Desktop cleanliness evaluation method and device, storage medium and computer equipment

Abstract

The invention discloses a desktop tidiness evaluation method and device, a storage medium and computer equipment, and the method comprises the steps: obtaining a desktop image through shooting a target desktop region, recognizing the types of articles on a desktop through an image recognition model, and then carrying out the evaluation of the desktop tidiness according to the types of the articles. And when the evaluation result does not meet the preset requirement, calling a large model to generate and output desktop arrangement prompt information so as to guide the user to perform desktop arrangement. According to the method, the desktop cleanliness is evaluated through an automatic means, so that the manpower and time cost can be saved, and the evaluation task can be completed more quickly and more accurately. Through a mode of automatically outputting the desktop arrangement prompt information, a user is helped to quickly improve the desktop tidiness, the learning and working environment quality of the user is improved, the user can be promoted to develop a good desktop management habit, the user is helped to develop a living habit of keeping the desktop tidy and orderly, and the user experience is improved. Therefore, high efficiency and concentration are kept in daily work and study.

Description

Technical Field

[0001] The present application relates to the technical field of equipment monitoring, and in particular to a desktop cleanliness evaluation method, device, storage medium and computer equipment. Background Art

[0002] In daily work and life, keeping the desktop clean is important for improving work efficiency, creating a comfortable environment, and enhancing personal image. A clean and tidy learning environment can help children concentrate better and avoid being distracted by debris or stains, thereby improving their learning efficiency. Secondly, good desktop hygiene habits can also cultivate children's respect and sense of responsibility for personal and public spaces, which is not only important in daily life, but also in their future social interactions. In addition, keeping clean and hygienic can reduce the risk of disease transmission and allow children to learn and grow in a healthy and safe environment. Therefore, it is very important to cultivate and adhere to good desktop hygiene habits.

[0003] However, most of the current methods rely on manual subjective judgment, which is not only inefficient but also prone to forgetfulness. Summary of the invention

[0004] In view of this, the embodiments of the present application provide a desktop cleanliness assessment method, device, storage medium and computer equipment. By performing desktop cleanliness assessment through automated means, manpower and time costs can be greatly saved. Compared with traditional manual assessment methods, this method can complete the assessment task more quickly and accurately.

[0005] According to one aspect of the present application, a method for evaluating desktop neatness is provided, the method comprising:

[0006] Acquire a desktop image obtained by photographing a target desktop area;

[0007] Performing object recognition on the desktop image by using an image recognition model to determine the type of each object contained on the target desktop;

[0008] Evaluate the cleanliness of the target desktop according to the type of each item;

[0009] If the cleanliness evaluation result of the target desktop does not meet the preset cleanliness requirement, the large model is called to generate desktop tidying prompt information based on the cleanliness evaluation result, and the desktop tidying prompt information is output.

[0010] Optionally, the cleanliness of the target desktop is evaluated according to the type of each object, including:

[0011] Determine the number of items and the number of item types corresponding to each item according to the item type, calculate the item quantity score corresponding to the item quantity and the item type quantity score corresponding to the item type quantity, and perform weighted summation of the item quantity score and the item type quantity score to obtain a first neatness score of the target desktop;

[0012] According to the item type, the category deduction score corresponding to the item type is obtained, and based on the preset category basic score and the category deduction score corresponding to each item type, a second neatness score of the target desktop is determined.

[0013] Optionally, the first cleanliness score = 0.6 / number of items + 0.4 / number of item types, wherein when the number of items is 0, the number of items and the number of item types are both 1; the second cleanliness score = 1-the deduction score of each item type.

[0014] Optionally, performing object recognition on the desktop image by using an image recognition model to determine the object type of each object contained on the target desktop includes:

[0015] The desktop image is subjected to object instance segmentation and instance type recognition through the image recognition model to determine the instance segmentation result and item type of each item.

[0016] Optionally, evaluating the cleanliness of the target desktop according to the type of each object further includes:

[0017] According to the instance segmentation results of each item and the item type, the instance segmentation results of other item types are removed from the instance segmentation results of the item type being desktop to determine the vacant area of ​​the desktop, and the ratio of the vacant area of ​​the desktop to the desktop area is calculated as the third neatness score of the target desktop.

[0018] Optionally, after evaluating the cleanliness of the target desktop according to the type of each object, the method further includes:

[0019] Determine an overall cleanliness score according to the first cleanliness score, the second cleanliness score, and the third cleanliness score, and respectively determine whether the first cleanliness score, the second cleanliness score, the third cleanliness score, and the overall cleanliness score are less than their respective corresponding score thresholds;

[0020] When at least one of the first neatness score, the second neatness score, the third neatness score and the overall neatness score is less than the corresponding score threshold, it is determined that the neatness evaluation result of the target desktop does not meet the preset neatness requirement.

[0021] Optionally, the image recognition model includes a basic image recognition model, a discriminator, a feature recognition module, and a secondary image recognition model; performing object instance segmentation and instance type recognition on the desktop image through the image recognition model to determine the instance segmentation result and the object type of each object, including:

[0022] Performing object instance segmentation and instance type recognition on the desktop image by using the basic image recognition model to determine an instance segmentation result of each object among the objects on the target desktop and a confidence level that each object belongs to each of a plurality of preset object types;

[0023] For each object, the discriminator discriminates whether the confidence that the object belongs to each preset object category satisfies an object category determination condition of the object, and determines the object category of the object based on the confidence that the object belongs to each preset object category when it is determined that the object category determination condition is satisfied, and inputs the instance segmentation result corresponding to the object into the feature recognition module when it is determined that the object category determination condition is not satisfied;

[0024] The feature recognition module is used to extract features from the instance segmentation result of the object, and the secondary image recognition model is used to determine the object type based on the extracted object features.

[0025] Optionally, determining the object type of the object based on the extracted object features by the secondary image recognition model includes:

[0026] The feature similarities between the object feature and each preset object feature in a preset object feature library are calculated respectively through the secondary image recognition model, and the object type of the object is determined based on the feature similarities, wherein the preset object feature library includes preset object features corresponding to at least one object under each preset object type.

