Sensor-based intelligent ring gesture recognition method and system

Abstract

The application relates to the technical field of human-computer interaction, in particular to a sensor-based intelligent ring gesture recognition method and system, which comprises the steps of acquiring hand state information, motion trajectory monitoring, posture change monitoring, gesture trigger condition judgment, pressure feature extraction and matching, dynamic optimization and function mapping and the like. Through multi-sensor cooperative collection and data fusion analysis, the application improves the real-time performance and stability of gesture recognition; the dynamic optimization mechanism is used to enhance the adaptability in a complex environment, and the gesture function association table is used to dynamically update user habits and optimize the recognition strategy. The application can significantly improve the accuracy and adaptability of gesture recognition and meet the needs of diversified application scenarios.

Description

Technical Field

[0001] This invention belongs to the field of human-computer interaction technology, specifically a sensor-based intelligent ring gesture recognition method and system. Background Technology

[0002] With the rapid development of smart wearable devices, gesture recognition technology based on smart rings has gradually become an important research direction in the field of human-computer interaction due to its portability and strong interactivity. As a miniaturized and lightweight wearable device, the smart ring can capture and recognize hand gestures by integrating multiple sensors, thus providing users with a natural and intuitive way to interact with computers. This technology shows broad application prospects in scenarios such as virtual reality (VR), augmented reality (AR), smart home control, and accessibility assistive devices. However, existing gesture recognition methods based on smart rings still have shortcomings in terms of sensor data acquisition, processing efficiency, and recognition accuracy, affecting user experience and system performance, and limiting their further promotion in diverse application scenarios.

[0003] A search revealed that CN111766941B, a method and system for gesture recognition based on a smart ring (published on November 9, 2021), proposes a technical solution that integrates a motion sensor and a camera to collect motion data of the wearing hand and image data of the non-wearing hand, and then uses a microprocessor to analyze the two types of data to achieve bi-handed gesture recognition. While this solution achieves bi-handed gesture recognition to a certain extent, the computational complexity is high due to the need to process both motion and image data simultaneously, which may lead to a decrease in real-time performance, especially in high-frequency gesture interaction scenarios. Furthermore, this solution relies on a camera to collect image data of the non-wearing hand, which may be affected by changes in ambient light or occlusion, thus reducing the stability and accuracy of gesture recognition. These limitations significantly restrict the application of this technology in dynamic environments or under complex lighting conditions.

[0004] Another related technology is a gesture monitoring method and system based on a smart ring, published on April 8, 2025 (CN117111732B). This patent improves the accuracy of gesture monitoring and data processing efficiency by optimizing the sampling frequency of the sensor and dynamically adjusting the sampling strategy based on the changes between current data and reference data. However, this technical solution mainly focuses on the monitoring of single-handed gestures and does not fully consider the recognition needs of two-handed coordinated gestures. In addition, its optimization of the sampling frequency relies on preset gesture command sequences, which may have certain limitations when dealing with complex or custom gestures, making it difficult to meet the needs of diverse application scenarios. For example, when users need to dynamically adjust gesture commands or use personalized gestures for interaction, this solution may not provide sufficient flexibility and adaptability.

[0005] The aforementioned problems indicate that existing smart ring-based gesture recognition technologies still have shortcomings in areas such as multi-hand collaborative recognition, real-time performance, environmental adaptability, and the ability to handle complex gestures. Specifically, existing technologies face challenges in the following aspects: First, the sensor data acquisition mechanism lacks flexibility, making it difficult to dynamically adjust the sampling strategy according to actual application scenarios, leading to data redundancy or loss of key information; second, data processing efficiency is low, especially in high-frequency gesture interaction scenarios, making it difficult to guarantee real-time performance; third, existing technologies have poor adaptability to environmental changes, particularly under complex lighting conditions or in dynamic environments, where recognition accuracy and stability significantly decrease; finally, existing solutions have limitations in supporting diverse gesture needs, making it difficult to meet users' recognition requirements for complex or custom gestures.

