first output result corresponding to the first inspection result according to the model output indication information, and uses the first output result as the test result of the target interface.

[0170] In a feasible implementation manner, the processor 1001 is further configured to execute:

[0171] Get the interface image of the target interface under the historical version to be tested, where the historical version to be tested is the version compiled before the current version to be tested;

[0172] The interface image of the target interface in the historical version to be tested is used as the standard image.

[0173] In a feasible implementation scheme, the model capability indication information is used to instruct the large language model to check the interface image according to the test description information and the standard image; the model output indication information is used to indicate the output result of the large language model under different inspection results.

[0174] In a feasible implementation scheme, the processor 1001 is specifically used to input the test description information, the task instruction information and the interface image into the large language model, and the large language model performs the task corresponding to the test description information on the interface image based on the task instruction information to obtain the test result of the target interface:

[0175] Input the test description information, task instruction information, standard pictures and interface pictures into the large language model;

[0176] The large language model compares and checks the interface image with the test description information according to the model capability indication information to obtain a second inspection result;

[0177] The large language model compares and checks the interface image with the standard image according to the model capability indication information to obtain a third inspection result;

[0178] The large language model determines the second output result corresponding to the second inspection result and the third output result corresponding to the third inspection result according to the model output indication information, and uses the second output result and the third output result as the test results of the target interface.

[0179] In a feasible implementation scheme, the model capability indication information is used to instruct the large language model to check the standard image according to the test description information, and the model output indication information is used to indicate the output result of the large language model under different checking results.

[0180] In a feasible implementation manner, the processor 1001 is further configured to execute:

[0181] The large language model compares and checks the standard image and the test description information according to the model capability indication information to obtain a fourth inspection result;

[0182] The large language model determines a fourth output result corresponding to the fourth inspection result according to the model output indication information, and uses the fourth output result as the test result of the target interface.

[0183] In a feasible implementation scheme, the test description information includes test description information of multiple checkpoints;

[0184] The processor 1001 is further configured to execute:

[0185] If there are abnormal results in the test results of the target interface, the test description information of the checkpoint corresponding to the abnormal result is corrected to obtain the corrected test description information.

[0186] In a feasible implementation, when the method is applied to an electronic device, a test configuration interface is provided through a graphical user interface of the electronic device, and the test configuration interface includes a first configuration area, a second configuration area, and a third configuration area;

[0187] When executing to obtain the interface image of the target interface under the current version to be tested, the processor 1001 is specifically used to:

[0188] In response to a first selection instruction for the first configuration area, obtaining an interface image of a target interface under the current version to be tested indicated by the first selection instruction;

[0189] When executing to obtain the test description information of the target interface in the game interface test scenario, the processor 1001 is specifically used to:

[0190] In response to a configuration instruction for the second configuration area, acquiring test description information in the second configuration area;

[0191] After executing the task corresponding to the test description information for the interface image and obtaining the test result of the target interface, the processor 1001 is specifically used to:

[0192] The test results are displayed in the third configuration area.

[0193] In a feasible implementation manner, after executing and obtaining the test result of the target interface, the processor 1001 is further configured to:

[0194] If the test result indicates that there is an error in the checkpoint of the target interface, problem verification indication information is generated according to the test result, the target interface and the standard image of the target interface, so as to correct the error in the checkpoint based on the problem verification information.

[0195] In a feasible implementation scheme, when the processor 1001 executes the acquisition of the interface image of the target interface under the current version to be tested, it is specifically used to:

[0196] The virtual scene under the current version to be tested is frozen to obtain a frozen virtual scene, and according to the position information of each test point in the virtual scene, the interface picture of each test point in the frozen virtual scene under a preset viewing angle is collected.

[0197] This application does not rely on standard images as a judgment benchmark. It only needs to check the checkpoints in the interface image based on the test description information, and compared with manual judgment, judgment through a large language model is more efficient. In addition, the test description information in this application can be reused after being written once. Even if the version is updated again, regression testing can be quickly carried out on the updated interface image based on the test description information.

