Naked eye 3D interaction guiding method and system based on attitude sensing data

Abstract

The invention relates to the technical field of man-machine interaction, and discloses a naked eye 3D interaction guiding method and system based on posture perception data, and the method comprises the steps: obtaining a continuous posture sequence of a user in a naked eye 3D display space; recognizing an action mode according to the motion vector and acceleration change of the key joint points in the continuous attitude sequence; generating a predicted interaction track according to the spatial orientation and the change trend of the action mode; comparing the predicted interaction track with a preset interaction area in real time, and taking the interaction area as a target interaction area when a main path section of the predicted interaction track is within a triggering range of the interaction area; calling interaction logic bound with the target interaction area in the naked eye 3D display space to determine a target interaction event; generating visual guidance information synchronized with the naked eye 3D display space according to the target interaction event; according to the invention, the accuracy of naked eye 3D interaction guidance can be improved.

Description

Technical Field

[0001] This invention relates to the field of human-computer interaction technology, and in particular to a naked-eye 3D interactive guidance method and system based on posture perception data. Background Technology

[0002] In practical applications of naked-eye 3D interaction technology, existing technologies lack the ability to deeply mine spatiotemporal features in processing users' original posture perception data, making it difficult to effectively remove redundant information and abnormal interference from the data. This results in insufficient completeness and accuracy of the generated continuous posture sequences. Consequently, the subsequent recognition of user action patterns lacks reliable data support, making it impossible to accurately capture the motion vectors and acceleration change patterns of key joints. This leads to a deviation between the action pattern recognition results and the user's actual intentions, affecting the basic effectiveness of interactive guidance.

[0003] Existing technologies, in the process of generating predicted interaction trajectories and matching target interaction areas based on action patterns, lack a comprehensive consideration of the constraints of the naked-eye 3D display space boundary and the user's historical behavioral characteristics. This results in insufficient continuity and foresight of the predicted interaction trajectory, and low accuracy in comparison with the preset interaction area. At the same time, the generation of visual guidance information fails to fully adapt to the real-time state of the naked-eye 3D display space, and the temporal synchronization and spatial overlay rendering effect with the main scene are poor. This not only fails to provide users with accurate and timely interaction guidance, but may also reduce the smoothness of naked-eye 3D interaction and user experience due to the fuzziness or delay of the guidance information. Therefore, how to improve the smoothness of naked-eye 3D interaction has become an urgent problem to be solved. Summary of the Invention

[0004] This invention provides a naked-eye 3D interactive guidance method and system based on posture perception data to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides a naked-eye 3D interactive guidance method based on posture perception data, comprising: S1. Obtain the user's original posture perception data in the naked-eye 3D display space, and analyze the spatiotemporal characteristics of the original posture perception data to obtain the user's continuous posture sequence. S2. Identify the user's action pattern based on the motion vectors and acceleration changes of key joints in the continuous posture sequence across consecutive time frames; S3. Based on the spatial orientation and change trend of the action pattern in the naked-eye 3D display space, generate the user's predicted interaction trajectory; S4. Compare the spatial coordinates of the predicted interaction trajectory with the preset interaction area in real time. When the main path segment of the predicted interaction trajectory is within the trigger range of the interaction area, the interaction area is taken as the user's target interaction area. S5. Invoke the interaction logic bound to the target interaction area in the naked-eye 3D display space to determine the user's target interaction event; S6. Based on the target interaction event, generate visual guidance information synchronized with the naked-eye 3D display space, and overlay and render the visual guidance information onto the spatial position of the target interaction area to obtain the user's interaction guidance image.

[0006] In a preferred embodiment, the step of acquiring the user's original posture perception data within the naked-eye 3D display space and analyzing the spatiotemporal characteristics of the original posture perception data to obtain the user's continuous posture sequence includes: Collect the user's original pose point cloud data in the naked-eye 3D display space, align the timestamps of the original pose point cloud data, and obtain the user's synchronized pose point cloud data. Redundant frames are removed from the synchronized attitude point cloud data to obtain the user's standard attitude data; In the spatiotemporal dimension, the standard attitude data is segmented by a sliding window to obtain attitude data fragments containing spatial coordinate information. Extract the attitude key points from the attitude data segment, and obtain the user's motion trajectory segment based on the motion direction and displacement of the attitude key points; The motion trajectory segments are smoothly spliced ​​together in chronological order, and abnormal jump points in the spliced ​​motion trajectory are removed to obtain the user's continuous posture sequence.