[0006] Therefore, this invention aims to improve the real-time performance, stability, and adaptability of the system by optimizing the sensor data acquisition and processing mechanism, while supporting diverse gesture recognition needs to meet the requirements of modern human-computer interaction for efficient and accurate gesture recognition technology. This invention will focus on solving problems existing in current technologies such as high computational complexity, poor environmental adaptability, insufficient bimanual recognition capability, and limited ability to process complex gestures, providing an innovative solution for the development of smart ring gesture recognition technology. Summary of the Invention

[0007] This invention provides a sensor-based smart ring gesture recognition method, the main purpose of which is to improve the real-time performance, stability and adaptability of gesture recognition, while supporting diverse gesture needs to meet the requirements of modern human-computer interaction for efficient and accurate gesture recognition technology.

[0008] To achieve the above objectives, this invention provides a sensor-based smart ring gesture recognition method, comprising: acquiring hand state information of a user wearing a smart ring, wherein the smart ring includes a multi-axis accelerometer, a gyroscope, and a pressure sensor; using the multi-axis accelerometer to monitor the motion trajectory of the user's hand to obtain a first motion data sequence; using the gyroscope to monitor the posture changes of the user's hand to obtain a second posture data sequence; determining whether the first motion data sequence and the second posture data sequence meet a preset gesture triggering condition; when it is determined that the first motion data sequence and the second posture data sequence meet the gesture triggering condition, using the pressure sensor to acquire a pressure distribution map of the user's fingers, and performing feature extraction on the pressure distribution map to obtain a pressure feature set; and performing gesture matching on the pressure feature set according to a pre-built gesture feature library. The process involves: obtaining the target gesture; when the first motion data sequence and the second posture data sequence do not meet the gesture triggering condition, combining the first motion data sequence and the second posture data sequence to generate a comprehensive motion feature vector, and dynamically adjusting and optimizing the comprehensive motion feature vector to obtain an optimized motion feature vector; using a pre-built gesture classification model to predict gestures on the optimized motion feature vector to obtain candidate gestures; determining whether the target gesture or the candidate gesture is detected; if the target gesture or the candidate gesture is not detected, returning to the steps described above of using a multi-axis accelerometer to monitor the user's hand motion trajectory and obtain the first motion data sequence; and when the target gesture or the candidate gesture is detected, using a pre-built gesture function association table to perform function mapping on the target gesture or candidate gesture to obtain the target function, and executing the target function.

[0009] Optionally, the step of performing feature extraction on the pressure distribution map to obtain a pressure feature set includes: performing image segmentation processing on the pressure distribution map to obtain multiple pressure regions; performing a pressure intensity-based grading operation on the multiple pressure regions to obtain a pressure level map; performing shape analysis on each pressure region in the pressure level map to obtain a pressure shape feature set; and fusing the pressure shape feature set with the pressure level map to obtain a pressure feature set.

[0010] Optionally, the step of performing gesture matching on the pressure feature set according to the pre-constructed gesture feature library to obtain the target gesture includes: performing directional analysis on each pressure feature in the pressure feature set to obtain a directional feature set; evaluating the spatial distribution of the directional feature set according to a preset spatial position weight to obtain a spatial distribution feature set; comparing the similarity of the spatial distribution feature set with reference gesture features in the gesture feature library to obtain the reference gesture feature with the highest matching degree; and defining the gesture corresponding to the reference gesture feature with the highest matching degree as the target gesture.

[0011] Optionally, the step of combining the first motion data sequence and the second posture data sequence to generate a comprehensive motion feature vector includes: performing time series analysis on the first motion data sequence to obtain motion trend features; performing angle change analysis on the second posture data sequence to obtain posture change features; performing feature concatenation operation on the motion trend features and the posture change features to obtain an initial comprehensive motion feature vector; and performing dimensionality reduction processing on the initial comprehensive motion feature vector based on principal component analysis to obtain a comprehensive motion feature vector.