[0027] Optionally, calling the large model to generate desktop tidying prompt information based on the neatness evaluation result includes:

[0028] The desktop image is annotated based on the instance segmentation results of each object and the object type, and a prompt word is generated based on the annotated desktop image, the neatness evaluation result and a preset prompt word template, wherein the preset prompt word template includes desktop organization guidance information, and the desktop organization guidance information is used to guide the large model to generate a desktop organization method for improving desktop neatness based on the annotated desktop image and the neatness evaluation result;

[0029] The large model is called to generate the desktop arrangement prompt information based on the prompt word.

[0030] Optionally, after outputting the desktop organization prompt information, the method further includes:

[0031] After a preset tidying time, reacquire a new desktop image obtained by photographing the target desktop area, and perform a tidyness evaluation on the target desktop based on the new desktop image;

[0032] Compare the new cleanliness assessment result corresponding to the target desktop with the previous cleanliness assessment result to determine the cleanliness improvement of the target desktop, and if the cleanliness improvement of the target desktop meets the preset improvement requirements, issue desktop organization reward resources to the user account corresponding to the target desktop.

[0033] Optionally, after evaluating the cleanliness of the target desktop according to the type of each object, the method further includes:

[0034] If the cleanliness evaluation result of the target desktop meets the preset cleanliness requirement, a first desktop cleanliness reward resource is issued to the user account corresponding to the target desktop;

[0035] Accordingly, after evaluating the cleanliness of the target desktop based on the new desktop image, the method further includes:

[0036] If the new cleanliness evaluation result meets the preset cleanliness requirement, a second desktop cleanliness reward resource is issued to the user account corresponding to the target desktop, wherein the value of the first desktop cleanliness reward resource is greater than the value of the second desktop cleanliness reward resource, the value of the second desktop cleanliness reward resource is greater than the value of the desktop tidying reward resource, and the value of the first desktop cleanliness reward resource is greater than the sum of the values ​​of the second desktop cleanliness reward resource and the desktop tidying reward resource, and the number of times the desktop tidying reward resource is allocated within a preset desktop evaluation duration is not greater than 1;

[0037] If the new neatness evaluation result does not meet the preset neatness requirement, the large model is called to generate new desktop organization prompt information based on the new neatness evaluation result and output it.

[0038] According to another aspect of the present application, a desktop neatness assessment device is provided, the device comprising:

[0039] An acquisition module, used to acquire a desktop image obtained by photographing a target desktop area;

[0040] A recognition module, used to perform object recognition on the desktop image through an image recognition model to determine the type of each object contained on the target desktop;

[0041] An evaluation module, used for evaluating the cleanliness of the target desktop according to the type of each object;

[0042] The prompt module is used to call the large model to generate desktop tidying prompt information based on the tidyness evaluation result if the tidyness evaluation result of the target desktop does not meet the preset tidyness requirement, and output the desktop tidying prompt information.

[0043] Optionally, the evaluation module is further used to:

[0044] Determine the number of items and the number of item types corresponding to each item according to the item type, calculate the item quantity score corresponding to the item quantity and the item type quantity score corresponding to the item type quantity, and perform weighted summation of the item quantity score and the item type quantity score to obtain a first neatness score of the target desktop;

[0045] According to the item type, the category deduction score corresponding to the item type is obtained, and based on the preset category basic score and the category deduction score corresponding to each item type, a second neatness score of the target desktop is determined.

[0046] Optionally, the first cleanliness score = 0.6 / number of items + 0.4 / number of item types, wherein when the number of items is 0, the number of items and the number of item types are both 1; the second cleanliness score = 1-the deduction score of each item type.

[0047] Optionally, the identification module is further used to:

[0048] The desktop image is subjected to object instance segmentation and instance type recognition through the image recognition model to determine the instance segmentation result and item type of each item.

[0049] Optionally, the evaluation module is further used to:

[0050] According to the instance segmentation results of each item and the item type, the instance segmentation results of other item types are removed from the instance segmentation results of the item type being desktop to determine the vacant area of ​​the desktop, and the ratio of the vacant area of ​​the desktop to the desktop area is calculated as the third neatness score of the target desktop.

[0051] Optionally, the evaluation module is further used to:

[0052] Determine an overall cleanliness score according to the first cleanliness score, the second cleanliness score, and the third cleanliness score, and respectively determine whether the first cleanliness score, the second cleanliness score, the third cleanliness score, and the overall cleanliness score are less than their respective corresponding score thresholds;

[0053] When at least one of the first neatness score, the second neatness score, the third neatness score and the overall neatness score is less than the corresponding score threshold, it is determined that the neatness evaluation result of the target desktop does not meet the preset neatness requirement.

[0054] Optionally, the image recognition model includes a basic image recognition model, a discriminator, a feature recognition module, and a secondary image recognition model; the recognition module is further used to:

[0055] Performing object instance segmentation and instance type recognition on the desktop image by using the basic image recognition model to determine an instance segmentation result of each object among the objects on the target desktop and a confidence level that each object belongs to each of a plurality of preset object types;

[0056] For each object, the discriminator discriminates whether the confidence that the object belongs to each preset object category satisfies an object category determination condition of the object, and determines the object category of the object based on the confidence that the object belongs to each preset object category when it is determined that the object category determination condition is satisfied, and inputs the instance segmentation result corresponding to the object into the feature recognition module when it is determined that the object category determination condition is not satisfied;

[0057] The feature recognition module is used to extract features from the instance segmentation result of the object, and the secondary image recognition model is used to determine the object type based on the extracted object features.

[0058] Optionally, the identification module is further used to:

[0059] The feature similarities between the object feature and each preset object feature in a preset object feature library are calculated respectively through the secondary image recognition model, and the object type of the object is determined based on the feature similarities, wherein the preset object feature library includes preset object features corresponding to at least one object under each preset object type.

[0060] Optionally, calling the large model to generate desktop tidying prompt information based on the neatness evaluation result includes:

[0061] The desktop image is annotated based on the instance segmentation results of each object and the object type, and a prompt word is generated based on the annotated desktop image, the neatness evaluation result and a preset prompt word template, wherein the preset prompt word template includes desktop organization guidance information, and the desktop organization guidance information is used to guide the large model to generate a desktop organization method for improving desktop neatness based on the annotated desktop image and the neatness evaluation result;

[0062] The large model is called to generate the desktop arrangement prompt information based on the prompt word.

