Model calling method, device and electronic device for target detection

Abstract

The present application provides a model calling method, apparatus, electronic device, computer-readable storage medium and computer program product for object detection; the method includes: presenting a model calling interface, and presenting a screening function item for a model for object detection in the model calling interface; in response to a model screening operation triggered based on the screening function item, presenting call entrances for at least one alternative model and model information of each alternative model, where the model information includes at least the model usage status of the alternative model; wherein the alternative model is obtained by screening based on at least two screening dimensions; in response to a trigger operation for the call entrance of the target model among at least one alternative model, when the model usage status of the target model is a normal status, presenting an object detection page corresponding to the target model. Through the present application, more model information can be presented flexibly and orderly.

Description

Technical Field

[0001] The present application relates to computer technology, and in particular, to a model calling method, device, and electronic device for target detection. Background Art

[0002] In the process of service-ifying a target detection model as the core, since the provided target detection models are divided by multiple target types, such as ship recognition, aircraft recognition, oil tank recognition, vehicle recognition, etc., and there are also significant differences between models for the same target type, such as model recognition for pictures of different precisions. Usually, the approach is to train a model for a specific target and specific precision to adapt to the corresponding recognition task. If there are M types of targets and N types of precisions, M * N models are required.

[0003] For the target detection model service provided by the related art, to avoid information clutter, the usual approach is to directly present the target detection type information of the model to the user for the user to select the corresponding model for target detection. The amount of model information presented in this service method is relatively small. Summary of the Invention

[0004] Embodiments of the present application provide a model calling method, device, electronic device, computer-readable storage medium, and computer program product for target detection, which can flexibly and orderly present more model information.

[0005] The technical solution of the embodiments of the present application is implemented as follows:

[0006] Embodiments of the present application provide a model calling method for target detection, including:

[0007] Present a model calling interface, and in the model calling interface, present a screening function item for the model for target detection;

[0008] Wherein, the screening function item is used to implement screening of the model from at least two screening dimensions;

[0009] In response to a model screening operation triggered based on the screening function item, present a call entry for at least one alternative model and the model information of each alternative model, where the model information includes at least the model usage status of the alternative model;

[0010] Wherein, the alternative model is screened based on the at least two screening dimensions;

[0011] In response to a trigger operation on the call entry of the target model among the at least one alternative model, when the model usage status of the target model is in a normal state, present a target detection page corresponding to the target model;

[0012] Among them, the target detection page is used to call the target model to implement corresponding target detection.

[0013] An embodiment of the present application provides a model calling device for target detection, including:

[0014] An interface presentation module, configured to present a model calling interface, and in the model calling interface, present a screening function item for a model used for target detection;

[0015] Among them, the screening function item is used to implement screening of the model from at least two screening dimensions;

[0016] A model information presentation module, configured to present a call entry for at least one alternative model and model information of each of the alternative models in response to a model screening operation triggered based on the screening function item, where the model information at least includes the model usage status of the alternative model;

[0017] Among them, the alternative model is obtained by screening based on the at least two screening dimensions;

[0018] A page presentation module, configured to present a target detection page corresponding to the target model when the model usage status of the target model is in a normal state in response to a trigger operation on the call entry of the target model among the at least one alternative model;

[0019] Among them, the target detection page is used to call the target model to implement corresponding target detection.

[0020] In the above solution, the interface presentation module is further configured to present a confirmation function item and screening function items for at least two screening conditions included in each of the screening dimensions in the model calling interface, and each screening function item corresponds to a screening condition; correspondingly, the model calling device for target detection further includes: a model screening module, configured to respectively receive a screening condition selection operation triggered based on the corresponding at least two screening conditions for each of the screening dimensions, and use the selected screening conditions as the screening conditions for the corresponding screening dimensions; based on the selected screening conditions for each of the screening dimensions, in response to a confirmation instruction triggered based on the confirmation function item, trigger the model screening operation to perform alternative model screening based on the selected screening conditions for each of the screening dimensions.

[0021] In the above solution, the interface presentation module is further configured to present a model call interface, and in the model call interface, present a filtering function item corresponding to at least two filtering conditions included in the i-th filtering dimension among the at least two filtering dimensions; where i is a positive integer, and i ∈ [1, N - 1], and N is the number of filtering dimensions; upon receiving a filtering condition selection operation for the i-th filtering dimension triggered based on the at least two filtering conditions, use the selected filtering condition as the filtering condition for the i-th filtering dimension; present a filtering condition selection interface for the (i + 1)-th filtering dimension, and in the filtering condition selection interface, present a filtering function item corresponding to at least two filtering conditions included in the (i + 1)-th filtering dimension; upon receiving a filtering condition selection operation for the (i + 1)-th filtering dimension, when the value of i + 1 is the same as the value of N, trigger the model filtering operation to perform alternative model filtering based on the filtering conditions selected for each of the filtering dimensions.

[0022] In the above solution, the model filtering module is further configured to, in response to a trigger operation for the filtering function item, present a condition selection list for the filtering function item, and present at least two selection function items in the condition selection list, each selection function item corresponding to a filtering condition, and each filtering condition corresponding to the at least two filtering dimensions; in response to a selection operation for a target selection function item among the at least two selection function items, use the selected filtering condition as the filtering condition for the at least two filtering dimensions, and trigger the model filtering operation to perform alternative model filtering based on the selected filtering condition.

[0023] In the above solution, the interface presentation module is further configured to present a confirmation function item and a filtering function item corresponding to at least two filtering conditions included in each of the filtering dimensions in the model call interface, each filtering function item corresponding to a filtering dimension; correspondingly, the model filtering module is further configured to, in response to a trigger operation for the filtering function item, present a condition selection list for the filtering function item, and present at least two selection function items in the condition selection list, each selection function item corresponding to a filtering condition; in response to a selection operation for a target selection function item among the at least two selection function items, use the selected filtering condition as the filtering condition for the corresponding filtering dimension; based on the filtering conditions selected for each of the filtering dimensions, in response to a confirmation instruction triggered based on the confirmation function item, trigger the model filtering operation to perform alternative model filtering based on the filtering conditions selected for each of the filtering dimensions.

[0024] In the above solution, the model information presentation module is further configured to obtain the user permissions of the user corresponding to the currently logged-in account; in response to a model screening operation triggered based on the screening function item, when the user permissions of the user are ordinary user permissions, the number of alternative models is at least two, and the model usage statuses corresponding to at least two of the alternative models include a normal status and an abnormal status, present the call entry and corresponding model information of the alternative models in the normal status among the at least two alternative models.

[0025] In the above solution, the model information presentation module is further configured to obtain the user permissions of the user corresponding to the currently logged-in account; in response to a model screening operation triggered based on the screening function item, when the user permissions of the user are administrator permissions, the number of alternative models is at least two, and the model usage statuses corresponding to at least two of the alternative models include a normal status and an abnormal status, present the call entry and corresponding model information of the alternative models in the normal status in a first presentation style, and present the call entry and corresponding model information of the alternative models in the abnormal status in a second presentation style.

[0026] In the above solution, the model call device for object detection further includes: a model status update module, configured to, when the model usage status of the target model is in the internal test status, present the object detection page corresponding to the target model in the internal test status; obtain the object detection page corresponding to the target model based on the internal test status, and the object detection result obtained by calling the target model for object detection; determine the detection accuracy of the target model according to the object detection result, and the detection accuracy is used to update the usage status of the target model. The model information presentation module is further configured to obtain the user permissions of the user corresponding to the currently logged-in account; in response to a model screening operation triggered based on the screening function item, when the user permissions of the user are internal test user permissions and the model usage status corresponding to the alternative model includes the internal test status, present the call entry and corresponding model information of the alternative models in the internal test status.