[0198] The embodiment of the present application further provides a computer-readable storage medium, on which a computer program is stored. The computer program is executed when a processor is running, and the processor performs the following steps:

[0199] Obtain test description information of the target interface in the game interface test scenario, and obtain an interface image of the target interface in the current version to be tested, wherein the test description information is used to describe the checkpoints for the target interface and the preset appearance features corresponding to the checkpoints;

[0200] Generate task indication information corresponding to the game interface test scenario, the task indication information including: model capability indication information and model output indication information;

[0201] The test description information, task instruction information and interface image are input into the large language model, and the large language model executes the task corresponding to the test description information on the interface image based on the task instruction information to obtain the test result of the target interface.

[0202] In a feasible implementation scheme, the model capability indication information is used to instruct the large language model to check the interface image according to the test description information; the model output indication information is used to indicate the output results of the large language model under different inspection results.

[0203] In a feasible implementation scheme, when the processor inputs the test description information, the task instruction information and the interface image into the large language model, and the large language model performs the task corresponding to the test description information on the interface image based on the task instruction information to obtain the test result of the target interface, it is specifically used to:

[0204] Input the test description information, task instruction information and interface pictures into the large language model;

[0205] The large language model compares and checks the interface image and the test description information according to the model capability indication information to obtain a first inspection result;

[0206] The large language model determines a first output result corresponding to the first inspection result according to the model output indication information, and uses the first output result as the test result of the target interface.

[0207] In one feasible embodiment, the processor is further configured to execute:

[0208] Get the interface image of the target interface under the historical version to be tested, where the historical version to be tested is the version compiled before the current version to be tested;

[0209] The interface image of the target interface in the historical version to be tested is used as the standard image.

[0210] In a feasible implementation scheme, the model capability indication information is used to instruct the large language model to check the interface image according to the test description information and the standard image; the model output indication information is used to indicate the output result of the large language model under different inspection results.

[0211] In a feasible implementation scheme, when the processor inputs the test description information, the task instruction information and the interface image into the large language model, and the large language model performs the task corresponding to the test description information on the interface image based on the task instruction information to obtain the test result of the target interface, it is specifically used to:

[0212] Input the test description information, task instruction information, standard pictures and interface pictures into the large language model;

[0213] The large language model compares and checks the interface image with the test description information according to the model capability indication information to obtain a second inspection result;

[0214] The large language model compares and checks the interface image with the standard image according to the model capability indication information to obtain a third inspection result;

[0215] The large language model determines the second output result corresponding to the second inspection result and the third output result corresponding to the third inspection result according to the model output indication information, and uses the second output result and the third output result as the test results of the target interface.

[0216] In a feasible implementation scheme, the model capability indication information is used to instruct the large language model to check the standard image according to the test description information, and the model output indication information is used to indicate the output result of the large language model under different checking results.

[0217] In one feasible embodiment, the processor is further configured to execute:

[0218] The large language model compares and checks the standard image and the test description information according to the model capability indication information to obtain a fourth inspection result;

[0219] The large language model determines a fourth output result corresponding to the fourth inspection result according to the model output indication information, and uses the fourth output result as the test result of the target interface.

[0220] In a feasible implementation scheme, the test description information includes test description information of multiple checkpoints;

[0221] The processor is further configured to execute:

[0222] If there are abnormal results in the test results of the target interface, the test description information of the checkpoint corresponding to the abnormal result is corrected to obtain the corrected test description information.

[0223] In a feasible implementation, when the method is applied to an electronic device, a test configuration interface is provided through a graphical user interface of the electronic device, and the test configuration interface includes a first configuration area, a second configuration area, and a third configuration area;

[0224] When the processor executes to obtain the interface image of the target interface under the current version to be tested, it is specifically used to:

[0225] In response to a first selection instruction for the first configuration area, obtaining an interface image of a target interface under the current version to be tested indicated by the first selection instruction;

[0226] When executing to obtain the test description information of the target interface in the game interface test scenario, the processor is specifically used to:

[0227] In response to a configuration instruction for the second configuration area, acquiring test description information in the second configuration area;

[0228] After executing the task corresponding to the test description information for the interface image and obtaining the test result of the target interface, the processor is specifically used to:

[0229] The test results are displayed in the third configuration area.