[0063] Optionally, the acquisition module is further used to reacquire a new desktop image obtained by photographing the target desktop area after a preset sorting time;

[0064] The recognition module is further used to perform object recognition on the new desktop image through an image recognition model to determine the object type of each object contained on the target desktop;

[0065] The evaluation module is further used to evaluate the neatness of the target desktop based on the item types of each item corresponding to the new desktop image;

[0066] The device also includes: a reward module, which is used to compare the new cleanliness evaluation result corresponding to the target desktop with the previous cleanliness evaluation result to determine the cleanliness improvement of the target desktop, and when the cleanliness improvement of the target desktop meets the preset improvement requirements, issue desktop organization reward resources to the user account corresponding to the target desktop.

[0067] Optionally, the reward module is further used to:

[0068] If the cleanliness evaluation result of the target desktop meets the preset cleanliness requirement, a first desktop cleanliness reward resource is issued to the user account corresponding to the target desktop;

[0069] If the new cleanliness evaluation result meets the preset cleanliness requirement, a second desktop cleanliness reward resource is issued to the user account corresponding to the target desktop, wherein the value of the first desktop cleanliness reward resource is greater than the value of the second desktop cleanliness reward resource, the value of the second desktop cleanliness reward resource is greater than the value of the desktop tidying reward resource, and the value of the first desktop cleanliness reward resource is greater than the sum of the values ​​of the second desktop cleanliness reward resource and the desktop tidying reward resource, and the number of times the desktop tidying reward resource is allocated within a preset desktop evaluation duration is not greater than 1;

[0070] If the new neatness evaluation result does not meet the preset neatness requirement, the large model is called to generate new desktop organization prompt information based on the new neatness evaluation result and output it.

[0071] According to another aspect of the present application, a storage medium is provided, on which a computer program is stored, and when the program is executed by a processor, the above-mentioned desktop neatness evaluation method is implemented.

[0072] According to another aspect of the present application, a computer device is provided, including a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, wherein the processor implements the above-mentioned desktop neatness evaluation method when executing the program.

[0073] By means of the above technical scheme, a desktop cleanliness assessment method, device, storage medium and computer equipment provided in the embodiment of the present application obtains a desktop image by photographing the target desktop area, uses an image recognition model to identify the types of items on the desktop, and then evaluates the desktop cleanliness according to the types of items. When the evaluation result does not meet the preset requirements, the large model is called to generate and output desktop tidying prompt information to guide the user to tidy the desktop. The embodiment of the present application evaluates the cleanliness of the desktop by automated means, which can greatly save manpower and time costs. Compared with the traditional manual evaluation method, this method can complete the evaluation task more quickly and accurately. By automatically outputting desktop tidying prompt information, it is helpful to help users quickly improve the cleanliness of the desktop, improve the quality of the user's learning and working environment, and can also encourage users to develop good desktop management habits, which helps users develop living habits of keeping the desktop clean and orderly, so as to maintain efficiency and concentration in daily work and study.

[0074] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0076] Figure 1 A schematic diagram of a process of evaluating a desktop neatness provided by an embodiment of the present application is shown;

[0077] Figure 2 A schematic diagram of a process of another desktop neatness evaluation method provided in an embodiment of the present application is shown;

[0078] Figure 3 A schematic structural diagram of a desktop neatness assessment device provided in an embodiment of the present application is shown. DETAILED DESCRIPTION

[0079] The present application will be described in detail below with reference to the accompanying drawings and in combination with embodiments. It should be noted that the embodiments and features in the embodiments of the present application can be combined with each other without conflict.

[0080] In this embodiment, a method for evaluating the neatness of a desktop is provided. Figure 1 As shown, the method includes:

[0081] Step 101: Acquire a desktop image obtained by photographing a target desktop area.

[0082] Step 102: Perform object recognition on the desktop image through an image recognition model to determine the type of each object contained on the target desktop.

[0083] Step 103: Evaluate the cleanliness of the target desktop according to the type of each object.

[0084] Step 104: If the cleanliness evaluation result of the target desktop does not meet the preset cleanliness requirement, the large model is called to generate desktop tidying prompt information based on the cleanliness evaluation result, and the desktop tidying prompt information is output.

[0085] The desktop tidiness assessment method provided in the embodiment of the present application is intended to evaluate and prompt the user to tidy up the desktop by automated means to achieve the preset tidiness standard. First, the target desktop area is photographed by a camera or other image acquisition device to obtain the image of the area, providing basic data for subsequent image recognition. In actual application scenarios, the target desktop area can be photographed by a camera in a learning machine arranged in front of the table to obtain a desktop image. Secondly, an image recognition model (such as a Yolov5 model) is used to identify objects in the acquired desktop image. The model can identify the objects in the image and determine their types (such as books, stationery, water cups, etc.). This step is the key to assessing tidiness, because the judgment of tidiness usually depends on the types and quantity of objects on the desktop. Then, according to the identified types of objects, the target desktop is assessed for tidiness, and the desktop tidiness can be assessed based on factors such as the number, type distribution, and placement of objects. For example, too much debris or disordered placement may reduce the tidiness score. Finally, if the tidiness assessment result of the target desktop does not meet the preset tidiness requirements, a large model (such as a large language model, a multimodal model, etc.) is called to generate a desktop tidying prompt message. These prompts may include information that reminds the user that the current desktop is not tidy enough, and may also include specific tidying suggestions, such as "Please move the water cup to one side of the desktop", "Put the books in categories", etc. Then, these prompts are conveyed to the user through display devices or voice output.

[0086] By applying the technical solution of this embodiment, the desktop image is obtained by photographing the target desktop area, the type of items on the desktop is identified using an image recognition model, and then the desktop cleanliness is evaluated according to the type of items. When the evaluation result does not meet the preset requirements, the large model is called to generate and output desktop tidying prompt information to guide the user to tidy the desktop. The embodiment of the present application uses automated means to evaluate the cleanliness of the desktop, which can greatly save manpower and time costs. Compared with traditional manual evaluation methods, this method can complete the evaluation task more quickly and accurately. By automatically outputting desktop tidying prompt information, it is helpful to help users quickly improve the cleanliness of their desktops, improve the quality of their learning and working environments, and encourage users to develop good desktop management habits, which helps users develop living habits of keeping their desktops clean and orderly, so as to maintain efficiency and concentration in daily work and study.

[0087] Further, as a refinement and extension of the specific implementation of the above embodiment, in order to fully illustrate the specific implementation process of this embodiment, another desktop neatness evaluation method is provided, such as Figure 2 As shown, the method includes:

[0088] Step 201: Acquire a desktop image obtained by photographing a target desktop area.