[0027] In the above solution, the model call device for object detection further includes: a model management module, configured to present a model management interface, and in the model management interface, present a status change function item and the model for object detection; in response to a usage status change operation for a first model triggered based on the status change function item, when the user permissions of the user corresponding to the currently logged-in account are administrator permissions, change the model usage status of the first model.

[0028] In the above solution, the model management module is further configured to present a model management interface, and in the model management interface, present a routing update function item and the model for object detection; in response to an update operation on the routing of the second model triggered based on the routing update function item, when the user permission of the user corresponding to the currently logged-in account is the administrator permission, update the routing of the second model; wherein, the update includes routing modification and routing deletion.

[0029] In the above solution, the model management module is further configured to present a model management interface, and in the model management interface, present an information import function item and the routing information of the model for object detection; in response to an import operation on new routing information triggered based on the information import function item, when the user permission of the user corresponding to the currently logged-in account is the administrator permission, change the routing information to the new routing information, and based on the new routing information, update the routing of the model for object detection; wherein, the update includes routing modification, routing deletion and routing addition.

[0030] In the above solution, the model calling device for object detection further includes: an object detection module, configured to present an image upload function item on the object detection page; in response to an image upload operation triggered based on the image upload function item, obtain the uploaded image to be detected; call the target model to perform object detection on the image to be detected, and when the object detection is completed, present the detection result of the corresponding object detection.

[0031] An embodiment of the present application provides an electronic device, including:

[0032] A memory, configured to store executable instructions;

[0033] A processor, configured to implement the model calling method for object detection provided by the embodiment of the present application when executing the executable instructions stored in the memory.

[0034] An embodiment of the present application provides a computer-readable storage medium, storing executable instructions, which are used to cause a processor to implement the model calling method for object detection provided by the embodiment of the present application when executed.

[0035] An embodiment of the present application provides a computer program product, including a computer program, characterized in that the computer program implements the model calling method for object detection provided by the embodiment of the present application when executed by a processor.

[0036] The embodiment of the present application has the following beneficial effects:

[0037] Compared with the service method of providing target detection type information of the model for users to select the model in the prior art, in the embodiment of the present application, through the model call interface, a screening function item for the model for target detection is presented, and in response to the model screening operation triggered based on the screening function item, a call entry for at least one alternative model and model information including the model usage status are presented. In response to the trigger operation for the call entry of the target model among the at least one alternative models, when the model usage status of the target model is in a normal state, a target detection page corresponding to the target model is presented, so as to provide a target detection model service for users. By setting the screening function item, it is possible to avoid information interference caused by too much messy information presented on the interface at the same time, thereby overcoming the defect of less model information presented in the prior art and realizing the flexible and orderly presentation of more model information. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 FIG. is an alternative structural schematic diagram of a model call system for target detection provided by an embodiment of the present application;

[0039] Figure 2 FIG. is an alternative structural schematic diagram of an electronic device provided by an embodiment of the present application;

[0040] Figure 3 FIG. is an alternative flowchart of a model call method for target detection provided by an embodiment of the present application;

[0041] Figure 4 FIG. is an alternative schematic diagram of a model call interface provided by an embodiment of the present application;

[0042] Figure 5 FIG. is an alternative schematic diagram of a model call interface provided by an embodiment of the present application;

[0043] Figure 6 FIG. is an alternative schematic diagram of a model call interface provided by an embodiment of the present application;

[0044] Figure 7 FIG. is an alternative schematic diagram of a model call interface provided by an embodiment of the present application;

[0045] Figure 8 FIG. is an alternative interface schematic diagram of presenting alternative models provided by an embodiment of the present application;

[0046] Figures 9A-9D FIG. is an alternative interface schematic diagram of a target detection page provided by an embodiment of the present application;

[0047] Figure 10 FIG. is an alternative schematic diagram of a model management interface provided by an embodiment of the present application;

[0048] Figure 11 is an optional schematic diagram of the model management interface provided by an embodiment of the present application;

[0049] Figure 12 is an optional schematic diagram of the model information corresponding to the routing table of the model provided by an embodiment of the present application;

[0050] Figure 13 is an optional flowchart of the model calling method for target detection provided by an embodiment of the present application;

[0051] Figure 14 is an optional interface schematic diagram of the model calling interface provided by an embodiment of the present application. Detailed implementation manners

[0052] In order to make the objectives, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the accompanying drawings. The described embodiments should not be construed as limitations on the present application. 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 application.

[0053] In the following description, reference is made to "some embodiments", which describe a subset of all possible embodiments. However, it can be understood that "some embodiments" can be the same subset or different subsets of all possible embodiments, and can be combined with each other without conflict.

[0054] In the following description, the terms "first / second / third" are only used to distinguish similar objects and do not represent a specific order for the objects. It can be understood that "first / second / third" can be interchanged with a specific order or sequence when allowed, so that the embodiments of the present application described herein can be implemented in an order other than that illustrated or described herein.

[0055] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of this application and are not intended to limit this application.

[0056] Before further elaborating on the embodiments of the present application, the nouns and terms involved in the embodiments of the present application are described. The nouns and terms involved in the embodiments of the present application are subject to the following explanations.

[0057] 1) Target detection, also known as target extraction, is an image segmentation technology based on the geometric and statistical features of the target. It combines the segmentation and recognition of the target into one, that is, finding all the objects of interest in the image, including two subtasks of object localization and object classification, and determining the category and location of the object at the same time.

[0058] 2) Ground sample distance (GSD), also known as ground sampling interval, in photogrammetry and remote sensing, refers to the pixel size represented by ground distance units in a digital image, with the unit of meters / pixel.

[0059] The embodiments of the present application provide a model calling method, device, electronic device, computer-readable storage medium, and computer program product for target detection, which can flexibly and orderly present more model information.

[0060] First, the model calling system for target detection provided by the embodiments of the present application will be described. Refer to Figure 1 , Figure 1 FIG. 11 is an optional schematic architecture diagram of the model calling system 100 for target detection provided by the embodiments of the present application. The terminal 400 is connected to the server 200 through the network 300. The network 300 can be a wide area network, a local area network, or a combination of the two, and uses a wireless link to implement data transmission. In some embodiments, the terminal 400 can be a laptop computer, a tablet computer, a desktop computer, a smart phone, a dedicated messaging device, a portable gaming device, a smart speaker, a smart watch, etc., but is not limited thereto. The server 200 can be an independent physical server, a server cluster or a distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. The network 300 can be a wide area network, a local area network, or a combination of the two. The terminal 400 and the server 200 can be directly or indirectly connected through wired or wireless communication methods, which are not limited in the embodiments of the present application.

[0061] The terminal 400 is configured to, in response to a startup operation for the target detection client, start the target detection client, present a login interface within the target detection client, and present login function items within the login interface. In response to a login operation triggered based on the login function items, obtain the login account and login password carried by the login operation, and send the login account and login password to the server 200, so that the server 200 verifies the login account and login password. When the verification passes, return a login instruction; in response to the login instruction, present a model invocation interface. In the model invocation interface, present function items for screening models for target detection. For each screening dimension, respectively receive a screening condition selection operation triggered based on at least two corresponding screening conditions, use the selected screening conditions as the screening conditions for the corresponding screening dimension, and based on the selected screening conditions for each of the screening dimensions, in response to a confirmation instruction triggered based on the confirmation function item, trigger the model screening operation, and send a model screening instruction to the server 200.

[0062] The server 200 is configured to, in response to the model screening instruction, perform alternative model screening for the screening conditions of each of the screening dimensions, and send at least one alternative model obtained by the screening to the terminal 400.

[0063] The terminal 400 is further configured to, in response to a trigger operation for the invocation entry of the target model among the at least one alternative models, when the model usage status of the target model is in a normal state, present a target detection page corresponding to the target model, and present an image upload function item in the target detection page. In response to an image upload operation triggered based on the image upload function item, obtain the uploaded image to be detected and send it to the server 200.

