Display method of key information interface, inference model establishment method and electronic device

Through artificial intelligence technology, key information inference model is established, key information categories and projects are inferred using window pixel data, which solves the problems of display information in the existing technology and the difficulty of equipment updating, and achieves the improvement of flexibility and user experience.

CN119088649APending Publication Date: 2024-12-06ACER INC
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202310657767.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2023-06-05
Publication Date
2024-12-06

AI Technical Summary

Technical Problem

The prior art is difficult to adapt to the needs of different applications and games without affecting user operations, dynamically display key information of electronic devices, and updates of new devices and applications make it difficult to set display information.

Method used

Using artificial intelligence technology to establish a key information inference model, infer key information categories and projects that need to be displayed through window pixel data, and use application layer, data analysis unit, artificial intelligence operation unit and hardware layer monitoring unit to achieve quick display and optimization.

Benefits of technology

It realizes adaptive display of key information under different applications and games, reduces the difficulty of adapting to new devices and application updates, and improves user experience and device flexibility.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119088649A_ABST
    Figure CN119088649A_ABST
Patent Text Reader

Abstract

The invention discloses a key information interface display method, a key information inference model establishment method and an electronic device applying the key information interface display method and the key information inference model establishment method. The electronic device comprises an application layer operation unit, a data analysis unit, an artificial intelligence operation unit and a hardware layer monitoring unit. If the shortcut key is triggered, the application layer operation unit detects a window of interest. The data analysis unit is used for obtaining at least one window picture of the attention window and obtaining a plurality of pixel data according to the window picture. The artificial intelligence operation unit utilizes a key information inference model to infer at least one key information category and at least one key information item of a key information interface according to the pixel data. The hardware layer monitoring unit is used for capturing an information content corresponding to each key information item.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present disclosure relates to a method for displaying an interface, a method for establishing an inference model, and an electronic device using the same, and in particular to a method for displaying a key information interface, a method for establishing a key information inference model, and an electronic device using the same. Background Art

[0002] With the rapid development of technology, many electronic devices can run different applications or games. When people run these applications or games, they may need to obtain some key information of the electronic device, such as hardware load, wireless network status, screen brightness, etc. In order not to affect the user's operation, the industry is committed to developing a quick display method for key information.

[0003] In addition, when users execute different applications or different games, the key information they need may not be the same. Therefore, how to adaptively display the key information required for different applications and games is also a major challenge currently faced by the industry. Summary of the invention

[0004] The present disclosure is related to a method for displaying a key information interface, a method for training and optimizing a key information inference model, and an electronic device using the same. The key information inference model is established using artificial intelligence technology, so that the key information category and key information items that need to be displayed on the key information interface can be directly inferred from the pixel data of the window screen. Even if new electronic devices are continuously launched, or new games, new streaming media applications, or new tool programs are continuously launched, the key information interface can be directly inferred using pixel data through the artificial intelligence technology of this embodiment.

[0005] According to one aspect of the present disclosure, a method for displaying a key information interface is provided. The method for displaying a key information interface includes the following steps. Determine whether a shortcut key is triggered. If the shortcut key is triggered, detect a focus window. Obtain at least one window screen based on the focus window. Obtain a plurality of pixel data based on the window screen. Utilize a key information inference model to infer at least one key information category and at least one key information item of each key information category using these pixel data. Capture an information content corresponding to each key information item. Display the key information category, key information item, and information content on the key information interface.

[0006] According to another aspect of the present disclosure, a method for establishing a key information inference model is provided. The training and optimization method of the key information inference model includes the following steps. Based on a plurality of applications, a plurality of pixel data of at least one window screen of each application is obtained. Based on these applications, at least one key information category and at least one key information item of each key information category that need to be displayed on a key information interface when operating each application are obtained. The key information inference model is trained with the corresponding pixel data, the key information categories and the key information items.

[0007] According to another aspect of the present disclosure, an electronic device is provided. The electronic device includes an application layer operation unit, a data analysis unit, an artificial intelligence operation unit and a hardware layer monitoring unit. The application layer operation unit is used to detect whether a shortcut key is triggered. If the shortcut key is triggered, the application layer operation unit detects a focus window. The data analysis unit is used to obtain at least one window screen of the focus window, and obtain a plurality of pixel data based on the window screen. The artificial intelligence operation unit uses a key information inference model to infer at least one key information category of a key information interface and at least one key information item of each key information category with these pixel data. The hardware layer monitoring unit is used to capture an information content corresponding to each key information item. The application layer operation unit is further used to display key information categories, key information items and information content on the key information interface.