[0012] Optionally, the step of using a pre-built gesture classification model to predict gestures from the optimized motion feature vector to obtain candidate gestures includes: inputting the optimized motion feature vector into the pre-built gesture classification model to obtain preliminary gesture prediction results; evaluating the confidence of the preliminary gesture prediction results to obtain a confidence score; performing a filtering operation on the confidence score according to a preset confidence threshold to obtain a set of high-confidence gestures; and selecting the gesture with the highest frequency of occurrence from the set of high-confidence gestures as candidate gestures.

[0013] Optionally, before utilizing the pre-built gesture function association table, the method further includes: when starting the pre-built gesture input service, acquiring the user's repeated gesture actions for the target device; extracting features from the repeated gesture actions to obtain input gesture features; acquiring the gesture names configured by the user and the corresponding device function links, and constructing the association relationship between the device function links and the input gesture features; and inputting the association relationship into the pre-built gesture function association table according to the gesture names.

[0014] Optionally, after determining whether the target gesture or the candidate gesture has been detected, the method further includes: acquiring successfully detected target gestures or candidate gestures within a preset time period to obtain historical gesture records; identifying gesture usage patterns in the historical gesture records; sorting the input gesture features in the gesture function association table according to the gesture usage patterns based on the priority of usage frequency to obtain updated gesture features; and updating the association relationships in the gesture function association table according to the updated gesture features.

[0015] To achieve the above objectives, the present invention also provides a sensor-based smart ring gesture recognition system, comprising: a hand state acquisition module for acquiring hand state information of a user wearing a smart ring, wherein the smart ring includes a multi-axis accelerometer, a gyroscope, and a pressure sensor; a motion monitoring module for monitoring the motion trajectory of the user's hand using the multi-axis accelerometer to obtain a first motion data sequence, and monitoring the posture changes of the user's hand using the gyroscope to obtain a second posture data sequence; and a gesture recognition module for determining whether the first motion data sequence and the second posture data sequence meet preset gesture triggering conditions, and when it is determined that the first motion data sequence and the second posture data sequence meet the gesture triggering conditions, acquiring a pressure distribution map of the user's fingers using the pressure sensor, and performing feature extraction on the pressure distribution map. A pressure feature set is obtained, and a gesture matching is performed on the pressure feature set according to a pre-built gesture feature library to obtain a target gesture; a dynamic optimization module is used to generate a comprehensive motion feature vector by combining the first motion data sequence and the second posture data sequence when it is determined that the first motion data sequence and the second posture data sequence do not meet the gesture triggering condition, and to dynamically adjust and optimize the comprehensive motion feature vector to obtain an optimized motion feature vector; a function mapping module is used to return to the above steps of using a multi-axis accelerometer to monitor the motion trajectory of the user's hand to obtain a first motion data sequence when the target gesture or the candidate gesture is not detected, and when the target gesture or the candidate gesture is detected, to perform function mapping on the target gesture or the candidate gesture using a pre-built gesture function association table to obtain a target function, and to execute the target function.

[0016] To address the aforementioned problems, the present invention also provides an electronic device comprising: a memory storing at least one instruction; and a processor executing the instruction stored in the memory to implement the sensor-based smart ring gesture recognition method described above.

[0017] To address the aforementioned problems, the present invention also provides a computer-readable storage medium storing at least one instruction, which is executed by a processor in an electronic device to implement the sensor-based smart ring gesture recognition method described above.

[0018] To address the problems described in the background section, this invention first employs a multi-sensor collaborative acquisition method to comprehensively perceive the user's hand state. Specifically, to improve the real-time performance and stability of gesture recognition, this invention achieves accurate judgment of gesture triggering conditions through the fusion analysis of pressure sensor data and motion sensor data. Simultaneously, a dynamic optimization mechanism enhances gesture recognition capabilities in complex environments. Furthermore, this invention introduces a dynamic update mechanism for the gesture function association table, which can optimize the gesture recognition strategy based on user habits, thereby further improving the system's adaptability and user experience. Therefore, this invention can significantly improve the real-time performance, stability, and adaptability of gesture recognition, meeting the needs of diverse application scenarios. Attached Figure Description