[0064] The server 200 is further configured to call the target model to perform target detection on the image to be detected. When the target detection ends, send the detection result of the corresponding target detection to the terminal 400, so that the terminal 400 presents the detection result of the target detection within the target detection page.

[0065] See Figure 2 , Figure 2 FIG. is an optional structural schematic diagram of the electronic device 500 provided by an embodiment of the present application. See Figure 2 , Figure 2 FIG. is an optional schematic diagram of the structure of the electronic device 500 provided by an embodiment of the present application. In practical applications, the electronic device 500 may be implemented as Figure 1 the terminal 400 or the server 200 in Figure 1 Taking the server 200 shown in FIG. as an example, the electronic device for the model invocation method for target detection according to the embodiment of the present application will be described. Figure 2The illustrated electronic device 500 includes: at least one processor 510, a memory 550, at least one network interface 520, and a user interface 530. Each component in the electronic device 500 is coupled together through a bus system 540. It can be understood that the bus system 540 is used to implement the connection and communication between these components. In addition to the data bus, the bus system 540 also includes a power bus, a control bus, and a status signal bus. However, for the sake of clear illustration, in Figure 2 all kinds of buses are labeled as the bus system 540.

[0066] The processor 510 can be an integrated circuit chip with signal processing capabilities, such as a general-purpose processor, a digital signal processor (DSP, Digital Signal Processor), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Among them, the general-purpose processor can be a microprocessor or any conventional processor, etc.

[0067] The user interface 530 includes one or more output devices 531 that enable the presentation of media content, including one or more speakers and / or one or more visual display screens. The user interface 530 also includes one or more input devices 532, including user interface components that facilitate user input, such as a keyboard, a mouse, a microphone, a touch screen display, a camera, other input buttons, and controls.

[0068] The memory 550 can be removable, non-removable, or a combination thereof. Exemplary hardware devices include solid-state memory, hard disk drives, optical disc drives, etc. The memory 550 optionally includes one or more storage devices that are physically located away from the processor 510.

[0069] The memory 550 includes volatile memory or non-volatile memory, and can also include both volatile and non-volatile memory. The non-volatile memory can be a read-only memory (ROM, Read Only Memory), and the volatile memory can be a random access memory (RAM, Random Access Memory). The memory 550 described in the embodiments of the present application is intended to include any suitable type of memory.

[0070] In some embodiments, the memory 550 is capable of storing data to support various operations. Examples of such data include programs, modules, and data structures, or subsets or supersets thereof, which are illustrated below.

[0071] An operating system 551, including system programs for processing various basic system services and performing hardware-related tasks, such as a framework layer, a core library layer, a driver layer, etc., for implementing various basic services and processing hardware-based tasks;

[0072] A network communication module 552 for reaching other computing devices via one or more (wired or wireless) network interfaces 520. Exemplary network interfaces 520 include: Bluetooth, Wireless Fidelity (WiFi), and Universal Serial Bus (USB), etc.;

[0073] A presentation module 553 for enabling the presentation of information (e.g., a user interface for operating peripheral devices and displaying content and information) via one or more output devices 531 associated with the user interface 530 (such as a display screen, a speaker, etc.);

[0074] An input processing module 554 for detecting and translating one or more user inputs or interactions from one of one or more input devices 532.

[0075] In some embodiments, the model calling device for object detection provided by the embodiments of the present application can be implemented in software. Figure 2 Shown is a model calling device 555 for object detection stored in the memory 550, which can be software in the form of a program and a plugin, etc., including the following software modules: an interface presentation module 5551, a model information presentation module 5552, and a page presentation module 5553. These modules are logical, so they can be combined arbitrarily or further split according to the functions implemented. The functions of each module will be described below.

[0076] In other embodiments, the model calling device for object detection provided by the embodiments of the present application can be implemented in hardware. As an example, the model calling device for object detection provided by the embodiments of the present application can be a processor in the form of a hardware decoding processor, which is programmed to execute the model calling method for object detection provided by the embodiments of the present application. For example, a processor in the form of a hardware decoding processor can adopt one or more Application Specific Integrated Circuits (ASICs), DSPs, Programmable Logic Devices (PLDs), Complex Programmable Logic Devices (CPLDs), Field-Programmable Gate Arrays (FPGAs), or other electronic components.

[0077] The model calling method for object detection provided by the embodiments of the present application will be described in combination with the exemplary applications and implementations of the terminals provided by the embodiments of the present application. Refer to Figure 3 ,Figure 3 is an alternative flowchart of the model calling method for object detection provided by an embodiment of the present application, which will be described in conjunction with Figure 3 the steps shown.

[0078] Refer to Figure 3 , Figure 3 is an alternative flowchart of the model calling method for object detection provided by an embodiment of the present application, which will be described in conjunction with Figure 3 the steps shown.

[0079] Step 101, the terminal presents a model calling interface, and in the model calling interface, presents a screening function item for the model for object detection. Among them, the screening function item is used to implement the screening of the model from at least two screening dimensions.

[0080] In actual implementation, when object detection needs to be performed, the user can trigger the startup operation of the object detection client based on the input device of the terminal. The terminal responds to the startup operation for the object detection client, starts the object detection client, and presents a model calling interface in the object detection client. Among them, a screening function item for screening the model for object detection is presented in the model calling interface. In actual implementation, at least two screening dimensions of the model can be screened with one screening function item, or the condition selection for each screening dimension can be completed respectively through the screening function items corresponding to the screening dimensions one by one, so as to jointly implement the screening of the model.

[0081] In some embodiments, based on Figure 3 , step 101 can also be implemented in the following manner: the terminal presents a model calling interface, and in the model calling interface, presents a confirmation function item and screening function items for at least two screening conditions included in each of the screening dimensions, and each screening function item corresponds to one screening condition. Correspondingly, the terminal can also execute: for each of the screening dimensions, the terminal respectively receives a screening condition selection operation triggered based on the corresponding at least two screening conditions, and uses the selected screening conditions as the screening conditions for the corresponding screening dimensions; based on the selected screening conditions for each of the screening dimensions, in response to a confirmation instruction triggered based on the confirmation function item, trigger the model screening operation to perform alternative model screening based on the selected screening conditions for each of the screening dimensions.

[0082] Exemplarily, refer to Figure 4 , Figure 4It is an optional schematic diagram of the model call interface provided by an embodiment of the present application. Here, at least two screening function items corresponding to screening dimensions are presented in the model call interface. The at least two screening dimensions include model type, picture GSD supported by the model, and model accuracy (picture accuracy supported by the model). Each screening dimension corresponds to at least two screening function items, and each screening function item corresponds to a screening condition. The terminal obtains the selected screening condition in response to a selection operation based on the screening function item. After the terminal obtains the screening conditions for each screening dimension, in response to a confirmation operation for the confirmation function item presented in the model call interface, the terminal triggers a model screening operation, and the terminal then performs alternative model screening based on the selected screening conditions for each screening dimension. It should be noted that Figure 4 The presented style of the screening function item is text and a circle pattern. The text is used to prompt the screening condition corresponding to the screening function item, and the circle pattern is used to implement the selection of the screening function item. When the screening function item is selected, a dot appears in the center of the circle pattern. When the screening function item is not selected, the circle pattern is a hollow circle graph. In some embodiments, other styles can also be used to implement the screening function item, and the embodiments of the present application do not make specific limitations on this.

[0083] In some embodiments, based on Figure 3 , step 101 can also be implemented in the following manner: The terminal presents a model call interface, and in the model call interface, presents screening function items corresponding to at least two screening conditions included in the i-th screening dimension among the at least two screening dimensions; where i is a positive integer, and i ∈ [1, N - 1], and N is the number of screening dimensions. Correspondingly, the terminal can also execute: The terminal receives a screening condition selection operation for the i-th screening dimension triggered based on the at least two screening conditions, and uses the selected screening condition as the screening condition for the i-th screening dimension; presents a screening condition selection interface for the (i + 1)-th screening dimension, and in the screening condition selection interface, presents screening function items corresponding to at least two screening conditions included in the (i + 1)-th screening dimension; receives a screening condition selection operation for the (i + 1)-th screening dimension. When the value of i + 1 is the same as the value of N, the terminal triggers the model screening operation to perform alternative model screening based on the selected screening conditions for each of the screening dimensions.