[0008] In order to better understand the above and other aspects of the present disclosure, embodiments are given below and described in detail with reference to the accompanying drawings: BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Figure 1 A key information interface showing a game being played on an electronic device according to an embodiment is shown.

[0010] Figure 2 A key information interface of an electronic device performing word processing according to another embodiment is shown.

[0011] Figure 3 A key information interface showing a game being played on an electronic device according to another embodiment is shown.

[0012] Figure 4 A key information interface showing that an electronic device is watching a video according to another embodiment is shown.

[0013] Figure 5 A schematic diagram of a key information inference model according to an embodiment is shown.

[0014] Figure 6 A state machine diagram of a key information inference model according to an embodiment is shown.

[0015] Figure 7 A block diagram of an electronic device according to an embodiment is shown.

[0016] Figure 8 A flow chart of a method for training a key information inference model according to an embodiment is shown.

[0017] Fig. 9 A flow chart of a method for displaying a key information interface according to an embodiment is shown.

[0018] Fig.10 A schematic diagram of an artificial intelligence operation unit according to an embodiment is shown.

[0019] Fig.11 A flow chart of a method for optimizing a key information inference model according to an embodiment is shown.

[0020] Description of Reference Numerals

[0021] 100,300,400,700:Electronic devices

[0022] 710: Application layer operation unit

[0023] 720: Data Analysis Unit

[0024] 730: Artificial Intelligence Operation Unit

[0025] 731: Status Monitor

[0026] 732:Optimizer

[0027] 740: Hardware layer monitoring unit

[0028] API: Application

[0029] CG, CGi, CGj: Key information categories

[0030] CTj: Information content

[0031] FM11,FM12,FM3,FM4,FMi,FMj:Window screen

[0032] IT,ITi,ITj: Key Information Items

[0033] KY: shortcut keys

[0034] M1: Training Mode

[0035] M2: Inference Mode

[0036] M3: Optimization mode

[0037] MD: Key Information Inference Model

[0038] PX, PXi, PXj: pixel data

[0039] S810, S820, S830, S910, S920, S930, S940, S950, S960, S970, S111, S112: Steps

[0040] UI: Key information interface

[0041] WDj: Focus on the window DETAILED DESCRIPTION

[0042] Please refer to Figure 1 , which depicts a key information interface UI of an electronic device 100 playing a game according to an embodiment. The electronic device 100 is, for example, a laptop computer. When a user uses the electronic device 100 to play a game, the key information interface UI can be opened to display key information. The key information interface UI is displayed on the top layer, covering all components displayed on the screen, so it is also called Overlay user interface (Overlay UI). Figure 1 As shown, the key information category CG displayed in the key information interface UI is, for example, "Communication" and "Device loading". The key information item IT displayed in the key information category CG of "Communication" is, for example, "WiFi ON". The key information item IT displayed in the key information category CG of "Device loading" is, for example, "CPU 20%", "GPU 30%", "FPS--".

[0043] Please refer to Figure 2 , which illustrates a key information interface UI of the electronic device 100 in the process of document processing according to another embodiment. When the user is using the electronic device 100 to process documents, the key information interface UI can be opened to display key information. Figure 2 As shown, the key information category CG displayed in the key information interface UI is, for example, "I / O" and "Storage". The key information items IT displayed in the key information category CG of "I / O" are, for example, "USB1 ON" and "hdmiON". The key information items IT displayed in the key information category CG of "Storage" are, for example, "RAM 20%" and "VRAM 30%".

[0044] That is, when the user executes different applications, the key information interface UI can adaptively display the required key information categories CG and key information items IT.

[0045] Please refer to Figure 3, which depicts a key information interface UI of an electronic device 300 playing a game according to another embodiment. The electronic device 300 is, for example, a handheld game console. When a user is playing a game using the electronic device 300, the key information interface UI can be opened to display key information. The key information interface UI is displayed on the top layer, covering all components displayed on the screen.

[0046] Please refer to Figure 4 , which depicts a key information interface UI of an electronic device 400 that is watching a video according to another embodiment. The electronic device 400 is, for example, a tablet computer. When a user is watching a video using the electronic device 400, the key information interface UI can be opened to display key information. The key information interface UI is displayed on the top layer, covering all components displayed on the screen.