[0089] Step 202: Use the image recognition model to perform object instance segmentation and instance type recognition on the desktop image to determine instance segmentation results and item types of each item.

[0090] In the embodiment of the present application, the image recognition model will first analyze the desktop image, identify and segment each independent item in the image through a deep learning algorithm. This process is similar to "circling" each item in the image, ensuring that each item is individually identified and processed to avoid confusion or overlap between items affecting the recognition result. The instance segmentation technology can be accurate to the boundary of the item, making the subsequent type recognition more accurate. After completing the item instance segmentation, the image recognition model will further identify the type of each segmented item instance. Specifically, the features of each item instance (such as shape, color, texture, etc.) can be compared with the item type feature library pre-learned by the model to determine the item type to which each item belongs. This process is similar to "labeling each circled item" and clearly indicating what each item is (such as a book, a water cup, a pen holder, etc.). By combining item instance segmentation and instance type recognition, the image recognition model can accurately identify the item type of each item on the target desktop, providing reliable data support for subsequent neatness assessment.

[0091] In the embodiment of the present application, optionally, the image recognition model includes a basic image recognition model, a discriminator, a feature recognition module, and a secondary image recognition model; step 202 includes:

[0092] Step 202-1: performing object instance segmentation and instance type recognition on the desktop image by using the basic image recognition model to determine an instance segmentation result of each object among the objects on the target desktop and a confidence level that each object belongs to each of multiple preset object types;

[0093] Step 202-2: for each object, the discriminator discriminates whether the confidence that the object belongs to each preset object category satisfies the object category determination condition of the object, and determines the object category of the object based on the confidence that the object belongs to each preset object category when it is determined that the object category determination condition is satisfied, and inputs the instance segmentation result corresponding to the object into the feature recognition module when it is determined that the object category determination condition is not satisfied;

[0094] Step 202-3: extract features from the instance segmentation result of the object through the feature recognition module, and determine the object type of the object based on the extracted object features through the secondary image recognition model.

[0095] In this embodiment, in order to improve the accuracy of the image recognition model in identifying the type of item, a discriminator, a feature recognition module, and a secondary image recognition model can be added on the basis of the basic image recognition model to obtain a complete image recognition model. Of course, the basic image recognition model can also be directly used as an image recognition model. Take the image recognition model including the basic image recognition model, the discriminator, the feature recognition module, and the secondary image recognition model as an example: the basic image recognition model first performs preliminary processing on the desktop image, performs the basic tasks of item instance segmentation and instance type recognition, and the basic image recognition model can select the Yolov5 model or other types of image recognition models. For each identified item instance, the basic image recognition model will provide a confidence list, which contains the confidence that the item belongs to each of the multiple preset item types. The discriminator receives the confidence list output by the basic image recognition model and further analyzes and judges the confidence of each item. It can evaluate whether the item type of the item can be directly determined according to the preset item type determination conditions (such as confidence threshold). For example, if the confidence of an item satisfies the determination condition (for example, only one item type has a confidence that exceeds the confidence threshold, or there are multiple item types whose confidence exceeds the confidence threshold but the difference between the largest confidence and the second largest confidence is greater than or equal to a preset value), the discriminator determines the item type of the item based on the confidence, for example, determining the item type corresponding to the largest confidence as the item type of the item. If the determination condition is not met (i.e., the confidence of all item types is lower than the threshold, or the confidence distribution is too dispersed, or there are multiple item types whose confidence exceeds the confidence threshold but the difference between the largest confidence and the second largest confidence is less than a preset value), the discriminator passes the instance segmentation result corresponding to the item to the feature recognition module for further analysis. The feature recognition module receives the item instance segmentation results from the discriminator that cannot directly determine the item type, extracts features for these items, and uses the extracted features for further analysis and recognition. Based on the object features extracted by the feature recognition module, the secondary image recognition model performs a deeper identification of the item types. Ultimately, the secondary image recognition model determines the item types of those items that are not directly identified by the basic model. By combining these components, the image recognition model in the embodiment of the present application can efficiently perform the tasks of item instance segmentation and instance type identification of desktop images. It first attempts to use the basic model and discriminator for fast and accurate identification; for items that are difficult to identify, it uses the feature recognition module and the secondary image recognition model for a deeper analysis and identification. This staged processing method not only improves the accuracy of recognition, but also optimizes the overall computational efficiency and performance.

[0096] In an embodiment of the present application, optionally, in step 202-3, the item type of the item is determined based on the extracted item features through the secondary image recognition model, including: calculating the feature similarities between the item features and each preset item feature in a preset item feature library through the secondary image recognition model, and determining the item type of the item based on the feature similarities, wherein the preset item feature library includes preset item features corresponding to at least one item under each preset item type.

[0097] In this embodiment, the secondary image recognition model first receives the item features extracted by the feature recognition module, and then the model compares these features with each preset item feature in the preset item feature library. The preset item feature library is a database containing multiple item category features, and at least one or more representative features corresponding to each item are stored under each item category. For example, the item category features corresponding to the pen in the preset item feature library can be obtained by extracting features from multiple pen images of different shapes taken at different angles. The secondary image recognition model uses a similarity measurement method (such as cosine similarity, Euclidean distance, etc.) to calculate the similarity between the extracted item features and each feature in the preset item feature library. After calculating all similarities, the secondary image recognition model will find the preset item feature that is most similar to the extracted item feature, and determine the item category of the item based on the most similar preset item feature. In some cases, the extracted item feature may have a high similarity with multiple preset item features in the preset item feature library, or a low similarity with all preset item features. In order to handle these boundary conditions, the secondary image recognition model can also adopt some additional strategies, such as setting a similarity threshold, using a classifier for secondary judgment, etc. For example, the item feature has a high similarity with multiple preset item features in the preset item feature library (for example, multiple similarities are greater than 90%), and for another example, the similarity between the item feature and all the preset item features in the preset item feature library is low (for example, the similarity is less than 50%). The item type of the item can be further determined in combination with the confidence of multiple preset item types output by the basic image recognition model. Specifically, the highest similarity between the item feature and each preset item type can be determined based on the similarity between the item feature and each preset item feature, and the highest similarity corresponding to each preset item type is used as the similarity between the item and each preset item type, and the item type of the item is comprehensively determined based on the similarity between the item and each preset item type and the confidence that the item belongs to each preset item type. In this way, the embodiment of the present application can use the information in the preset item feature library to perform a deeper item type recognition on items that cannot be directly recognized by the basic model. This not only improves the accuracy of recognition, but also expands the scope of application of the image recognition model, enabling it to handle more diverse and complex item recognition tasks.