[0084] In the embodiments of the present application, the terminal presents the filtering function items of each filtering dimension in a page-turning manner. Specifically, after presenting the model call interface, the terminal presents the filtering function items corresponding to the first filtering dimension among at least two filtering dimensions in the model call interface. It should be noted that one filtering dimension corresponds to multiple filtering conditions. Here, one filtering condition corresponds to one filtering function item. In response to the selection operation for the filtering function item of the first filtering dimension, the terminal takes the selected filtering condition as the filtering condition of the first filtering dimension and presents the filtering condition selection interface of the second filtering dimension to select the filtering conditions of the next filtering dimension. Here, the terminal can switch from the currently presented model call interface to the filtering condition selection interface of the next filtering dimension in a page-turning manner, or cancel the filtering function items presented in the currently presented model call interface and present the filtering function items of the next filtering dimension. The embodiments of the present application do not specifically limit the switching form of the filtering function items of different dimensions.

[0085] Exemplarily, referring to Figure 5 , Figure 5 is an optional schematic diagram of the model call interface provided by the embodiments of the present application. As Figure 5 shown, the order of each filtering dimension is model type, GPS, and model accuracy. Here, the terminal first presents the filtering function items corresponding to the first filtering dimension in the model call interface, that is, the filtering function items corresponding to the model type. Here, the model type includes three types: road network detection, vehicle detection, and oil tank detection, and three filtering function items are correspondingly presented. In response to the selection operation for the filtering function item, the terminal automatically switches to the filtering condition selection interface corresponding to the next filtering dimension, and presents the filtering function items corresponding to the next filtering dimension in the filtering condition selection interface. Here, the second filtering dimension is GSD, and there are three filtering function items corresponding to the filtering conditions of 0.3 dpi, 0.7 dpi, and 1.0 dpi. When the currently presented filtering condition selection interface is the last filtering dimension among all filtering dimensions, the terminal, in response to the selection operation for the filtering function item of the current filtering dimension, obtains the filtering condition of the current filtering dimension and automatically triggers the model filtering operation to filter the alternative models based on the selected filtering conditions for each filtering dimension.

[0086] In some embodiments, based on Figure 3The terminal can also perform the following: in response to a trigger operation on the filtering function item, the terminal presents a conditional selection list of the filtering function item, and presents at least two selection function items in the conditional selection list. Each selection function item corresponds to a filtering condition, and each filtering condition corresponds to the at least two filtering dimensions; in response to a selection operation on a target selection function item among the at least two selection function items, the selected filtering condition is used as the filtering condition for the at least two filtering dimensions, and the model filtering operation is triggered to perform alternative model filtering based on the selected filtering condition.

[0087] Here, the conditional selection list can be presented in the form of the following list, or in the form of a pull-up list, or in a floating form, etc. The present application does not specifically limit the presentation form and presentation position of the conditional selection list. Exemplarily, referring to Figure 6 , Figure 6 is an optional schematic diagram of the model calling interface provided by an embodiment of the present application. Here, a filtering function item corresponding to three filtering dimensions including model type, GSD, and model accuracy is simultaneously presented in the model calling interface. The filtering function item is used to implement conditional filtering for all filtering dimensions. In actual implementation, when the user triggers the filtering function item, a conditional selection list corresponding to the filtering function item is presented. Here, at least two selection function items are presented in the conditional selection list. Each selection function item corresponds to a filtering condition. The filtering condition is, for example, "road network detection, 0.7 dpi, 3.0 m", which includes sub-filtering conditions of a single dimension corresponding to all filtering dimensions, and the sub-filtering conditions of each filtering dimension jointly form the filtering condition. In the embodiment of the present application, the terminal directly triggers the model filtering operation to filter the alternative models in response to the selection operation on the target selection function item. In some embodiments, a confirmation function item is also presented in the model calling interface. The terminal obtains the filtering conditions for the at least two filtering dimensions in response to the selection operation, and triggers the model filtering operation in response to the confirmation operation on the confirmation function item.

[0088] In some embodiments, based on Figure 3, Step 101 can also be executed in the following manner: The terminal presents a confirmation function item and screening function items for at least two screening conditions included in each of the screening dimensions in the model call interface, and each of the screening function items corresponds to one screening dimension. Correspondingly, the terminal can also execute: The terminal respectively performs the following processing for each of the screening function items: In response to a trigger operation for the screening function item, the terminal presents a condition selection list for the screening function item, and presents at least two selection function items in the condition selection list, and each of the selection function items corresponds to one screening condition; In response to a selection operation for a target selection function item among the at least two selection function items, the selected screening condition is used as the screening condition for the corresponding screening dimension; Based on the screening conditions for each of the screening dimensions selected, in response to a confirmation instruction triggered based on the confirmation function item, the terminal triggers the model screening operation to perform alternative model screening based on the screening conditions for each of the screening dimensions selected.

[0089] In the embodiment of the present application, the terminal presents screening function items corresponding one by one to at least two screening dimensions in the model call interface, the number of screening function items is the same as the number of screening dimensions, and a confirmation function item is presented within the model call interface. In actual implementation, the user can trigger the screening function items corresponding to each screening dimension in any order. The terminal, in response to a trigger operation for a certain screening function item triggered by the user, presents a condition selection list corresponding to the screening function item, and presents at least two screening conditions corresponding to the screening dimension within the condition selection list.

[0090] In actual implementation, the terminal can, in response to a confirmation operation for the confirmation function item, that is, trigger the model screening operation. Here, if the user does not select the screening conditions corresponding to a certain screening dimension, the terminal defaults that all screening conditions of this screening dimension are selected. If no screening conditions are selected for all screening dimensions and the confirmation function item is triggered for a confirmation operation, the terminal screens out all models.

[0091] Exemplarily, referring to Figure 7 , Figure 7 is an optional schematic diagram of the model call interface provided by the embodiment of the present application. There are screening function items corresponding to three screening dimensions presented within the model call interface. For any one of the screening function items, the terminal presents a corresponding condition selection list in response to a trigger operation for the screening function item, and the user can then select the corresponding screening conditions for the condition selection list. By selecting the screening conditions for each screening dimension one by one, the terminal obtains the screening conditions corresponding to each screening dimension respectively. After the confirmation function item is triggered, the terminal triggers the model screening operation.

[0092] Step 102: In response to a model screening operation triggered based on the screening function item, present the call entry of at least one alternative model and the model information of each of the alternative models. The model information includes at least the model usage status of the alternative models, and the alternative models are obtained by screening based on the at least two screening dimensions.

[0093] Here, the call entry of the alternative model can be presented in any form. For example, the call path of the model can be bound to a graphic, and this graphic is used as the call entry of the alternative model. The user can trigger the call operation of the alternative model by triggering this graphic. Here, the model information can include the model usage status of the model. The model usage status can be, for example, the normal status, the internal test status, or the offline status, etc. Exemplarily, referring to Figure 5 , Figure 5 the shown alternative models include Model 1 and Model 3. Here, a rectangular graphic is bound to the call path of the alternative model, and the bound rectangle is used as the call entry of the model. Here, the terminal also presents the model information of the model below the alternative model. Here, the model usage statuses of Model 1 and Model 3 are both in the normal status. By presenting the model information, it can intuitively prompt the usage status of the model or other information of the model, such as the version information, for the user to further select the alternative model.