[0019] Figure 1 This is a schematic flowchart of a sensor-based smart ring gesture recognition method provided in an embodiment of the present invention;

[0020] Figure 2 A functional block diagram of a sensor-based intelligent ring gesture recognition system provided in an embodiment of the present invention;

[0021] Figure 3 This is a schematic diagram of the structure of an electronic device that implements the sensor-based smart ring gesture recognition method according to an embodiment of the present invention. Detailed Implementation

[0022] This invention provides a sensor-based smart ring gesture recognition method and system, aiming to improve the real-time performance, stability, and adaptability of gesture recognition through multi-sensor collaborative data acquisition and data fusion analysis, and to meet the demands of modern human-computer interaction for efficient and accurate gesture recognition technology. The following is in conjunction with the appendix... Figure 1 To be continued Figure 3 The specific embodiments of the present invention will be described in detail.

[0023] like Figure 1 The diagram illustrates a flowchart of a sensor-based smart ring gesture recognition method according to an embodiment of the present invention. The method first acquires hand state information through a smart ring worn on the user's finger. The smart ring integrates a multi-axis accelerometer, a gyroscope, and a pressure sensor. These sensors are used to monitor hand movement trajectory, posture changes, and finger pressure distribution, respectively. In practical applications, after the user wears the smart ring, the system initializes and activates the sensor modules to ensure accurate data acquisition. The multi-axis accelerometer captures the hand's movement trajectory in three-dimensional space and generates a first motion data sequence; the gyroscope monitors hand posture changes and generates a second posture data sequence; and the pressure sensor records the pressure distribution applied by the fingers. Through the data acquisition from these sensors, the system can comprehensively perceive the user's hand state.

[0024] After acquiring the first motion data sequence and the second posture data sequence, the system determines whether these data meet preset gesture triggering conditions. Gesture triggering conditions are threshold ranges or pattern templates pre-defined based on the motion characteristics and posture change patterns of common gestures. For example, when a user makes a specific gesture, their hand movement trajectory and posture changes usually exhibit certain regularities, which can be extracted through data analysis and used as triggering conditions. If the first motion data sequence and the second posture data sequence meet the preset gesture triggering conditions, the system proceeds to the next step: acquiring a pressure distribution map of the user's fingers using a pressure sensor and performing feature extraction. Feature extraction includes image segmentation of the pressure distribution map to obtain multiple pressure regions; then, a pressure intensity-based grading operation is performed on these pressure regions to generate a pressure level map; next, shape analysis is performed on each pressure region to obtain a pressure shape feature set; finally, the pressure shape feature set is fused with the pressure level map to form a complete pressure feature set describing the pressure distribution characteristics. The core of this process lies in extracting key information reflecting gesture characteristics through multi-level analysis of the pressure distribution map.

[0025] After obtaining the pressure feature set, the system matches gestures against a pre-built gesture feature library. This library stores various reference gesture features, generated through training and optimization on a large number of gesture samples. The matching process involves performing directional analysis on each pressure feature in the pressure feature set to obtain a directional feature set; then, based on preset spatial position weights, evaluating the spatial distribution of the directional feature set to generate a spatial distribution feature set; finally, comparing the similarity of the spatial distribution feature set with the reference gesture features in the gesture feature library, identifying the reference gesture feature with the highest matching degree, and defining its corresponding gesture as the target gesture. The key to this matching process is improving the accuracy of gesture recognition through a comprehensive evaluation of directionality and spatial distribution. To further illustrate, suppose a user makes a "pinch" gesture; its pressure distribution features might be concentrated in the fingertip area, and the directional features show a tendency to converge towards the center. Through the above matching algorithm, the system can quickly identify this gesture and define it as the target gesture.