[0098] Step 203: Determine the number of items and the number of item types corresponding to each item according to the item type, calculate the item quantity score corresponding to the item quantity and the item type quantity score corresponding to the item type number, perform weighted summation on the item quantity score and the item type quantity score to obtain a first neatness score for the target desktop; obtain the type deduction score corresponding to the item type according to the item type, and determine the second neatness score for the target desktop based on a preset type basic score and the type deduction score corresponding to each item type; according to the instance segmentation result of each item and the item type, remove the instance segmentation results of other item types from the instance segmentation result of the item type being the desktop to determine the vacant area of ​​the desktop, and calculate the proportion of the vacant area of ​​the desktop relative to the desktop area as the third neatness score of the target desktop.

[0099] In this embodiment, the tidiness score of the target desktop may specifically include the first, second, and third tidiness scores. Calculate the first tidiness score: According to the item types of each item determined in the previous step, count the number of each item, and calculate the total number of different item types. According to a preset scoring standard or function, the number of items is converted into an item quantity score. The more the number, the lower the score may be. Similarly, convert the number of item types into an item type quantity score. The more types, the less tidy the desktop is, and therefore the lower the score. According to the importance of the item quantity score and the item type quantity score, different weights are assigned to them, and the two weighted scores are added to obtain the first tidiness score of the target desktop. Optionally, the first tidiness score = 0.6 / number of items + 0.4 / number of item types, wherein when the number of items is 0, the number of items and the number of item types are both 1. Calculate the second neatness score: assign a deduction score to each category according to the degree of influence of each item type on the neatness, preset a category basic score, indicating the score when there are no items or only a small number of items that do not affect the neatness, and deduct the corresponding score from the basic score according to the deduction score of each item type on the desktop to obtain the second neatness score. Optionally, the second neatness score = 1-category deduction score of each item type. Calculate the third neatness score: use instance segmentation technology to segment the desktop from the background, and remove all other instance segmentation results of all item types from the instance segmentation results of the desktop according to the instance segmentation results of each item and the item type. The remaining area is the vacant area of ​​the desktop, or the desktop area is identified by the maximum connected area to obtain the vacant area of ​​the desktop. Calculate the area or number of pixels of the vacant area of ​​the desktop, and compare it with the area or number of pixels of the entire desktop area to obtain the proportion of the vacant area relative to the desktop area, and determine the third neatness score. The higher the proportion of vacant area, the cleaner the desktop is, so the higher the third neatness score. Finally, the first tidiness score, the second tidiness score and the third tidiness score can be combined in some way (such as weighted average) to obtain a final tidiness score of the target desktop. This score can be used to evaluate the tidiness of the desktop, and can also be used as a reference for users to tidy up the desktop or perform other related operations.

[0100] Step 204: determine an overall neatness score based on the first neatness score, the second neatness score, and the third neatness score, and respectively determine whether the first neatness score, the second neatness score, the third neatness score, and the overall neatness score are less than their respective corresponding score thresholds; when at least one of the first neatness score, the second neatness score, the third neatness score, and the overall neatness score is less than their respective corresponding score thresholds, determine that the cleanliness evaluation result of the target desktop does not meet the preset cleanliness requirement.

[0101] In this embodiment, according to the preset weight or evaluation method, the first neatness score, the second neatness score and the third neatness score are weighted summed or comprehensively evaluated, and the final result is the overall neatness score. For the first neatness score, the second neatness score, the third neatness score and the overall neatness score, corresponding scoring thresholds are set respectively, and these thresholds are determined according to the preset neatness requirements or standards, and are used to judge whether the score is up to standard. Each score is compared with its corresponding threshold. If the score is less than the threshold, it is considered that the scoring item does not meet the preset neatness requirements. If at least one of the first neatness score, the second neatness score, the third neatness score and the overall neatness score is less than the corresponding scoring threshold, it can be determined that the cleanliness evaluation result of the target desktop does not meet the preset cleanliness requirements.

[0102] Step 205: If the cleanliness evaluation result of the target desktop does not meet the preset cleanliness requirement, the desktop image is labeled based on the instance segmentation result of each object and the type of object, and prompt words are generated based on the labeled desktop image, the cleanliness evaluation result and a preset prompt word template, wherein the preset prompt word template includes desktop organization guide information, and the desktop organization guide information is used to guide the large model to generate a desktop organization method for improving desktop cleanliness based on the labeled desktop image and the cleanliness evaluation result; the large model is called to generate the desktop organization prompt information based on the prompt word, and the desktop organization prompt information is output.

[0103] In this embodiment, after determining that the tidiness evaluation result of the target desktop does not meet the preset tidiness requirements, a series of measures are taken to generate and output desktop tidying prompt information. Specifically, the desktop image is annotated using the instance segmentation results of each item obtained in the previous step and the determined item type. The annotation may include drawing a bounding box around each item, marking the item name or type, and possible other related information (such as the number of items, location, etc.). Design one or more preset prompt word templates, which contain desktop tidying guidance information. The desktop tidying guidance information may be suggestions on how to tidy up the desktop, which items need to be moved or discarded, and possible tidying order. Fill the annotated desktop image and the tidiness evaluation result into the preset prompt word template to generate prompt words containing specific desktop tidying guidance information. Select a large model (such as GPT, BERT, etc.) that has been trained and can understand and generate natural language text to ensure that the large model has sufficient context understanding ability and can generate reasonable desktop tidying prompt information based on the prompt word. The generated prompt word is passed as input to the large model, and the large model generates specific desktop tidying prompt information based on the content of the prompt word, combined with its internal knowledge and training data. The desktop organization prompt information generated by the large model is displayed to the user in the form of text, image or other forms, so as to guide the user to effectively organize the desktop. Through the above steps, when the cleanliness evaluation result of the target desktop does not meet the preset requirements, the system can automatically generate and output targeted desktop organization prompt information to help users improve the cleanliness of their desktop.