[0094] In some embodiments, based on Figure 3 , Step 102 can also be implemented in the following manner: The terminal obtains the user permissions of the user corresponding to the currently logged-in account; in response to a model screening operation triggered based on the screening function item, when the user permissions of the user are ordinary user permissions, the number of alternative models is at least two, and the model usage statuses corresponding to at least two of the alternative models include the normal status and the abnormal status, present the call entry of the alternative model in the normal status among the at least two alternative models and the corresponding model information.

[0095] It should be noted that the logged-in account is the user account required to log in to the target detection client. When the user uses the target detection client, the user first enters the logged-in account and the corresponding account password to log in. After successful login, the terminal presents a model call interface in the target detection client. In this application, the user permissions of the logged-in account include ordinary user permissions, internal test user permissions, and administrator permissions. When the user permissions of the logged-in account are ordinary user permissions, the terminal screens alternative models among the models in the normal status in response to a model screening operation triggered based on a user operation. When the user permissions of the logged-in user are internal test user permissions, alternative models are screened among the models in the normal status and the internal test status. When the user permissions of the logged-in user are administrator permissions, alternative models are screened among the models in the normal status, the internal test status, and the offline status.

[0096] In some embodiments, based on Figure 3 , step 102 can also be implemented in the following manner: The terminal obtains the user permissions of the user corresponding to the currently logged-in account; in response to a model screening operation triggered based on the filtering function item, when the user permissions of the user are administrator permissions, the number of alternative models is at least two, and the model usage statuses corresponding to at least two of the alternative models include a normal status and an abnormal status, the call entry of the alternative models in the normal status and the corresponding model information are presented in a first presentation style, and the call entry of the alternative models in the abnormal status and the corresponding model information are presented in a second presentation style.

[0097] In actual implementation, when the user permissions of the user corresponding to the logged-in account are administrator permissions, all alternative models with different model usage statuses are presented. If the model usage statuses of the filtered alternative models include both a normal status and an abnormal status, the alternative models with different model usage statuses are presented in different presentation styles. Exemplarily, referring to Figure 8 , Figure 8 is an optional interface schematic diagram for presenting alternative models provided by an embodiment of the present application. Here, the terminal presents alternative models with different model usage statuses in images of different colors. In some embodiments, the abnormal status can also include multiple statuses, such as an internal test status or a deactivated status. Here, models in the internal test status are only provided for trial use by some users, such as users with internal test user permissions or users with administrator permissions, and the deactivated status indicates that the model is no longer available.

[0098] Step 103, in response to a trigger operation on the call entry of the target model among the at least one alternative model, when the model usage status of the target model is in the normal status, a target detection page corresponding to the target model is presented. Wherein, the target detection page is used to call the target model to implement corresponding target detection.

[0099] In some embodiments, based on Figure 3 , the terminal can also execute: The terminal responds to an image upload operation triggered based on the image upload function item, obtains the uploaded image to be detected; calls the target model to perform target detection on the image to be detected, and when the target detection is completed, presents the detection result of the corresponding target detection.

[0100] In actual implementation, the presentation style of the call entry of the alternative model is as Figure 5 or Figure 8As shown, in the embodiment of the present application, when the terminal responds to a trigger operation for the call entry of the alternative model, if the model usage status of the alternative model targeted by the trigger operation is the normal status, the terminal presents a target detection page in the display interface. Exemplarily, refer to Figures 9A-9D , Figures 9A-9D which is an optional interface schematic diagram of the target detection page provided by the embodiment of the present application. The terminal first presents an image selection function item in the target detection page, such as Figure 9A the function item with the text "Click to select an image" shown. The terminal responds to the click operation on the click-to-select-image, presents the corresponding image selection interface, and selects an image from the image selection interface for upload. Here, the uploaded image can be selected from the local of the terminal. As Figure 9B shown, after the to-be-detected image is successfully uploaded, the uploaded to-be-detected image can be presented in the target detection page, and a detection function item is presented in the target detection page, such as the function item with the text "Start detection" shown. The terminal responds to the trigger operation on the detection function item and performs target detection on the uploaded to-be-detected image. Here, the uploaded to-be-detected image is a remote sensing image of a road, and there are several vehicles parked on the road. Then, the terminal performs corresponding target detection on the to-be-detected image based on the selected alternative model. Here, the model type of the selected alternative model is vehicle detection, and the terminal uses the vehicle detection model to perform vehicle detection on the to-be-detected image, identifies the vehicles in the to-be-detected image, marks the recognition results, generates corresponding recognition results and outputs them to the target detection page for display, such as Figure 9D the detection results shown.

[0101] In some embodiments, based on Figure 3 , step 102 can also be implemented in the following manner: The terminal obtains the user permissions of the user corresponding to the currently logged-in account; in response to the model screening operation triggered based on the screening function item, when the user's user permissions are internal test user permissions and the model usage status corresponding to the alternative model includes the internal test status, the call entry of the alternative model in the internal test status and the corresponding model information are presented. Correspondingly, the terminal can also execute: when the model usage status of the target model is the internal test status, the terminal presents the target detection page corresponding to the target model in the internal test status; obtains the target detection results obtained by calling the target model for target detection based on the target detection page corresponding to the target model in the internal test status; and determines the detection accuracy of the target model according to the target detection results, and the detection accuracy is used to update the usage status of the target model.

[0102] In actual implementation, after the terminal performs object detection on the image to be detected using the target model in the internal test state, it obtains the object detection result of the target model for the image to be detected, and further performs an accuracy detection on this object detection result. Here, the image to be detected can be manually subjected to object detection, and the recognition result manually identified is compared with this object detection result to determine the detection accuracy. In some embodiments, the image to be detected can also be input into a target detection model with a relatively high accuracy, and the recognition result of this target detection model with a relatively high accuracy is compared with the object detection result of this target model to determine the detection accuracy of the target model. If the detection accuracy is higher than the accuracy threshold, the usage status of this target model is updated to the normal state; otherwise, it is changed to the offline state, and this target model is taken offline.

[0103] In some embodiments, based on Figure 3 , the terminal can also execute: The terminal presents a model management interface, and in the model management interface, presents a status change function item and the model for object detection; in response to a usage status change operation for the first model triggered based on the status change function item, when the user permission of the user corresponding to the currently logged-in account is the administrator permission, the model usage status of the first model is changed.

[0104] Exemplarily, referring to Figure 10 , Figure 10 is an optional schematic diagram of the model management interface provided by the embodiments of the present application. Here, the model management interface presents information on different dimensions of the model in the form of a topological structure. The administrator can trigger the edit function item to start re-editing the topological connection of each model, and change information such as the usage status of the model in the way of moving arrows.

[0105] In some embodiments, referring to Figure 11 , Figure 11 is an optional schematic diagram of the model management interface provided by the embodiments of the present application. Here, the model management interface presents the routing table of the model and presents the corresponding edit function item for the administrator to edit the routing table. In actual implementation, the administrator can directly perform text editing on the routing table and submit the edited routing table based on the submit function item. In particular, in response to the submit operation for the submit function item, the usage status of each model is automatically updated according to the information of the model recorded in the updated routing table.

[0106] In some embodiments, the terminal may further perform: The terminal presents a model management interface, and in the model management interface, presents a routing update function item and the model for target detection; in response to an update operation for the routing of the second model triggered based on the routing update function item, when the user permission of the user corresponding to the currently logged-in account is an administrator permission, update the routing of the second model; wherein the update includes routing modification and routing deletion.

[0107] Exemplarily, when the model management interface is a topology structure as Figure 10 shown, the administrator can update the routing of the model by changing the topology structure. Specifically, the terminal obtains the topology structure of the reconnected model in response to the administrator's connection operation on the connection line within the model management interface, and obtains the model information of each model based on the reconnected topology structure. The terminal updates the routing of each model with the model information corresponding to the reconnected topology structure in response to the administrator's submission operation on the submission function item.

