A multi-project switching method and system

Abstract

The application discloses a multi-project switching method, comprising the following steps: S1, controlling multiple project identifiers to be displayed in a first area of a man-machine interface in a rolling manner; S2, collecting a user image and analyzing the user image; when it is analyzed that the user switches from performing a predetermined gesture to stopping performing the predetermined gesture at T3 moment, locking the project identifier appearing in the first area at T1 moment in the first area. The multi-project switching method can shorten the average project switching time and greatly improve the operation efficiency without waiting for a fixed rolling period through natural action to intervene in the rolling process in real time.

Description

Technical Field

[0001] This invention relates to the field of machine vision, and more particularly to a method for switching between multiple items. Background Technology

[0002] In existing technologies, machine vision-based action recognition for switching between multiple items mainly includes the following approaches:

[0003] First, mapping is based on preset gestures, which are bound to project switching commands. However, too many commands lead to an excessive memory burden on users and are prone to accidental touches, resulting in a poor user experience.

[0004] Second, based on timing sensing, a certain action is used as an instruction, such as raising a hand for a few seconds to trigger a switch. However, it lacks a feedback mechanism, and users cannot intuitively perceive the switching process, which can easily lead to operation delays or misjudgments.

[0005] Third, the static interface selection relies on a paginated static icon arrangement, requiring multiple gestures to flip through pages to locate the target item, which is time-consuming and cannot meet real-time requirements.

[0006] Fourth, it is based on eye tracking, but it is severely affected by ambient light and has low recognition accuracy when the user's head moves.

[0007] The above four solutions share three common flaws:

[0008] First, it is inefficient and has a high rate of false triggering; simplifying gestures can easily lead to misoperation.

[0009] Secondly, reducing the false trigger rate will lead to more complex operations.

[0010] Third, the system fails to dynamically link user actions with the interface, making it difficult to match system responses with user intent.

[0011] Fourth, users cannot confirm whether the switching command was successfully triggered. This needs to be addressed urgently. Summary of the Invention

[0012] To address the technical problems existing in the background art, this invention proposes a multi-project switching method, comprising:

[0013] S1. Control the multiple item icons to be displayed in turn in the first area of ​​the human-machine interface;

[0014] S2. Collect user images and analyze them;

[0015] When it is determined that the user switched from executing the predetermined posture to stopping the predetermined posture at time T3, the item identifier that appeared in the first area at time T1 is locked in the first area.

[0016] Furthermore, S1 specifically includes:

[0017] S11. Collect user images and analyze them;

[0018] When it is determined that the user continuously performs a predetermined posture between time T1 and T2, multiple item icons are controlled to be displayed in turn in the first area of ​​the human-machine interface.

[0019] Furthermore, it also includes:

[0020] S3, the project corresponding to the project identifier for starting and locking.

[0021] Furthermore, the predetermined posture is specifically raising one's hand.

[0022] Furthermore, a multi-item switching system is proposed, characterized by comprising:

[0023] The display control module is used to control the sequential scrolling display of multiple item icons in the first area of ​​the human-machine interface.

[0024] The image acquisition module is configured to acquire user images in real time via a camera;

[0025] The attitude analysis module, connected to the image acquisition module, is used to analyze user images and determine whether the user has switched from executing a predetermined attitude to stopping executing a predetermined attitude at time T3.

[0026] The locking module, when detecting the attitude change, locks the project identifier that appears in the first area at time T1 in the first area;

[0027] The project execution module is used to start and lock the project corresponding to the project identifier.

[0028] Furthermore, the display control module is configured to trigger the scrolling display of multiple item identifiers when the attitude analysis module determines that the user is continuously performing a predetermined attitude between time T1 and T2.

[0029] Furthermore, the predetermined posture is a hand-raising action. The posture analysis module uses a skeletal key point detection algorithm to identify whether the angle between the user's arms and the vertical direction is greater than a preset threshold.

[0030] Furthermore, it also includes a feedback module for generating visual or tactile feedback signals when the project identifier is locked.