[0047] The present disclosure does not limit the types of electronic devices used above, that is, laptop computers (such as clerical type, business type, game competition type, or special field type laptop computers), desktop computers, handheld game consoles, tablet computers, and any consumer electronic products with computing and interactive functions (such as those with display functions and / or adaptive feedback interactive signals) are all applicable. In addition, as electronic devices are constantly being updated, the versions and file names of application programs may also be updated, and it is difficult to directly use the application program to set the key information category CG and key information item IT that the key information interface UI needs to display.

[0048] like Figure 1 , Figure 3 and Figure 4 As shown, the window images FM11, FM3, and FM4 for playing games or watching videos have obvious pixel changes. Figure 2 As shown, the window screen FM12 for word processing has less obvious pixel changes. The type of application currently executed by the electronic device 100, 300, 400 can be inferred through the pixel changes.

[0049] Please refer to Figure 5 , which depicts a schematic diagram of a key information inference model MD according to an embodiment. In this embodiment, after the key information inference model MD is established, the key information category CG and the key information item IT of the key information category CG that need to be displayed in the key information interface UI can be directly inferred from the pixel data PX of the window screens FM11, FM12, FM3, and FM4. In other words, even if new electronic devices, or new games, new streaming media applications, or new tool programs are launched in the future, through the artificial intelligence technology of this embodiment, the key information category CG and the key information item IT that need to be displayed in the key information interface UI can be directly inferred using the pixel data PX.

[0050] Please refer to Figure 6 , which illustrates a state machine diagram of a key information inference model MD according to an embodiment. In training mode M1, the key information inference model MD is trained, and after the training is completed, the inference mode M2 ​​can be entered. In inference mode M2, the inference operation of the key information inference model MD can be performed. During the operation of inference mode M2, if the user modifies the inference result of the key information inference model MD, the optimization mode M3 is entered. In optimization mode M3, the key information inference model MD can be optimized according to the user's modification.

[0051] Please refer to Figure 7 , which illustrates a block diagram of an electronic device 700 according to an embodiment. The electronic device 700 includes an application layer operation unit 710, a data analysis unit 720, an artificial intelligence operation unit 730 and a hardware layer monitoring unit 740. The application layer operation unit 710 is used to operate various applications and detect the operation of the applications. The data analysis unit 720 is used to obtain data from various applications, create packets and perform analysis. The artificial intelligence operation unit 730 is used to train the above-mentioned key information inference model MD, operate the above-mentioned key information inference model MD or optimize the above-mentioned key information inference model MD. The hardware layer monitoring unit 740 is used to monitor the information of various hardware components. The application layer operation unit 710, the data analysis unit 720, the artificial intelligence operation unit 730 and / or the hardware layer monitoring unit 740 are, for example, a circuit, a chip, a circuit board or a storage device storing program code.

[0052] The following Figure 7 The electronic device 700 is used as an example to illustrate the operation of the training mode M1, the inference mode M2 ​​and the optimization mode M3. Figure 7 and Figure 8 , Figure 8 A flowchart of a method for training a key information inference model MD according to an embodiment is shown. In the method for training the key information inference model MD, a training data set is first obtained in steps S810 to S820, and then training is performed in step S830.

[0053] In step S810, the data analysis unit 720 obtains a plurality of pixel data PXi of at least one window screen FMi of each application program APi according to a plurality of application programs APi. The pixel data PXi, for example, includes a red pixel value, a green pixel value, a blue pixel value, and an alpha value. In this step, the pixel data PXi may be recorded into a packet in the order of rows and columns.

[0054] Alternatively, in one embodiment, the pixel data PXi is, for example, the distribution of red pixel values, green pixel values, blue pixel values, and opacity, for example, the proportion of X consecutive identical pixel values.

[0055] Next, in step S820, the data analysis unit 720 obtains the key information categories CGi and key information items ITi that need to be displayed on the key information interface UI when operating each application APi based on these application APi. For different application APi, the key information categories CGi and key information items ITi that need to be displayed may not be exactly the same. When training the key information inference model MD, the key information categories CGi and key information items ITi that need to be displayed can be set in advance according to the characteristics of different application APi.