[0108] In some embodiments, exemplarily, when the model management interface is a routing table as Figure 11 shown, the administrator can update the routing information of the model by editing the routing table, and the terminal automatically updates the routing of each model based on the updated routing table in response to the submission operation of the submission function item. Refer to Figure 12 , Figure 12 which is an optional schematic diagram of the model information corresponding to the routing table of the model provided in the embodiments of the present application. After obtaining the updated routing table, the terminal automatically updates the routing of the model based on the correspondence between the model and the model information in the routing table to implement the update of the model usage status.

[0109] In some embodiments, the terminal may further perform: The terminal presents a model management interface, and in the model management interface, presents an information import function item and the routing information of the model for target detection; in response to an import operation for new routing information triggered based on the information import function item, when the user permission of the user corresponding to the currently logged-in account is an administrator permission, change the routing information to the new routing information, and update the routing of the model for target detection based on the new routing information; wherein the update includes routing modification, routing deletion, and routing addition.

[0110] Exemplarily, refer to Figure 11, an import function item is also presented in the model management interface. The administrator can trigger the import operation of the routing table through this import function item. The terminal responds to the import operation, receives the newly imported routing table, and responds to the submission operation for the submission function item, updating the routing table of the imported model to the new routing table to update the routing of the model based on the new routing table. Here, by importing the routing tables of all models with one click, simple and fast routing updates are achieved.

[0111] Next, continue to introduce the model calling method for target detection provided by the embodiments of the present application. Figure 13 is an optional flowchart of the model calling method for target detection provided by the embodiments of the present application. Refer to Figure 13 , the model calling method for target detection provided by the embodiments of the present application is jointly implemented by the terminal and the server.

[0112] Step 201, the terminal responds to the startup operation for the target detection client, starts the target detection client, and presents a login interface in the target detection client, and presents a login function item in the login interface.

[0113] Step 202, the terminal obtains the login account and login password carried by the login operation in response to the login operation triggered based on the login function item.

[0114] Step 203, the terminal sends the login account and login password to the server.

[0115] Step 204, the server verifies the login account and login password, and when the verification passes, sends a login instruction to the terminal.

[0116] Step 205, the terminal presents a model call interface in response to the login instruction, and in the model call interface, presents a screening function item for the model for target detection.

[0117] Step 206, for each screening dimension, the terminal respectively receives a screening condition selection operation triggered based on at least two corresponding screening conditions, and uses the selected screening conditions as the screening conditions for the corresponding screening dimension.

[0118] Step 207, based on the selected screening conditions for each of the screening dimensions, in response to the confirmation instruction triggered based on the confirmation function item, the terminal triggers the model screening operation and sends a model screening instruction to the server.

[0119] Step 208, the server performs alternative model screening for the screening conditions of each of the screening dimensions in response to the model screening instruction.

[0120] Step 209, the server sends at least one alternative model obtained by screening to the terminal.

[0121] Step 210, in response to a trigger operation on the invocation entry of the target model among the at least one alternative model, when the model usage status of the target model is normal, the terminal presents a target detection page corresponding to the target model.

[0122] Step 211, the terminal presents an image upload function item on the target detection page, and in response to an image upload operation triggered based on the image upload function item, obtains the uploaded image to be detected.

[0123] Step 212, the terminal sends the image to be detected to the server.

[0124] Step 213, the server invokes the target model to perform target detection on the image to be detected.

[0125] Step 214, when the target detection ends, send the detection result of the corresponding target detection to the terminal.

[0126] Step 215, the terminal presents the detection result of the target detection within the target detection page.

[0127] Next, an exemplary application of the embodiments of the present application in an actual application scenario will be described.

[0128] Refer to Figure 14 , Figure 14 which is an optional interface schematic diagram of the model invocation interface provided by the embodiments of the present application. In target detection, in response to a startup operation on the target detection client, the terminal starts the target detection client and presents the model invocation interface within the target detection client. Among them, the model invocation interface presents screening function items with at least two screening dimensions, and each screening dimension corresponds to a screening function item. Figure 14 The screening dimensions of the two screening function items shown are model type and the GSD of the pictures supported by the model. In some embodiments, the screening dimensions further include model version, etc., and in the Figure 14 shown model invocation interface, a screening function item corresponding to the screening dimension of the model version can be further presented. In the embodiments of the present application, the terminal can present a conditional selection list of the screening function items of the model type in response to a trigger operation on the screening function item corresponding to the model type. The conditional selection list presents at least one selection function item. Here, Figure 14 three selection function items are shown, and each selection function item corresponds to a screening condition, which are road network detection, vehicle detection, and oil tank detection respectively. The user can trigger the selection operation of the screening condition by clicking the corresponding selection function item, and use the selected screening condition as the screening condition for at least two screening dimensions. Exemplarily, Figure 14The selected filtering condition for the filtering dimension of model type is vehicle detection. Next, continue to select the filtering dimension of GSD to obtain the filtering condition of 0.3 dpi. It should be noted that the embodiments of the present application can select each filtering dimension in any order.

[0129] In the embodiments of the present application, after the terminal obtains the filtering conditions for each filtering dimension, it automatically filters the models available to the user to obtain the filtered alternative models. Here, the number of alternative models can be one or more. When there are multiple alternative models, the terminal can further present the filtered models in the display interface for the user to select, or can automatically select the optimal one from the multiple alternative models for invocation. In actual implementation, a target detection page is also presented in the model invocation interface. Here, the target detection page includes a first display area and a second display area. The first display area is used to display the uploaded image to be detected, and the second display area is used to display the target detection result. Exemplarily, as Figure 5 shown, the display area on the left is the first display area, in which the uploaded image to be detected has been displayed. The image to be detected is a remote sensing image of a road, and there are several vehicles parked on the road. The display area on the right is the second display area. After the terminal obtains the filtering conditions for each filtering dimension, in response to the trigger operation for the running function item, it directly and automatically performs vehicle detection on the image to be detected according to the filtered model, and displays the result of the target detection in the second display area to complete the vehicle detection of the image to be detected.

[0130] In actual implementation, for the models used for target detection, the terminal stores the routing information of each model through a routing table. Referring to Figure 11 , the routing table stores the model name, model type, model version, ground sampling distance, model path, and model usage status of the model. It should be noted that the terminal can set three user permissions for the target detection client, including ordinary user permission, internal test user permission, and administrator permission. When the user permission of the user logging in to the target detection client is ordinary user permission, the terminal filters the alternative models among the models in the normal state in response to the model filtering operation triggered by the user operation. When the user permission of the logged-in user is internal test user permission, the alternative models are filtered among the models in the normal state and the internal test state. When the user permission of the logged-in user is administrator permission, the alternative models are filtered among the models in the normal state, the internal test state, and the offline state to prompt the administrator of the usage status of each model.

[0131] In actual implementation, when the login user's permission is the administrator permission, the administrator can also enter the model management page to edit the routing table, and the terminal responds to the administrator's editing operation to add, delete, or change the model routing. In addition, the terminal can also respond to the administrator's status update operation to update the status of the model.

[0132] Next, the exemplary structure of the model calling device 555 for object detection provided in the embodiments of the present application implemented as software modules will be further described. In some embodiments, as Figure 2 shown, the software modules stored in the memory 540 in the model calling device 555 for object detection may include:

[0133] The interface presentation module 5551 is used to present the model calling interface and present the screening function items for the model for object detection in the model calling interface;

[0134] Among them, the screening function items are used to implement the screening of the model from at least two screening dimensions;

[0135] The model information presentation module 5552 is used to present the call entry of at least one alternative model and the model information of each alternative model in response to the model screening operation triggered based on the screening function items. The model information includes at least the model usage status of the alternative model;

[0136] Among them, the alternative model is screened based on the at least two screening dimensions;

[0137] The page presentation module 5553 is used to present the object detection page corresponding to the target model in response to the trigger operation for the call entry of the target model among the at least one alternative model when the model usage status of the target model is the normal status;

[0138] Among them, the object detection page is used to call the target model to implement the corresponding object detection.