[0031] Furthermore, a computer-readable storage medium is proposed, on which a computer program is stored, which, when executed by a processor, implements the multi-item switching method as described in any one of claims 1-4. In this invention, the proposed multi-item switching method allows users to intervene in the scrolling process in real time through natural actions without waiting for a fixed scrolling cycle, thus shortening the average item switching time and greatly improving operational efficiency. Attached Figure Description

[0032] Figure 1 This is a schematic diagram illustrating the steps of a multi-item switching method proposed in this invention;

[0033] Figure 2 This is a schematic diagram illustrating the steps of a multi-item switching method proposed in this invention. Detailed Implementation

[0034] This invention proposes a multi-item switching method based on pose recognition, comprising the following steps:

[0035] S1. Control multiple item icons to be displayed in turn in the first area of ​​the human-machine interface.

[0036] Specifically, by using timestamp synchronization technology, the interface scrolling speed is set to 3-5 item markers per second (configurable), ensuring that each marker enters the visible range of the first area within a preset time window (e.g., T1±200ms). Asynchronous rendering technology is used to separate the interface layers, and the scrolling animation frame rate is kept stable at over 60fps, which can avoid visual misalignment caused by stuttering.

[0037] S1 specifically includes: S11, acquiring user images, analyzing user images, and when it is determined that the user continuously performs a predetermined posture between time T1 and T2, controlling multiple item icons to be displayed in turn in the first area of ​​the human-machine interface.

[0038] Specifically: During the T1-T2 time period, the posture analysis module performs posture verification at preset time intervals. If the hand-raising angle meets the threshold 10 consecutive times, it is considered a valid trigger. A single detection failure tolerance (such as brief occlusion) is introduced, but the cumulative number of failures must be lower than a preset value. The scrolling speed can be adaptively adjusted based on user history data. For example, the default speed is 3 items / second for first-time users, which can be increased to 8 items / second for experienced users. This can be dynamically adjusted based on user lock success rate analysis using a machine learning model. Dual verification (duration + angle threshold) reduces the false trigger rate. Dynamic speed adjustment improves the success rate for novice users and increases efficiency for experienced users.

[0039] The predetermined posture is specifically raising a hand; specifically, a modified version of the OpenPose model is used to achieve real-time detection on an embedded device. Even under conditions of changing lighting and partial occlusion, the recognition accuracy is still higher than that of traditional solutions. The modified version of the OpenPose model reduces the consumption of computing resources and can run with low memory.

[0040] S2. Collect user images and analyze them. When the analysis shows that the user switched from executing a predetermined posture to stopping executing a predetermined posture at time T3, lock the item identifier that appeared in the first area at time T1 in the first area.

[0041] Specifically, an action timeline database needs to be constructed to record precise timestamps for T1 (scrolling trigger moment), T2 (scrolling stop moment), and T3 (user's stopping posture moment). A time backtracking algorithm is used to correlate and match the action termination signal at time T3 with the interface state at time T1, controlling the error range. When the system detects the user moving from a raised hand to a lowered arm (time T3), it captures a snapshot of the interface at time T1 and uses an image matching algorithm to identify the item identifier within the currently locked area. By using timestamp synchronization and backtracking technology, the matching error between action and interface state is reduced, solving the false locking problem caused by system latency in traditional solutions. Users no longer need to wait for a fixed scrolling cycle; they can intervene in the scrolling process in real time through natural actions, shortening the average item switching time and significantly improving operational efficiency.

[0042] S3, the project corresponding to the project identifier for starting and locking.

[0043] Specifically, after locking, a unique project ID hash value is generated, and the execution instruction is sent to the system kernel through inter-process communication (IPC) to ensure isolation from existing processes (avoiding resource conflicts). Preloading technology is adopted to partially cache the resources of adjacent projects during the rolling process, thereby compressing the project startup time and realizing full-link automation of "selection-locking-execution". The user operation steps are reduced, and the preloading technology can reduce the peak CPU utilization and memory fragmentation rate.

[0044] This invention proposes a multi-item switching system, comprising:

[0045] The display control module is used to control the sequential scrolling display of multiple item icons in the first area of ​​the human-machine interface.

[0046] The image acquisition module is configured to acquire user images in real time via a camera;

[0047] The attitude analysis module, connected to the image acquisition module, is used to analyze user images and determine whether the user has switched from executing a predetermined attitude to stopping executing a predetermined attitude at time T3.