[0056] Then, in step S830, the artificial intelligence operation unit 730 trains the key information inference model MD with the corresponding pixel data PXi, the key information category CGi and the key information item ITi. The key information inference model MD is trained, for example, with the distribution of the red pixel values, the green pixel values, the blue pixel values ​​and the opacity of the pixel data PXi. The artificial intelligence operation unit 730 is trained, for example, with a supervised learning algorithm, an unsupervised learning algorithm or a reinforcement learning algorithm.

[0057] In the embodiment using a supervised learning algorithm, the supervised learning algorithm can perform training and prediction based on a set of labeled examples provided. For example, the key information category CGi and key information item ITi in step S820 can be used as labels for training and prediction.

[0058] In the embodiment using an unsupervised learning algorithm, the data points are not labeled, and the unsupervised learning algorithm will label the data by organizing the data or describing its structure, so step S820 can be omitted.

[0059] In an embodiment using a reinforcement learning algorithm, learning can be done from the results and deciding the next action to take. After each action is completed, the reinforcement learning algorithm receives a response to help it determine whether the choice made is correct, neutral, or incorrect, and optimize the model accordingly. In this embodiment, the reinforcement learning algorithm can be used in optimization mode M3 to optimize the key information inference model MD. Optimization mode M3 is described after inference mode M2.

[0060] The training of the key information inference model MD is described by taking the execution in the electronic device 700 as an example. In another embodiment, the training of the key information inference model MD can also be completed by another independent server, and the trained key information inference model MD is set in the server instead of the electronic device 700.

[0061] After completing the training of the key information inference model MD through the above steps S810 to S830, the inference mode M2 ​​may be entered.

[0062] Please refer to Figure 7 and Fig. 9 , Fig. 9 A flow chart of a method for displaying a key information interface UI according to an embodiment is shown. In step S910, the application layer operation unit 710 determines whether a shortcut key KY is triggered. The shortcut key KY is, for example, a single physical key, a combination of several physical keys, a touch key of a touch panel, or a virtual key displayed in the air. If the shortcut key KY is triggered, then step S920 is entered.

[0063] In step S920, the application layer operation unit 710 detects a user's focus window WDj. When the user operates the electronic device 700, multiple applications may be opened. The application layer operation unit 710 detects the focus window WDj based on the current situation when the user presses the shortcut key KY. For example, the application layer operation unit 710 can use the topmost window as the focus window WDj. Alternatively, the application layer operation unit 710 can use the window with the largest range as the focus window WDj. Alternatively, the application layer operation unit 710 can use the window where a cursor is located as the focus window WDj.

[0064] Next, in step S930, the data analysis unit 720 obtains at least one window image FMj according to the focused window WDj. The data analysis unit 720 may obtain a single window image FMj when the user presses the shortcut key KY. Alternatively, the data analysis unit 720 may obtain a plurality of consecutive window images FMj when the user presses the shortcut key KY to avoid misjudgment of the inference caused by a sudden insertion of an advertisement or a special scene image of a film.

[0065] Then, in step S940, the data analysis unit 720 obtains a plurality of pixel data PXj according to the window image FMj. The pixel data PXj includes, for example, a red pixel value, a green pixel value, a blue pixel value, and an alpha value. In this step, the pixel data PXj may be recorded in the order of rows and columns.

[0066] Alternatively, in one embodiment, the pixel data PXj is, for example, the distribution of red pixel values, green pixel values, blue pixel values, and opacity, for example, the proportion of X consecutive identical pixel values.

[0067] Next, in step S950, the artificial intelligence operation unit 730 uses the key information inference model MD to infer at least one key information category CGj and at least one key information item ITj from the pixel data PXj. The key information inference model MD, for example, is inferred based on the distribution of the red pixel values, the green pixel values, the blue pixel values, and the opacity of the pixel data PXj.

[0068] Then, in step S960 , the hardware layer monitoring unit 740 extracts an information content CTj corresponding to each key information item ITj.

[0069] Next, in step S970 , the application layer operation unit 710 displays the key information category CGj, the key information item ITj and the information content CTj on the key information interface UI.

[0070] The inference of the above-mentioned key information inference model MD is explained by taking the execution in the electronic device 700 as an example. In another embodiment, the key information inference model MD can be set in the server instead of in the electronic device 700. When the electronic device 700 needs to perform inference on the key information interface UI, it can upload the pixel data PXj to the server through the network, and the key information category CGj and the key information item ITj are inferred by the key information inference model MD in the server. After the key information category CGj and the key information item ITj are transmitted back to the electronic device 700, the electronic device 700 then uses the hardware layer monitoring unit 740 to capture the information content CTj corresponding to each key information item ITj, so as to display the key information category CGj, key information item ITj and information content CTj on the key information interface UI.