[0139] In some embodiments, the interface presentation module 5551 is further configured to present a confirmation function item and screening function items for at least two screening conditions included in each of the screening dimensions in the model invocation interface, with each screening function item corresponding to a screening condition; correspondingly, the model invocation device for object detection further includes: a model screening module, configured to respectively receive, for each of the screening dimensions, a screening condition selection operation triggered based on the corresponding at least two screening conditions, and use the selected screening conditions as the screening conditions for the corresponding screening dimensions; and in response to a confirmation instruction triggered based on the confirmation function item, trigger the model screening operation to perform alternative model screening based on the selected screening conditions for each of the screening dimensions.

[0140] In some embodiments, the interface presentation module 5551 is further configured to present a model invocation interface, and in the model invocation interface, present screening function items for at least two screening conditions included in the i-th screening dimension among the at least two screening dimensions; where i is a positive integer and i ∈ [1, N - 1], and N is the number of screening dimensions; receive a screening condition selection operation for the i-th screening dimension triggered based on the at least two screening conditions, and use the selected screening conditions as the screening conditions for the i-th screening dimension; present a screening condition selection interface for the (i + 1)-th screening dimension, and in the screening condition selection interface, present screening function items for at least two screening conditions included in the (i + 1)-th screening dimension; receive a screening condition selection operation for the (i + 1)-th screening dimension, and when the value of i + 1 is the same as the value of N, trigger the model screening operation to perform alternative model screening based on the selected screening conditions for each of the screening dimensions.

[0141] In some embodiments, the model screening module is further configured to, in response to a trigger operation on the screening function item, present a condition selection list for the screening function item, and present at least two selection function items in the condition selection list, with each selection function item corresponding to a screening condition, and each screening condition corresponding to the at least two screening dimensions; in response to a selection operation on a target selection function item among the at least two selection function items, use the selected screening conditions as the screening conditions for the at least two screening dimensions, and trigger the model screening operation to perform alternative model screening based on the selected screening conditions.

[0142] In some embodiments, the interface presentation module 5551 is further configured to present a confirmation function item and screening function items for at least two screening conditions included in each of the screening dimensions in the model call interface, and each of the screening function items corresponds to one screening dimension; correspondingly, the model screening module is further configured to, in response to a trigger operation on the screening function item, present a condition selection list for the screening function item, and present at least two selection function items in the condition selection list, and each of the selection function items corresponds to one screening condition; in response to a selection operation on a target selection function item among the at least two selection function items, use the selected screening condition as the screening condition for the corresponding screening dimension; based on the screening conditions for each of the screening dimensions selected, in response to a confirmation instruction triggered based on the confirmation function item, trigger the model screening operation to perform alternative model screening based on the screening conditions for each of the screening dimensions selected.

[0143] In some embodiments, the model information presentation module 5552 is further configured to obtain the user permissions of the user corresponding to the currently logged-in account; in response to a model screening operation triggered based on the screening function item, when the user permissions of the user are ordinary user permissions, the number of alternative models is at least two, and the model usage statuses corresponding to the at least two alternative models include a normal status and an abnormal status, present a call entry for the alternative models in the normal status and the corresponding model information among the at least two alternative models.

[0144] In some embodiments, the model information presentation module 5552 is further configured to obtain the user permissions of the user corresponding to the currently logged-in account; in response to a model screening operation triggered based on the screening function item, when the user permissions of the user are administrator permissions, the number of alternative models is at least two, and the model usage statuses corresponding to the at least two alternative models include a normal status and an abnormal status, present the call entry for the alternative models in the normal status and the corresponding model information in a first presentation style, and present the call entry for the alternative models in the abnormal status and the corresponding model information in a second presentation style.

[0145] In some embodiments, the model information presentation module 5552 is further configured to obtain the user permissions of the user corresponding to the currently logged-in account; in response to a model screening operation triggered based on the screening function item, when the user permissions of the user are internal test user permissions and the model usage status corresponding to the alternative model includes the internal test status, present the call entry of the alternative model in the internal test status and the corresponding model information; correspondingly, the model calling device for object detection further includes: a model status update module, configured to present the object detection page corresponding to the target model in the internal test status when the model usage status of the target model is the internal test status; obtain the object detection page corresponding to the target model based on the internal test status, and the object detection result obtained by calling the target model for object detection; determine the detection accuracy of the target model according to the object detection result, and the detection accuracy is used to update the usage status of the target model.

[0146] In some embodiments, the model calling device for object detection further includes: a model management module, configured to present a model management interface, and in the model management interface, present a status change function item and the model for object detection; in response to a usage status change operation for the first model triggered based on the status change function item, when the user permissions of the user corresponding to the currently logged-in account are administrator permissions, change the model usage status of the first model.

[0147] In some embodiments, the model management module is further configured to present a model management interface, and in the model management interface, present a routing update function item and the model for object detection; in response to an update operation for the routing of the second model triggered based on the routing update function item, when the user permissions of the user corresponding to the currently logged-in account are administrator permissions, update the routing of the second model; wherein, the update includes routing modification and routing deletion.

[0148] In some embodiments, the model management module is further configured to present a model management interface, and in the model management interface, present an information import function item and the routing information of the model for object detection; in response to an import operation for new routing information triggered based on the information import function item, when the user permissions of the user corresponding to the currently logged-in account are administrator permissions, change the routing information to the new routing information, and update the routing of the model for object detection based on the new routing information; wherein, the update includes routing modification, routing deletion, and routing addition.

[0149] In some embodiments, the model invocation device for object detection further includes: an object detection module, configured to present an image upload function item on the object detection page; in response to an image upload operation triggered based on the image upload function item, obtain the uploaded image to be detected; invoke the object model to perform object detection on the image to be detected, and when the object detection ends, present the detection result of the corresponding object detection.

[0150] It should be noted that the description of the device in the embodiments of the present application is similar to the description of the above method embodiments, and has similar beneficial effects as the method embodiments, so details are not described herein.

[0151] The embodiments of the present application provide a computer-readable storage medium storing executable instructions, where the executable instructions, when executed by a processor, cause the processor to execute the method provided by the embodiments of the present application. For example, as Figure 3 shown in the model invocation method for object detection.

[0152] The embodiments of the present application provide a computer program product, including a computer program, characterized in that when the computer program is executed by a processor, it implements the model invocation method for object detection provided by the embodiments of the present application.

[0153] In some embodiments, the computer-readable storage medium may be a memory such as FRAM, ROM, PROM, EPROM, EEPROM, flash memory, magnetic surface memory, optical disc, or CD-ROM; or it may be various devices including one or any combination of the above memories.

[0154] In some embodiments, the executable instructions may be in the form of a program, software, software module, script, or code, written in any form of programming language (including compiled or interpreted languages, or declarative or procedural languages), and may be deployed in any form, including being deployed as an independent program or being deployed as a module, component, subroutine, or other unit suitable for use in a computing environment.

[0155] As an example, the executable instructions may or may not correspond to a file in the file system, and may be stored as a part of a file that stores other programs or data. For example, they may be stored in one or more scripts in a Hyper Text Markup Language (HTML) document, stored in a single file dedicated to the program being discussed, or stored in multiple cooperating files (for example, files storing one or more modules, subroutines, or code portions).

[0156] As an example, the executable instructions may be deployed to execute on one computing device, or on multiple computing devices located at one location, or on multiple computing devices distributed across multiple locations and interconnected by a communication network.

[0157] In summary, through the embodiments of the present application, more model information can be presented flexibly and orderly.

[0158] The above is only the embodiments of the present application and is not intended to limit the protection scope of the present application. Any modifications, equivalent replacements, and improvements made within the spirit and scope of the present application are included in the protection scope of the present application.