[0048] The locking module, when detecting the attitude change, locks the project identifier that appears in the first area at time T1 in the first area;

[0049] The project execution module is used to start and lock the project corresponding to the project identifier.

[0050] Furthermore, the display control module is configured to trigger the scrolling display of multiple item identifiers when the attitude analysis module determines that the user is continuously performing a predetermined attitude between time T1 and T2.

[0051] Furthermore, the predetermined posture is a hand-raising action. The posture analysis module uses a skeletal key point detection algorithm to identify whether the angle between the user's arms and the vertical direction is greater than a preset threshold.

[0052] Furthermore, it also includes a feedback module for generating visual or tactile feedback signals when the project identifier is locked.

[0053] The image acquisition module supports multi-camera fusion (RGB + depth), eliminating viewpoint distortion through calibration algorithms to acquire a wide-angle view with a preset peak coverage. The feedback module integrates multimodal output, LED aperture color change (visual), and reduced feedback latency. A spatiotemporal convolutional network is employed to continuously analyze pose change trends in a 5-frame image sequence, resisting transient noise interference. Dynamic angle threshold calibration is supported, automatically adjusting the angle determination threshold based on the user's distance from the camera's depth information. Multi-sensor fusion improves the effective recognition rate in complex environments, enhancing system robustness; multimodal feedback improves the efficiency of operation confirmation perception, especially suitable for visually impaired users, thereby expanding the user experience dimension.

[0054] Furthermore, a computer-readable storage medium is proposed, on which a computer program is stored, which, when executed by a processor, implements a multi-item switching method.

[0055] The core algorithm is encapsulated into a neural network processing unit instruction set, and the quantization model is pre-installed in the storage medium, which improves the inference speed. The differential update mechanism allows only the incremental package of algorithm parameters to be downloaded instead of the complete program. It supports offline operation, has high functional availability in weak network environments, and offers deployment flexibility. Differential updates reduce monthly maintenance traffic consumption, making it suitable for large-scale deployment of IoT devices with low maintenance costs.

[0056] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A method for switching between multiple items, characterized in that, include: S1. Control the multiple item icons to be displayed in turn in the first area of ​​the human-machine interface; S2. Collect user images and analyze them; When it is determined that the user switched from executing the predetermined posture to stopping the predetermined posture at time T3, the item identifier that appeared in the first area at time T1 is locked in the first area.

2. The multi-item switching method as described in claim 1, characterized in that, S1 specifically includes: S11. Collect user images and analyze them; When it is determined that the user continuously performs a predetermined posture between time T1 and T2, multiple item icons are controlled to be displayed in turn in the first area of ​​the human-machine interface.

3. The multi-item switching method as described in claim 1, characterized in that, Also includes: S3, the project corresponding to the project identifier for starting and locking.

4. The multi-item switching method as described in claim 1, characterized in that, The intended posture is to raise one's hand.

5. A multi-item switching system, characterized in that, include: The display control module is used to control the sequential scrolling display of multiple item icons in the first area of ​​the human-machine interface. The image acquisition module is configured to acquire user images in real time via a camera; The attitude analysis module, connected to the image acquisition module, is used to analyze user images and determine whether the user has switched from executing a predetermined attitude to stopping executing a predetermined attitude at time T3. The locking module, when detecting the attitude change, locks the project identifier that appears in the first area at time T1 in the first area; The project execution module is used to start and lock the project corresponding to the project identifier.

6. The multi-item switching system according to claim 5, characterized in that, The display control module is configured to trigger the scrolling display of multiple item identifiers when the attitude analysis module determines that the user is continuously performing a predetermined attitude between time T1 and T2.

7. The multi-item switching system according to claim 5, characterized in that, The predetermined posture is a raised hand gesture. The posture analysis module uses a skeletal key point detection algorithm to identify whether the angle between the user's arms and the vertical direction is greater than a preset threshold.

8. The multi-item switching system according to claim 5, characterized in that, It also includes a feedback module for generating visual or tactile feedback signals when the project identifier is locked.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the multi-item switching method as described in any one of claims 1-4.

Images