[0071] Through the above steps S910 to S970, the inference of the content of the key information interface UI can be completed. The user can quickly check the required key information in the key information interface UI. When the user executes different applications, the key information interface UI can adaptively display the required key information category CGj, key information item ITj and information content CTj.

[0072] Since the present embodiment directly utilizes the pixel data PXj for inference, even when the version or file name of the application is updated, the key information category CGj, key information item ITj and information content CTj that need to be displayed in the key information interface UI can be adaptively inferred.

[0073] In some cases, the user may modify the key information interface UI, such as removing certain key information categories CGj or removing certain key information items ITj. At this time, the optimization mode M3 may be entered.

[0074] Please refer to Fig.10 and Fig.11 , Fig.10 FIG. 7 is a schematic diagram showing an artificial intelligence operation unit 730 according to an embodiment of the present invention. Fig.11 A flow chart of an optimization method of a key information inference model MD according to an embodiment is shown. The artificial intelligence operation unit 730 includes a state monitor 731 and an optimizer 732. The state monitor 731 is used to detect and monitor the inference results of the key information inference model MD. The optimizer 732 is used to optimize the key information inference model MD.

[0075] In step S111, the state monitor 731 is used to detect whether the key information category CGj or key information item ITj displayed on the key information interface UI is removed. If the key information category CGj or key information item ITj displayed on the key information interface UI is removed, the process proceeds to step S112.

[0076] In step S112, the optimizer 732 adjusts the weight value of the removed key information category CGj, the weight value of the removed key information item ITj, and the weight value of the current pixel data PXj to optimize the key information inference model MD. For example, the optimizer 732 can reduce the weight value of the key information category CGj to a first degree, reduce the weight value of the key information item ITj to a second degree, and reduce the weight value of the pixel data PXj to a third degree. The second degree is greater than the first degree, and the first degree is greater than the third degree. After the weight value is reduced, the probability of it being inferred can be reduced. In this embodiment, the weight value of the key information item ITj is reduced the most, so the probability of the key information item ITj being inferred in the future is greatly reduced.

[0077] After the artificial intelligence operation unit 730 optimizes the key information inference model MD in the electronic device 700 operated by the user, the key information inference model MD can be stored in the electronic device 700. In this way, the user can obtain an inference result that better meets the needs of the user when using the optimized key information inference model MD for inference next time.

[0078] The optimization of the above-mentioned key information inference model MD is explained by taking the execution in the electronic device 700 as an example. In another embodiment, the key information inference model MD can be set in the server instead of being set in the electronic device 700. When the electronic device 700 needs to optimize the key information interface UI, it can upload the removed key information category CGj or the removed key information item ITj to the server through the network, and the server optimizes the key information inference model MD.

[0079] According to the above embodiment, various electronic devices 700 and various application programs APi can establish key information inference models MD accordingly. After the user presses the shortcut key KY, the key information category CGj and key information item ITj that need to be displayed on the key information interface UI can be directly inferred from the pixel data PXj of the window screen WDj, and the corresponding information content CTj can be captured. Even if new electronic devices are constantly launched, or new games, new streaming media applications or new tool programs are constantly launched, through the artificial intelligence technology of this embodiment, the key information interface UI can be directly inferred using the pixel data PXj.

[0080] In addition, the key information inference model MD can be adaptively optimized according to the user's habits, so that when the optimized key information inference model MD is inferred, an inference result that better meets the needs of this user can be obtained.

[0081] In summary, although the present disclosure has been disclosed as above by way of embodiments, it is not intended to limit the present disclosure. A person with ordinary knowledge in the technical field to which the present disclosure belongs may make various changes and modifications without departing from the spirit and scope of the present disclosure. Therefore, the protection scope of the present disclosure shall be determined by the definition of the appended claims.

Claims

1. A method for displaying a key information interface, characterized in that: include: Determine whether a shortcut key is triggered; If the shortcut key is triggered, a focus window is detected; According to the focused window, obtaining at least one window image; Acquire a plurality of pixel data according to the at least one window image; Using a key information inference model, inferring at least one key information category and at least one key information item of each key information category from the pixel data; Retrieving information content corresponding to each key information item; as well as The at least one key information category, the at least one key information item and the at least one information content are displayed on the key information interface.