[0026] If the first motion data sequence and the second posture data sequence do not meet the preset gesture triggering conditions, the system will combine these two sets of data to generate a comprehensive motion feature vector. The generation process of the comprehensive motion feature vector includes performing time series analysis on the first motion data sequence to extract motion trend features; simultaneously performing angle change analysis on the second posture data sequence to extract posture change features; then concatenating the motion trend features and posture change features to generate an initial comprehensive motion feature vector; finally, performing dimensionality reduction processing based on principal component analysis on the initial comprehensive motion feature vector to obtain the final comprehensive motion feature vector. Principal component analysis is a commonly used dimensionality reduction method. Its core idea is to project high-dimensional data into a low-dimensional space through linear transformation while preserving the main features of the data. The formula is as follows:

[0027] Y = XW

[0028] Where X represents the original data matrix, W represents the principal component projection matrix, and Y represents the dimensionality-reduced data matrix. Through principal component analysis, the system can retain key information in the comprehensive motion feature vector while reducing computational complexity.

[0029] After obtaining the comprehensive motion feature vector, the system dynamically adjusts and optimizes it to improve gesture recognition capabilities in complex environments. The optimization process includes noise filtering and feature enhancement operations on the comprehensive motion feature vector. The optimized motion feature vector is then input into a pre-built gesture classification model to predict candidate gestures. The gesture classification model can employ deep learning algorithms, such as convolutional neural networks or recurrent neural networks, to achieve efficient classification tasks. Specifically, the system inputs the optimized motion feature vector into the gesture classification model to obtain preliminary gesture prediction results; then, it evaluates the confidence of the preliminary gesture prediction results to obtain a confidence score; based on a preset confidence threshold, it selects a set of high-confidence gestures; finally, it selects the gesture with the highest frequency from the high-confidence gesture set as a candidate gesture. The core of this process lies in improving the reliability of gesture prediction through confidence evaluation and frequency statistics.

[0030] After recognizing the target or candidate gesture, the system will determine whether a valid gesture has been detected. If no target or candidate gesture is detected, the system will return to the previous step and re-monitor the user's hand motion trajectory using a multi-axis accelerometer to generate a new first motion data sequence. If the target or candidate gesture is detected, the system will use a pre-built gesture function association table to map the gesture to a function, obtain the target function, and execute that function. The gesture function association table is generated through user-defined configuration. Its construction process includes starting a pre-built gesture input service, acquiring the user's repeated gesture actions for the target device; extracting features from these repeated gesture actions to obtain input gesture features; subsequently, acquiring the user-configured gesture name and corresponding device function link, constructing the association between the device function link and the input gesture features; and finally, recording the association in the pre-built gesture function association table based on the gesture name. For example, a user can define a "wave" gesture as corresponding to the "channel switching" function on a TV remote control and store this association in the gesture function association table. In practical applications, when the system detects a "wave" gesture, it will automatically execute the "channel switching" function.

[0031] To further enhance the system's adaptability and user experience, this invention also introduces a dynamic update mechanism for the gesture function association table. After determining whether a target gesture or candidate gesture has been detected, the system acquires successfully detected target gestures or candidate gestures within a preset time period, generating historical gesture records. Subsequently, it identifies gesture usage patterns in the historical gesture records and prioritizes the input gesture features in the gesture function association table based on usage frequency according to these patterns, obtaining updated gesture features. Finally, based on the updated gesture features, the association relationships in the gesture function association table are updated. The core of this dynamic update mechanism lies in continuously optimizing the gesture recognition strategy by analyzing users' gesture usage habits, thereby improving the system's intelligence level.

[0032] like Figure 2 The diagram shows a functional block diagram of a sensor-based smart ring gesture recognition system according to an embodiment of the present invention. The system includes a hand state acquisition module, a motion monitoring module, a gesture recognition module, a dynamic optimization module, and a function mapping module. The hand state acquisition module is responsible for collecting the user's hand state information through a multi-axis accelerometer, gyroscope, and pressure sensor in the smart ring; the motion monitoring module is responsible for monitoring and analyzing the first motion data sequence and the second posture data sequence; the gesture recognition module is responsible for determining the gesture triggering conditions and matching the pressure feature set; the dynamic optimization module is responsible for generating and optimizing the comprehensive motion feature vector; and the function mapping module is responsible for executing the target function according to the gesture function association table. The modules work closely together through data flow and control flow to jointly complete the gesture recognition task.