Claims

1. A method for model invocation for object detection, characterized in that, Including: Present a model call interface, and in the model call interface, present a screening function item for a model for object detection; Among them, the screening function item is used to implement the screening of models from at least two screening dimensions; In response to a model screening operation triggered based on the screening function item, present a call entry for at least one alternative model and model information of each of the alternative models, where the model information at least includes the model usage status of the alternative model; Among them, the alternative models are screened based on the at least two screening dimensions; In response to a trigger operation on the call entry of the target model among the at least one alternative model, when the model usage status of the target model is in a normal state, present a target detection page corresponding to the target model; Among them, the target detection page is used to call the target model to implement corresponding target detection.

2. The method according to claim 1, characterized in that, The presenting the model call interface and presenting a screening function item for a model for object detection in the model call interface includes: In the model call interface, present a confirmation function item and screening function items for at least two screening conditions included in each of the screening dimensions, and each screening function item corresponds to a screening condition; Correspondingly, the method further includes: For each of the screening dimensions, respectively receive a screening condition selection operation triggered based on the corresponding at least two screening conditions, and use the selected screening conditions as the screening conditions for the corresponding screening dimensions; Based on the selected screening conditions for each of the screening dimensions, in response to a confirmation instruction triggered based on the confirmation function item, trigger the model screening operation to screen alternative models based on the selected screening conditions for each of the screening dimensions.

3. The method according to claim 1, wherein The presenting the model call interface and presenting a screening function item for a model for object detection in the model call interface includes: Present a model call interface, and in the model call interface, present screening function items for at least two screening conditions included in the i-th screening dimension among the at least two screening dimensions; Among them, i is a positive integer, and i ∈ [1, N - 1], and N is the number of screening dimensions; Receive a screening condition selection operation for the i-th screening dimension triggered based on the at least two screening conditions, and use the selected screening conditions as the screening conditions for the i-th screening dimension; Present a screening condition selection interface for the (i + 1)-th screening dimension, and in the screening condition selection interface, present screening function items for at least two screening conditions included in the (i + 1)-th screening dimension; Receive a screening condition selection operation for the (i + 1)-th screening dimension. When the value of i + 1 is the same as the value of N, trigger the model screening operation to screen alternative models based on the selected screening conditions for each of the screening dimensions.

4. The method according to claim 1, characterized in that The method further includes: In response to a trigger operation on the screening function item, present a condition selection list of the screening function item, and in the condition selection list, present at least two selection function items, each selection function item corresponds to a screening condition, and each screening condition corresponds to the at least two screening dimensions; In response to a selection operation for a target selection function item among the at least two selection function items, the selected filtering condition is used as the filtering condition for the at least two filtering dimensions, and the model filtering operation is triggered to filter alternative models based on the selected filtering condition.

5. The method according to claim 1, wherein In the model calling interface, present filtering function items for models used for target detection, including: In the model calling interface, present a confirmation function item and filtering function items for at least two filtering conditions included in each of the filtering dimensions, with each filtering function item corresponding to one filtering dimension; Correspondingly, the method further includes: Perform the following processing for each of the filtering function items respectively: In response to a triggering operation for the filtering function item, present a condition selection list for the filtering function item, and present at least two selection function items in the condition selection list, with each selection function item corresponding to one filtering condition; In response to a selection operation for a target selection function item among the at least two selection function items, use the selected filtering condition as the filtering condition for the corresponding filtering dimension; Based on the selected filtering conditions for each of the filtering dimensions, in response to a confirmation instruction triggered based on the confirmation function item, trigger the model filtering operation to filter alternative models based on the selected filtering conditions for each of the filtering dimensions.

6. The method according to claim 1, wherein In response to a model filtering operation triggered based on the filtering function item, present a call entry for at least one alternative model and model information of each of the alternative models, including: Obtain the user permissions of the user corresponding to the currently logged-in account; In response to a model filtering operation triggered based on the filtering function item, when the user's user permissions are ordinary user permissions, the number of alternative models is at least two, and the model usage statuses corresponding to at least two of the alternative models include a normal status and an abnormal status, present the call entry for the alternative model in the normal status among the at least two alternative models and the corresponding model information.

7. The method according to claim 1, wherein In response to a model filtering operation triggered based on the filtering function item, present a call entry for at least one alternative model and model information of each of the alternative models, including: Obtain the user permissions of the user corresponding to the currently logged-in account; In response to a model filtering operation triggered based on the filtering function item, when the user's user permissions are administrator permissions, the number of alternative models is at least two, and the model usage statuses corresponding to at least two of the alternative models include a normal status and an abnormal status, Present the call entry for the alternative model in the normal status and the corresponding model information in a first presentation style, and present the call entry for the alternative model in the abnormal status and the corresponding model information in a second presentation style.

8. The method according to claim 1, wherein In response to a model filtering operation triggered based on the filtering function item, present a call entry for at least one alternative model and model information of each of the alternative models, including: Obtain the user permissions of the user corresponding to the currently logged-in account; In response to a model screening operation triggered based on the screening function item, when the user's user permission is an internal test user permission and the model usage status corresponding to the alternative model includes an internal test status, present a call entry for the alternative model in the internal test status and the corresponding model information; Correspondingly, the method further includes: When the model usage status of the target model is in the internal test status, present a target detection page corresponding to the target model in the internal test status; Obtain the target detection page corresponding to the target model based on the internal test status, and obtain the target detection result obtained by performing target detection on the target model; According to the target detection result, determine the detection accuracy of the target model, and the detection accuracy is used to update the usage status of the target model.

9. The method according to claim 1, characterized in that, The method further includes: Present a model management interface, and in the model management interface, present a status change function item and the model for target detection; In response to a usage status change operation for the first model triggered based on the status change function item, when the user permission corresponding to the currently logged-in account is an administrator permission, change the model usage status of the first model.

10. The method according to claim 1, characterized in that, The method further includes: Present a model management interface, and in the model management interface, present a routing update function item and the model for target detection; In response to a routing update operation for the second model triggered based on the routing update function item, when the user permission corresponding to the currently logged-in account is an administrator permission, update the routing of the second model; Wherein, the update includes routing modification and routing deletion.

11. The method according to claim 1, characterized in that The method further includes: Present a model management interface, and in the model management interface, present an information import function item and the routing information of the model for target detection; In response to an import operation for new routing information triggered based on the information import function item, when the user permission corresponding to the currently logged-in account is an administrator permission, change the routing information to the new routing information, and based on the new routing information, update the routing of the model for target detection; Wherein, the update includes routing modification, routing deletion, and routing addition.

12. The method according to claim 1, wherein The method further includes: Present an image upload function item on the target detection page; In response to an image upload operation triggered based on the image upload function item, obtain the uploaded image to be detected; Call the target model to perform target detection on the image to be detected, and when the target detection is completed, present the detection result of the corresponding target detection.

13. A model calling device for object detection, characterized in that, Includes: An interface presentation module, configured to present a model call interface, and in the model call interface, present a screening function item for the model for target detection; Wherein, the screening function item is used to implement screening of the model from at least two screening dimensions; A model information presentation module, configured to present a call entry for at least one alternative model and the model information of each alternative model in response to a model screening operation triggered based on the screening function item, and the model information includes at least the model usage status of the alternative model; Among them, the alternative model is obtained by screening based on the at least two screening dimensions; A page presentation module, configured to, in response to a trigger operation on an invocation entry of a target model among the at least one alternative model, when the model usage status of the target model is a normal status, present a target detection page corresponding to the target model; Among them, the target detection page is used to call the target model to implement corresponding target detection.

14. An electronic device, characterized in that, Comprising: A memory for storing executable instructions; A processor, configured to, when executing the executable instructions stored in the memory, implement the model invocation method for target detection according to any one of claims 1 to 12.

15. A computer-readable storage medium, characterized in that, Stored with executable instructions, configured to, when being executed by a processor, implement the model invocation method for target detection according to any one of claims 1 to 12.

16. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the model invocation method for target detection according to any one of claims 1 to 12.

Images