2. The method for displaying a key information interface according to claim 1, characterized in that: Each of the pixel data includes a red pixel value, a green pixel value, a blue pixel value and an opacity.

3. The method for displaying a key information interface as claimed in claim 2, characterized in that: The key information inference model is inferred based on the distribution of the red pixel values, the green pixel values, the blue pixel values ​​and the opacities of the pixel data.

4. The method for displaying a key information interface according to claim 1, characterized in that: The step of obtaining the focused window is to use the topmost window as the focused window.

5. The method for displaying a key information interface according to claim 1, characterized in that: The step of obtaining the focused window is to use the window with the largest area as the focused window.

6. The method for displaying a key information interface according to claim 1, characterized in that: The step of obtaining the focused window is to use the window where a cursor is located as the focused window.

7. A method for establishing a key information inference model, characterized in that: include: According to a plurality of applications, a plurality of pixel data of at least one window screen of each application is obtained; According to the applications, obtaining at least one key information category and at least one key information item of each key information category required to be displayed on a key information interface when operating each application; and The key information inference model is trained with the corresponding pixel data, the key information categories and the key information items.

8. The method for establishing a key information inference model as claimed in claim 7, characterized in that: Each of the pixel data includes a red pixel value, a green pixel value, a blue pixel value and an opacity.

9. The method for establishing a key information inference model as claimed in claim 8, characterized in that: The key information inference model is trained based on the distribution of the red pixel values, the green pixel values, the blue pixel values ​​and the opacities of the pixel data.

10. The method for establishing a key information inference model according to claim 8, characterized in that: Also includes: Determining whether the at least one key information category or the at least one key information item displayed on the key information interface is removed; and If the at least one key information category or the at least one key information item displayed on the key information interface is removed, the weight value of the removed at least one key information category, the weight value of the removed at least one key information item and the current weight value of the pixel data are adjusted to optimize the key information inference model.

11. The method for establishing a key information inference model according to claim 10, characterized in that: The weight value of the at least one key information category that is removed is reduced to a first degree, the weight value of the at least one key information item that is removed is reduced to a second degree, and the weight value of the current pixel data is reduced to a third degree, the second degree is greater than the first degree, and the first degree is greater than the third degree.

12. An electronic device, characterized in that: include: An application layer operation unit is used to detect whether a shortcut key is triggered. If the shortcut key is triggered, the application layer operation unit detects a focus window; A data analysis unit, used for obtaining at least one window frame of the concerned window, and obtaining a plurality of pixel data according to the at least one window frame; an artificial intelligence operation unit, which utilizes a key information inference model to infer at least one key information category of a key information interface and at least one key information item of each key information category from the pixel data; and A hardware layer monitoring unit is used to capture an information content corresponding to each of the key information items. The application layer operation unit is further used to display the at least one key information category, the at least one key information item and the at least one information content on the key information interface.

13. The electronic device according to claim 12, wherein: Each of the pixel data includes a red pixel value, a green pixel value, a blue pixel value and an opacity.

14. The electronic device according to claim 13, wherein: The key information inference model is inferred based on the distribution of the red pixel values, the green pixel values, the blue pixel values ​​and the opacities of the pixel data.

15. The electronic device according to claim 12, wherein: The application layer operation unit uses the uppermost window as the focus window.

16. The electronic device as claimed in claim 12, wherein: The application layer operation unit uses the window with the largest range as the focus window.

17. The electronic device as claimed in claim 12, wherein: The application layer operation unit uses a window where a cursor is located as the focus window.

18. The electronic device as claimed in claim 12, wherein: The AI ​​operation unit includes: a status monitor for detecting whether the at least one key information category or the at least one key information item displayed on the key information interface is removed; and An optimizer, if the at least one key information category or the at least one key information item displayed on the key information interface is removed, the optimizer adjusts the weight value of the removed at least one key information category, the weight value of the removed at least one key information item and the current weight value of the pixel data to optimize the key information inference model.

19. The electronic device as claimed in claim 18, characterized in that: The optimizer reduces the weight of the at least one key information category by a first degree, reduces the weight of the at least one key information item by a second degree, and reduces the weight of the pixel data by a third degree, wherein the second degree is greater than the first degree, and the first degree is greater than the third degree.