[0104] In an embodiment of the present application, optionally, step 205 also includes: step 206, after a preset tidying time, re-acquiring a new desktop image obtained by photographing the target desktop area, and performing a cleanliness assessment on the target desktop based on the new desktop image; comparing the new cleanliness assessment result corresponding to the target desktop with the previous cleanliness assessment result to determine the cleanliness improvement of the target desktop, and when the cleanliness improvement of the target desktop meets the preset improvement requirements, issuing desktop tidying reward resources to the user account corresponding to the target desktop.

[0105] In this embodiment, after the preset tidying time (such as 10 minutes), the target desktop area is re-photographed using a camera or other image acquisition device to obtain a new desktop image, ensuring that the shooting conditions (such as light, angle, etc.) are consistent with the previous shooting, so as to make an accurate comparison. The new desktop image is evaluated using the same neatness evaluation process as before (including image preprocessing, feature extraction, object recognition, score calculation and other steps) to obtain a new neatness evaluation result, including a first neatness score, a second neatness score, a third neatness score and an overall neatness score. Compare the new neatness evaluation result with the last neatness evaluation result (i.e., obtained in step 204), calculate the change in each score (such as the improvement of the overall neatness score) to evaluate the improvement of neatness. According to the preset improvement requirements (such as the overall neatness score is improved by a certain range, a specific scoring item reaches a certain standard, etc.), it is judged whether the improvement of neatness meets the requirements. If the requirements are met, the desktop tidying reward resources can be issued to the corresponding user account. The reward resources can be virtual currency, points, coupons, physical prizes, etc., and the specific form is determined according to the system settings and user preferences. When issuing rewards, users can be notified through system messages, emails, text messages, etc., and guided to view or use the rewards. Through the above steps, the system can motivate users to actively participate in desktop tidying activities, improve the neatness of the desktop, and provide users with positive feedback and rewards, which helps to cultivate users' good tidying habits and improve work and life efficiency.

[0106] In an embodiment of the present application, optionally, after step 204, the method further includes: if the cleanliness evaluation result of the target desktop meets the preset cleanliness requirement, issuing a first desktop cleanliness reward resource to the user account corresponding to the target desktop; accordingly, after step 206, the method further includes: if the new cleanliness evaluation result meets the preset cleanliness requirement, issuing a second desktop cleanliness reward resource to the user account corresponding to the target desktop, wherein the value of the first desktop cleanliness reward resource is greater than the value of the second desktop cleanliness reward resource, the value of the second desktop cleanliness reward resource is greater than the value of the desktop tidying reward resource, and the value of the first desktop cleanliness reward resource is greater than the sum of the values ​​of the second desktop cleanliness reward resource and the desktop tidying reward resource, and the number of allocations of the desktop tidying reward resource within a preset desktop evaluation duration is not greater than 1; if the new cleanliness evaluation result does not meet the preset cleanliness requirement, calling the large model to generate new desktop tidying prompt information based on the new cleanliness evaluation result and output it.

[0107] In this embodiment, in order to encourage users to keep the desktop clean and give a more detailed reward mechanism for the efforts to improve the cleanliness, reward resources with different values ​​can be issued in different situations. Specifically, if in step 204, the cleanliness evaluation result of the target desktop meets the preset cleanliness requirements for the first time, the first desktop cleanliness reward resource is issued to the user account corresponding to the target desktop. The value of the first desktop cleanliness reward resource is set to the highest, in order to commend the good tidying habits or one-time excellent tidying results that the user already has. In step 206, if after tidying, the new cleanliness evaluation result meets the preset cleanliness requirements, the second desktop cleanliness reward resource is issued to the user. The value of the second desktop cleanliness reward resource is lower than the first desktop cleanliness reward resource, but higher than the desktop tidying reward resource that may be obtained due to the subsequent improvement of cleanliness, so as to encourage users to actively respond to the tidying suggestions and successfully improve the cleanliness of the desktop. If the user has significantly improved the cleanliness after the preset tidying time (but not necessarily the preset requirements for the first time), the desktop tidying reward resource is issued. The value of the reward resource is set to the lowest, as a recognition and encouragement of the user's efforts to improve cleanliness. In order to avoid excessive rewards within a preset desktop evaluation period (such as one day, one week, etc.), the number of allocations of desktop organization reward resources is limited to no more than 1 time. Through such a reward mechanism design, the system can more effectively motivate users to maintain and improve the cleanliness of their desktops. Users can not only receive high rewards for their good organization habits or one-time excellent organization results, but also receive certain rewards for actively responding to organization suggestions and successfully improving cleanliness. At the same time, users who work hard to improve cleanliness can receive certain recognition and encouragement even if they fail to meet the preset requirements at one time. Such a mechanism helps to cultivate users' good organization habits, improve work and life efficiency, and enhance user satisfaction and loyalty to the system.

[0108] By applying the technical solution of this embodiment, through the automated cleanliness evaluation process, including steps such as image preprocessing, feature extraction, object recognition, and score calculation, the cleanliness of the desktop can be accurately evaluated, and targeted tidying suggestions can be provided to users. After the user tidies up according to the tidying suggestions, the system can re-evaluate the cleanliness to ensure that the cleanliness of the desktop is effectively improved. In addition, through the preset prompt word template and the desktop tidying prompt information generated by the large model, the system can provide users with personalized tidying suggestions to enhance the user's sense of participation and satisfaction. The design of the reward mechanism encourages users to actively participate in desktop tidying activities and continue to keep the desktop clean. Through reward resources of different values, the system can motivate users to continuously improve cleanliness, thereby cultivating good tidying habits.

[0109] Further, as Figure 1 The specific implementation of the method, the embodiment of the present application provides a desktop cleanliness evaluation device, such as Figure 3 As shown, the device comprises:

[0110] An acquisition module, used to acquire a desktop image obtained by photographing a target desktop area;

[0111] A recognition module, used to perform object recognition on the desktop image through an image recognition model to determine the type of each object contained on the target desktop;

[0112] An evaluation module, used for evaluating the cleanliness of the target desktop according to the type of each object;

[0113] The prompt module is used to call the large model to generate desktop tidying prompt information based on the tidyness evaluation result if the tidyness evaluation result of the target desktop does not meet the preset tidyness requirement, and output the desktop tidying prompt information.