[0033] like Figure 3 The diagram shows the structure of an electronic device implementing a sensor-based smart ring gesture recognition method according to an embodiment of the present invention. The electronic device includes a memory and a processor, wherein the memory stores at least one instruction, and the processor executes the instructions stored in the memory to implement the gesture recognition method described above. In practical applications, the electronic device can be a smartphone, tablet computer, or other terminal device with computing capabilities. By deploying the gesture recognition algorithm on the electronic device, users can flexibly use the smart ring for gesture interaction in different scenarios, thereby significantly improving the convenience and efficiency of human-computer interaction.

[0034] In summary, this invention achieves efficient and accurate gesture recognition through multi-sensor collaborative acquisition, data fusion analysis, and dynamic optimization mechanisms. Furthermore, by introducing a dynamic update mechanism for the gesture function association table, the system can continuously optimize the gesture recognition strategy based on user habits, thereby further improving the system's adaptability and user experience. In addition, this invention has a wide range of applications, applicable to multiple fields such as smart homes, virtual reality, and medical rehabilitation, possessing significant practical value and market potential.

Claims

1. A sensor-based smart ring gesture recognition method, characterized in that, The method includes: acquiring hand state information of a user wearing a smart ring, wherein the smart ring includes a multi-axis accelerometer, a gyroscope, and a pressure sensor; monitoring the motion trajectory of the user's hand using the multi-axis accelerometer to obtain a first motion data sequence; monitoring the posture changes of the user's hand using the gyroscope to obtain a second posture data sequence; determining whether the first motion data sequence and the second posture data sequence meet preset gesture triggering conditions; when it is determined that the first motion data sequence and the second posture data sequence meet the gesture triggering conditions, acquiring a pressure distribution map of the user's fingers using the pressure sensor, and performing feature extraction on the pressure distribution map to obtain a pressure feature set; performing gesture matching on the pressure feature set according to a pre-constructed gesture feature library to obtain a target gesture ... and performing gesture matching on the pressure distribution map according to a pre-constructed gesture feature library to obtain a target gesture. When the first motion data sequence and the second posture data sequence do not meet the gesture triggering condition, a comprehensive motion feature vector is generated by combining the first motion data sequence and the second posture data sequence, and the comprehensive motion feature vector is dynamically adjusted and optimized to obtain an optimized motion feature vector; a pre-built gesture classification model is used to predict gestures using the optimized motion feature vector to obtain candidate gestures; it is determined whether the target gesture or the candidate gesture is detected; when the target gesture or the candidate gesture is not detected, the process returns to the steps described above of using a multi-axis accelerometer to monitor the user's hand motion trajectory to obtain the first motion data sequence; when the target gesture or the candidate gesture is detected, a pre-built gesture function association table is used to perform function mapping on the target gesture or candidate gesture to obtain the target function and execute the target function.

2. The sensor-based smart ring gesture recognition method as described in claim 1, characterized in that, The step of extracting features from the pressure distribution map to obtain a pressure feature set includes: performing image segmentation on the pressure distribution map to obtain multiple pressure regions; performing a pressure intensity-based grading operation on the multiple pressure regions to obtain a pressure level map; performing shape analysis on each pressure region in the pressure level map to obtain a pressure shape feature set; and fusing the pressure shape feature set with the pressure level map to obtain a pressure feature set.

3. The sensor-based smart ring gesture recognition method as described in claim 1, characterized in that, The step of matching the pressure feature set with a pre-constructed gesture feature library to obtain the target gesture includes: performing directional analysis on each pressure feature in the pressure feature set to obtain a directional feature set; evaluating the spatial distribution of the directional feature set according to a preset spatial position weight to obtain a spatial distribution feature set; comparing the spatial distribution feature set with reference gesture features in the gesture feature library to obtain the reference gesture feature with the highest matching degree; and defining the gesture corresponding to the reference gesture feature with the highest matching degree as the target gesture.