[0230] In a feasible implementation manner, after executing and obtaining the test result of the target interface, the processor is further configured to:

[0231] If the test result indicates that there is an error in the checkpoint of the target interface, problem verification indication information is generated according to the test result, the target interface and the standard image of the target interface, so as to correct the error in the checkpoint based on the problem verification information.

[0232] In a feasible implementation scheme, when the processor executes the acquisition of the interface image of the target interface under the current version to be tested, it is specifically used to:

[0233] The virtual scene under the current version to be tested is frozen to obtain a frozen virtual scene, and according to the position information of each test point in the virtual scene, the interface picture of each test point in the frozen virtual scene under a preset viewing angle is collected.

[0234] This application does not rely on standard images as a judgment benchmark. It only needs to check the checkpoints in the interface image based on the test description information, and compared with manual judgment, judgment through a large language model is more efficient. In addition, the test description information in this application can be reused after being written once. Even if the version is updated again, regression testing can be quickly carried out on the updated interface image based on the test description information.

[0235] In the embodiment of the present application, the computer program can also execute other machine-readable instructions when run by the processor to execute other methods described in the embodiment. For the specific execution method steps and principles, please refer to the description of the embodiment, which will not be repeated here.

[0236] In the embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interfaces, and the indirect coupling or communication connection of the devices or units can be electrical, mechanical or other forms.

[0237] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0238] In addition, each functional unit in the embodiments provided in the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0239] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0240] It should be noted that similar numbers and letters represent similar items in the following figures. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. In addition, the terms "first", "second", "third", etc. are only used to distinguish the description and are not to be understood as indicating or implying relative importance.

[0241] Finally, it should be noted that the above-described embodiments are only specific implementation methods of the present application, which are used to illustrate the technical solutions of the present application, rather than to limit them. The protection scope of the present application is not limited thereto. Although the present application is described in detail with reference to the above-mentioned embodiments, ordinary technicians in the field should understand that any technician familiar with the technical field can still modify the technical solutions recorded in the above-mentioned embodiments within the technical scope disclosed in the present application, or can easily think of changes, or make equivalent replacements for some of the technical features therein; 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 the embodiments of the present application. They should all be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be based on the protection scope of the claims.

Claims

1. An image testing method, characterized in that: include: Obtain test description information of a target interface in a game interface test scenario, and obtain an interface image of the target interface in the current version to be tested, wherein the test description information is used to describe checkpoints for the target interface and preset appearance features corresponding to the checkpoints; Generate task indication information corresponding to the game interface test scenario, the task indication information including: model capability indication information and model output indication information; The test description information, the task instruction information and the interface image are input into a large language model, and the large language model executes the task corresponding to the test description information for the interface image based on the task instruction information to obtain a test result of the target interface.

2. The method according to claim 1, characterized in that The model capability indication information is used to instruct the large language model to check the interface image according to the test description information; the model output indication information is used to indicate the output results of the large language model under different inspection results.

3. The method according to claim 1, characterized in that The step of inputting the test description information, the task instruction information, and the interface image into a large language model, and having the large language model perform the task corresponding to the test description information on the interface image based on the task instruction information to obtain a test result of the target interface includes: Inputting the test description information, the task instruction information and the interface image into a large language model; The large language model compares and checks the interface image with the test description information according to the model capability indication information to obtain a first check result; The large language model determines a first output result corresponding to the first inspection result according to the model output indication information, and uses the first output result as a test result of the target interface.

4. The method according to claim 1, characterized in that: The method further comprises: Obtaining an interface image of the target interface under a historical version to be tested, wherein the historical version to be tested is a version compiled before the current version to be tested; The interface picture of the target interface in the historical version to be tested is used as the standard picture.