[0114] Optionally, the evaluation module is further used to:

[0115] Determine the number of items and the number of item types corresponding to each item according to the item type, calculate the item quantity score corresponding to the item quantity and the item type quantity score corresponding to the item type quantity, and perform weighted summation of the item quantity score and the item type quantity score to obtain a first neatness score of the target desktop;

[0116] According to the item type, the category deduction score corresponding to the item type is obtained, and based on the preset category basic score and the category deduction score corresponding to each item type, a second neatness score of the target desktop is determined.

[0117] Optionally, the first cleanliness score = 0.6 / number of items + 0.4 / number of item types, wherein when the number of items is 0, the number of items and the number of item types are both 1; the second cleanliness score = 1-the deduction score of each item type.

[0118] Optionally, the identification module is further used to:

[0119] The desktop image is subjected to object instance segmentation and instance type recognition through the image recognition model to determine the instance segmentation result and item type of each item.

[0120] Optionally, the evaluation module is further used to:

[0121] According to the instance segmentation results of each item and the item type, the instance segmentation results of other item types are removed from the instance segmentation results of the item type being desktop to determine the vacant area of ​​the desktop, and the ratio of the vacant area of ​​the desktop to the desktop area is calculated as the third neatness score of the target desktop.

[0122] Optionally, the evaluation module is further used to:

[0123] Determine an overall cleanliness score according to the first cleanliness score, the second cleanliness score, and the third cleanliness score, and respectively determine whether the first cleanliness score, the second cleanliness score, the third cleanliness score, and the overall cleanliness score are less than their respective corresponding score thresholds;

[0124] When at least one of the first neatness score, the second neatness score, the third neatness score and the overall neatness score is less than the corresponding score threshold, it is determined that the neatness evaluation result of the target desktop does not meet the preset neatness requirement.

[0125] Optionally, the image recognition model includes a basic image recognition model, a discriminator, a feature recognition module, and a secondary image recognition model; the recognition module is further used to:

[0126] Performing object instance segmentation and instance type recognition on the desktop image by using the basic image recognition model to determine an instance segmentation result of each object among the objects on the target desktop and a confidence level that each object belongs to each of a plurality of preset object types;

[0127] For each object, the discriminator discriminates whether the confidence that the object belongs to each preset object category satisfies an object category determination condition of the object, and determines the object category of the object based on the confidence that the object belongs to each preset object category when it is determined that the object category determination condition is satisfied, and inputs the instance segmentation result corresponding to the object into the feature recognition module when it is determined that the object category determination condition is not satisfied;

[0128] The feature recognition module is used to extract features from the instance segmentation result of the object, and the secondary image recognition model is used to determine the object type based on the extracted object features.

[0129] Optionally, the identification module is further used to:

[0130] The feature similarities between the object feature and each preset object feature in a preset object feature library are calculated respectively through the secondary image recognition model, and the object type of the object is determined based on the feature similarities, wherein the preset object feature library includes preset object features corresponding to at least one object under each preset object type.

[0131] Optionally, calling the large model to generate desktop tidying prompt information based on the neatness evaluation result includes:

[0132] The desktop image is annotated based on the instance segmentation results of each object and the object type, and a prompt word is generated based on the annotated desktop image, the neatness evaluation result and a preset prompt word template, wherein the preset prompt word template includes desktop organization guidance information, and the desktop organization guidance information is used to guide the large model to generate a desktop organization method for improving desktop neatness based on the annotated desktop image and the neatness evaluation result;

[0133] The large model is called to generate the desktop arrangement prompt information based on the prompt word.

[0134] Optionally, the acquisition module is further used to reacquire a new desktop image obtained by photographing the target desktop area after a preset sorting time;

[0135] The recognition module is further used to perform object recognition on the new desktop image through an image recognition model to determine the object type of each object contained on the target desktop;

[0136] The evaluation module is further used to evaluate the neatness of the target desktop based on the item types of each item corresponding to the new desktop image;

[0137] The device also includes: a reward module, which is used to compare the new cleanliness evaluation result corresponding to the target desktop with the previous cleanliness evaluation result to determine the cleanliness improvement of the target desktop, and when the cleanliness improvement of the target desktop meets the preset improvement requirements, issue desktop organization reward resources to the user account corresponding to the target desktop.

[0138] Optionally, the reward module is further used to:

[0139] If the cleanliness evaluation result of the target desktop meets the preset cleanliness requirement, a first desktop cleanliness reward resource is issued to the user account corresponding to the target desktop;

[0140] If the new cleanliness evaluation result meets the preset cleanliness requirement, a second desktop cleanliness reward resource is issued to the user account corresponding to the target desktop, wherein the value of the first desktop cleanliness reward resource is greater than the value of the second desktop cleanliness reward resource, the value of the second desktop cleanliness reward resource is greater than the value of the desktop tidying reward resource, and the value of the first desktop cleanliness reward resource is greater than the sum of the values ​​of the second desktop cleanliness reward resource and the desktop tidying reward resource, and the number of times the desktop tidying reward resource is allocated within a preset desktop evaluation duration is not greater than 1;

[0141] If the new neatness evaluation result does not meet the preset neatness requirement, the large model is called to generate new desktop organization prompt information based on the new neatness evaluation result and output it.

[0142] It should be noted that for other corresponding descriptions of the functional units involved in the desktop neatness evaluation device provided in the embodiment of the present application, reference can be made to Figure 1 to Figure 2 The corresponding description in the method will not be repeated here.

[0143] The embodiment of the present application also provides a computer device, which can be a personal computer, a server, a network device, etc. The computer device includes a bus, a processor, a memory and a communication interface, and can also include an input and output interface and a display device. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store location information. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, the steps in each method embodiment are implemented.

[0144] Those skilled in the art will appreciate that the structure of the above-mentioned computer device is only a partial structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components, or combine certain components, or have a different arrangement of components.

[0145] In one embodiment, a computer-readable storage medium is provided. The computer-readable storage medium may be non-volatile or volatile, and stores a computer program thereon. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0146] In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0147] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0148] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0149] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0150] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A method for evaluating desktop neatness, characterized in that: The method comprises: Acquire a desktop image obtained by photographing a target desktop area; Performing object recognition on the desktop image by using an image recognition model to determine the type of each object contained on the target desktop; Evaluate the cleanliness of the target desktop according to the type of each item; If the cleanliness evaluation result of the target desktop does not meet the preset cleanliness requirement, the large model is called to generate desktop tidying prompt information based on the cleanliness evaluation result, and the desktop tidying prompt information is output.