4. The sensor-based smart ring gesture recognition method as described in claim 1, characterized in that, The step of generating a comprehensive motion feature vector by combining the first motion data sequence and the second posture data sequence includes: performing time series analysis on the first motion data sequence to obtain motion trend features; performing angle change analysis on the second posture data sequence to obtain posture change features; performing feature concatenation operation on the motion trend features and the posture change features to obtain an initial comprehensive motion feature vector; and performing dimensionality reduction processing based on principal component analysis on the initial comprehensive motion feature vector to obtain a comprehensive motion feature vector.

5. The sensor-based smart ring gesture recognition method as described in claim 1, characterized in that, The step of using a pre-built gesture classification model to predict gestures from the optimized motion feature vector to obtain candidate gestures includes: inputting the optimized motion feature vector into the pre-built gesture classification model to obtain a preliminary gesture prediction result; evaluating the confidence of the preliminary gesture prediction result to obtain a confidence score; filtering the confidence score according to a preset confidence threshold to obtain a set of high-confidence gestures; and selecting the gesture with the highest frequency from the set of high-confidence gestures as a candidate gesture.

6. The sensor-based smart ring gesture recognition method as described in claim 1, characterized in that, Before utilizing the pre-built gesture function association table, the method further includes: acquiring repeated gesture actions of the user on the target device when starting the pre-built gesture input service; extracting features from the repeated gesture actions to obtain input gesture features; acquiring the gesture names configured by the user and the corresponding device function links to construct the association relationship between the device function links and the input gesture features; and inputting the association relationship into the pre-built gesture function association table according to the gesture names.

7. The sensor-based smart ring gesture recognition method as described in claim 1, characterized in that, After determining whether the target gesture or the candidate gesture has been detected, the method further includes: obtaining historical gesture records by acquiring successfully detected target gestures or candidate gestures within a preset time period; identifying gesture usage patterns in the historical gesture records; sorting the input gesture features in the gesture function association table according to the gesture usage patterns based on the priority of usage frequency to obtain updated gesture features; and updating the association relationships in the gesture function association table according to the updated gesture features.

8. A sensor-based intelligent ring gesture recognition system, characterized in that, The system includes: a hand state acquisition module for acquiring hand state information of a user wearing a smart ring, wherein the smart ring includes a multi-axis accelerometer, a gyroscope, and a pressure sensor; a motion monitoring module for monitoring the motion trajectory of the user's hand using the multi-axis accelerometer to obtain a first motion data sequence and monitoring the posture changes of the user's hand using the gyroscope to obtain a second posture data sequence; and a gesture recognition module for determining whether the first motion data sequence and the second posture data sequence meet preset gesture triggering conditions, and when it is determined that the first motion data sequence and the second posture data sequence meet the gesture triggering conditions, using the pressure sensor to acquire a pressure distribution map of the user's fingers and performing feature extraction on the pressure distribution map to obtain a pressure feature set, and performing hand recognition on the pressure feature set according to a pre-built gesture feature library. The target gesture is obtained through momentum matching; the dynamic optimization module is used to generate a comprehensive motion feature vector by combining the first motion data sequence and the second posture data sequence when it is determined that the first motion data sequence and the second posture data sequence do not meet the gesture triggering conditions, and to dynamically adjust and optimize the comprehensive motion feature vector to obtain an optimized motion feature vector; the optimized motion feature vector is used to predict the gesture using a pre-built gesture classification model to obtain a candidate gesture; the function mapping module is used to return to the above steps of using a multi-axis accelerometer to monitor the user's hand motion trajectory to obtain the first motion data sequence when the target gesture or the candidate gesture is not detected, and to use a pre-built gesture function association table to perform function mapping on the target gesture or the candidate gesture to obtain a target function and execute the target function when the target gesture or the candidate gesture is detected.

9. An electronic device, characterized in that, The electronic device includes a memory and a processor, the memory being used to store at least one instruction, and the processor being used to execute the instruction stored in the memory to implement the sensor-based smart ring gesture recognition method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores at least one instruction, which is executed by a processor in an electronic device to implement the sensor-based smart ring gesture recognition method as described in any one of claims 1 to 7.

Images