5. The method according to claim 4, characterized in that The model capability indication information is used to instruct the large language model to check the interface image according to the test description information and the standard image; the model output indication information is used to indicate the output results of the large language model under different inspection results.

6. The method according to claim 5, characterized in that The step of inputting the test description information, the task instruction information, and the interface image into a large language model, and having the large language model perform the task corresponding to the test description information on the interface image based on the task instruction information to obtain a test result of the target interface includes: Inputting the test description information, the task instruction information, the standard picture and the interface picture into a large language model; The large language model compares and checks the interface image with the test description information according to the model capability indication information to obtain a second inspection result; The large language model compares and checks the interface image with the standard image according to the model capability indication information to obtain a third inspection result; The large language model determines the second output result corresponding to the second inspection result and the third output result corresponding to the third inspection result according to the model output indication information, and uses the second output result and the third output result as the test results of the target interface.

7. The method according to claim 4, characterized in that The model capability indication information is used to instruct the large language model to check the standard image according to the test description information, and the model output indication information is used to indicate the output result of the large language model under different checking results.

8. The method according to claim 7, characterized in that The method further comprises: The large language model compares and checks the standard image with the test description information according to the model capability indication information to obtain a fourth inspection result; The large language model determines a fourth output result corresponding to the fourth inspection result according to the model output indication information, and uses the fourth output result as the test result of the target interface.

9. The method according to any one of claims 1 to 8, characterized in that: The test description information includes test description information of multiple checkpoints; The method further comprises: If there are abnormal results in the test results of the target interface, the test description information of the checkpoint corresponding to the abnormal result is corrected to obtain the corrected test description information.

10. The method according to claim 1, characterized in that When the method is applied to an electronic device, a test configuration interface is provided through a graphical user interface of the electronic device, wherein the test configuration interface includes a first configuration area, a second configuration area, and a third configuration area; The step of obtaining the interface image of the target interface under the current version to be tested includes: In response to a first selection instruction for the first configuration area, obtaining an interface picture of the target interface under the current version to be tested indicated by the first selection instruction; The step of obtaining the test description information of the target interface in the game interface test scenario includes: In response to a configuration instruction for the second configuration area, acquiring test description information in the second configuration area; After executing the task corresponding to the test description information on the interface image and obtaining the test result of the target interface, the method further includes: The test result is displayed in the third configuration area.

11. The method according to claim 1, characterized in that After obtaining the test result of the target interface, the method further includes: If the test result indicates that there is an error in the checkpoint of the target interface, problem verification indication information is generated according to the test result, the target interface and the standard image of the target interface, so as to correct the error of the checkpoint based on the problem verification information.

12. The method according to claim 1, characterized in that The step of obtaining the interface image of the target interface under the current version to be tested includes: The virtual scene under the current version to be tested is frozen to obtain a frozen virtual scene, and according to the position information of each test point in the virtual scene, an interface picture of each test point in the frozen virtual scene under a preset viewing angle is collected.

13. An image testing device, characterized in that: include: An acquisition module, used to acquire test description information of a target interface in a game interface test scenario, and to acquire an interface image of the target interface in the current version to be tested, wherein the test description information is used to describe checkpoints for the target interface and preset appearance features corresponding to the checkpoints; A generating module, used to generate task indication information corresponding to the game interface test scenario, wherein the task indication information includes: model capability indication information and model output indication information; The test module is used to input the test description information, the task instruction information and the interface image into a large language model, and the large language model executes the task corresponding to the test description information for the interface image based on the task instruction information to obtain the test result of the target interface.

14. An electronic device, characterized in that: include: A processor, a storage medium and a bus, wherein the storage medium stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor communicates with the storage medium via the bus, and the processor executes the machine-readable instructions to perform the steps of the image testing method as described in any one of claims 1 to 12.

15. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the image testing method according to any one of claims 1 to 12 are executed.

Images