2. The method according to claim 1, characterized in that The cleanliness of the target desktop is evaluated according to the type of each object, including: Determine the number of items and the number of item types corresponding to each item according to the item type, calculate the item quantity score corresponding to the item quantity and the item type quantity score corresponding to the item type quantity, and perform weighted summation of the item quantity score and the item type quantity score to obtain a first neatness score of the target desktop; According to the item type, the category deduction score corresponding to the item type is obtained, and based on the preset category basic score and the category deduction score corresponding to each item type, a second neatness score of the target desktop is determined.

3. The method according to claim 2, characterized in that The first cleanliness score=0.6 / number of items+0.4 / number of item types, wherein when the number of items is 0, the number of items and the number of item types are both 1; the second cleanliness score=1-the deduction score of each item type.

4. The method according to claim 2, characterized in that: The desktop image is subjected to object recognition by an image recognition model to determine the object type of each object contained on the target desktop, including: The desktop image is subjected to object instance segmentation and instance type recognition through the image recognition model to determine the instance segmentation result and item type of each item.

5. The method according to claim 4, characterized in that Evaluating the cleanliness of the target desktop according to the type of each item, further comprising: According to the instance segmentation results of each item and the item type, the instance segmentation results of other item types are removed from the instance segmentation results of the item type being desktop to determine the vacant area of ​​the desktop, and the ratio of the vacant area of ​​the desktop to the desktop area is calculated as the third neatness score of the target desktop.

6. The method according to claim 5, characterized in that After evaluating the cleanliness of the target desktop according to the types of the objects, the method further includes: Determine an overall cleanliness score according to the first cleanliness score, the second cleanliness score, and the third cleanliness score, and respectively determine whether the first cleanliness score, the second cleanliness score, the third cleanliness score, and the overall cleanliness score are less than their respective corresponding score thresholds; When at least one of the first neatness score, the second neatness score, the third neatness score and the overall neatness score is less than the corresponding score threshold, it is determined that the neatness evaluation result of the target desktop does not meet the preset neatness requirement.

7. The method according to claim 4, characterized in that The image recognition model includes a basic image recognition model, a discriminator, a feature recognition module, and a secondary image recognition model; the desktop image is subjected to item instance segmentation and instance type recognition by the image recognition model to determine the instance segmentation result and item type of each item, including: Performing object instance segmentation and instance type recognition on the desktop image by using the basic image recognition model to determine an instance segmentation result of each object among the objects on the target desktop and a confidence level that each object belongs to each of a plurality of preset object types; For each object, the discriminator discriminates whether the confidence that the object belongs to each preset object category satisfies an object category determination condition of the object, and determines the object category of the object based on the confidence that the object belongs to each preset object category when it is determined that the object category determination condition is satisfied, and inputs the instance segmentation result corresponding to the object into the feature recognition module when it is determined that the object category determination condition is not satisfied; The feature recognition module is used to extract features from the instance segmentation result of the object, and the secondary image recognition model is used to determine the object type based on the extracted object features.

8. The method according to claim 7, characterized in that Determining the type of the object based on the extracted object features by the secondary image recognition model includes: The feature similarities between the object feature and each preset object feature in a preset object feature library are calculated respectively through the secondary image recognition model, and the object type of the object is determined based on the feature similarities, wherein the preset object feature library includes preset object features corresponding to at least one object under each preset object type.

9. The method according to any one of claims 4 to 8, characterized in that The large model is called to generate desktop tidying prompt information based on the neatness evaluation result, including: The desktop image is annotated based on the instance segmentation results of each object and the object type, and a prompt word is generated based on the annotated desktop image, the neatness evaluation result and a preset prompt word template, wherein the preset prompt word template includes desktop organization guidance information, and the desktop organization guidance information is used to guide the large model to generate a desktop organization method for improving desktop neatness based on the annotated desktop image and the neatness evaluation result; The large model is called to generate the desktop arrangement prompt information based on the prompt word.

10. The method according to any one of claims 1 to 8, characterized in that After outputting the desktop organization prompt information, the method further includes: After a preset tidying time, reacquire a new desktop image obtained by photographing the target desktop area, and perform a tidiness evaluation on the target desktop based on the new desktop image; Compare the new cleanliness assessment result corresponding to the target desktop with the previous cleanliness assessment result to determine the cleanliness improvement of the target desktop, and if the cleanliness improvement of the target desktop meets the preset improvement requirements, issue desktop organization reward resources to the user account corresponding to the target desktop.

11. The method according to claim 10, characterized in that After evaluating the cleanliness of the target desktop according to the types of the objects, the method further includes: If the cleanliness evaluation result of the target desktop meets the preset cleanliness requirement, a first desktop cleanliness reward resource is issued to the user account corresponding to the target desktop; Accordingly, after evaluating the cleanliness of the target desktop based on the new desktop image, the method further includes: If the new cleanliness evaluation result meets the preset cleanliness requirement, a second desktop cleanliness reward resource is issued to the user account corresponding to the target desktop, wherein the value of the first desktop cleanliness reward resource is greater than the value of the second desktop cleanliness reward resource, the value of the second desktop cleanliness reward resource is greater than the value of the desktop tidying reward resource, and the value of the first desktop cleanliness reward resource is greater than the sum of the values ​​of the second desktop cleanliness reward resource and the desktop tidying reward resource, and the number of times the desktop tidying reward resource is allocated within a preset desktop evaluation duration is not greater than 1; If the new neatness evaluation result does not meet the preset neatness requirement, the large model is called to generate new desktop organization prompt information based on the new neatness evaluation result and output it.

12. A desktop neatness evaluation device, characterized in that: The device comprises: An acquisition module, used to acquire a desktop image obtained by photographing a target desktop area; A recognition module, used to perform object recognition on the desktop image through an image recognition model to determine the type of each object contained on the target desktop; An evaluation module, used for evaluating the cleanliness of the target desktop according to the type of each object; The prompt module is used to call the large model to generate desktop tidying prompt information based on the tidyness evaluation result if the tidyness evaluation result of the target desktop does not meet the preset tidyness requirement, and output the desktop tidying prompt information.

13. A storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 11 is implemented.

14. A computer device comprising a storage medium, a processor, and a computer program stored in the storage medium and executable on the processor, characterized in that: When the processor executes the computer program, the method according to any one of claims 1 to 11 is implemented.

Images