[0007] In a preferred embodiment, identifying the user's action pattern based on the motion vector and acceleration changes of key joints in the continuous posture sequence across consecutive time frames includes: Extract the spatial displacement sequence of key joints in consecutive time frames from the continuous attitude sequence; In the time dimension, differential analysis is performed on the spatial displacement sequence to obtain the instantaneous motion vector sequence and acceleration change sequence of the key joint points; The instantaneous motion vector sequence and the acceleration change sequence are normalized to obtain the user's motion state sequence; In terms of time sequence, the motion state sequence is segmented to obtain the user's motion segments, and the motion feature vector direction, amplitude change law and acceleration trend in the motion segments are integrated to obtain the user's structured motion feature set; The motion feature vectors in the structured motion feature set are compared with the baseline motion vectors of the preset action pattern in multiple dimensions to obtain a similarity measure between the motion feature vectors and the baseline motion vectors. Based on the similarity metric, analyze the overall matching degree between the user's motion state and the preset action pattern; The preset motion mode with the highest overall matching degree is selected as the user's current motion mode.

[0008] In a preferred embodiment, the formula for calculating the overall matching degree is: ; in, This indicates the overall matching degree. This indicates the number of the motion feature vectors. Indicates the first The aforementioned motion feature vectors Indicates the first The aforementioned reference motion vectors This represents the preset spatial similarity weight. This represents the preset time-series weights, and , Indicates the first The actual timestamps of the motion feature vectors. Indicates the first The reference timestamp of each of the reference motion vectors This represents the preset time decay factor. This refers to the similarity metric. Represents the magnitude of a vector. This represents an exponential function.

[0009] In a preferred embodiment, generating the user's predicted interaction trajectory based on the spatial orientation and change trend of the action pattern within the naked-eye 3D display space includes: A three-dimensional reference coordinate system for the naked-eye 3D display space is established with the midpoint of the naked-eye 3D display space as the origin and the boundary of the naked-eye 3D display space as the constraint condition. The action pattern is mapped to the three-dimensional reference coordinate system to obtain the user's spatial pointing vector set; Based on the changes in the pointing angle sequence of the spatial pointing vector set, identify the current main spatial pointing and angle change trend of the action pattern; By combining the user's historical behavior parameters, the current main space orientation and the angle change trend are extrapolated along the time axis of the three-dimensional reference coordinate system to obtain the user's continuous action evolution sequence; The user's initial predicted trajectory is obtained by smoothly connecting the continuous evolution sequence of the actions in chronological order. Verify the continuity and motion intent of the initial predicted trajectory. Figure 1Consistency is achieved to obtain the optimized predicted trajectory of the user; Based on the historical interaction trajectory within the naked-eye 3D display space, the optimized prediction trajectory is corrected to obtain the user's predicted interaction trajectory.

[0010] In a preferred embodiment, the step of comparing the spatial coordinates of the predicted interaction trajectory with a preset interaction area in real time, and when the main path segment of the predicted interaction trajectory is within the trigger range of the interaction area, designating the interaction area as the user's target interaction area, includes: In the three-dimensional reference coordinate system of the naked-eye 3D display space, the regional center coordinates, spatial orientation, and half-width information along the coordinate axes of the interactive area are integrated to obtain the boundary space description of the interactive area. According to the time sequence, the predicted interaction trajectory is discretized into trajectory sampling points, and the spatial connections between the trajectory sampling points are summarized to obtain a set of continuous path segments of the predicted interaction trajectory. The minimum spatial distance between the continuous path segment set and the boundary space description is calculated. When the minimum spatial distance is less than the trigger range of the interaction area, it indicates that the path segment is within the trigger range. The ratio of the number of path segments within the trigger range to the total number of path segments in the continuous path segment set is used as the trajectory region overlap of the predicted interaction trajectory. When the overlap of the trajectory regions exceeds a preset ratio threshold, the interaction region is used as the user's candidate interaction region. The stability of the candidate interaction regions is verified to obtain the user's target interaction region.

[0011] In a preferred embodiment, the step of invoking the interaction logic bound to the target interaction area in the naked-eye 3D display space to determine the user's target interaction event includes: Analyze the motion convergence characteristics of the predicted interaction trajectory before it enters the target interaction area, and evaluate the intensity of intent persistence of the user's action pattern. Based on the motion convergence characteristics and the intensity of intent persistence, a dynamic adaptive condition for event triggering within the target interaction area is constructed, and the original static triggering condition within the target interaction area is replaced by the dynamic adaptive condition. Based on the dynamic adaptability conditions, the candidate event types bound to the target interaction area are analyzed, and event types that are mistakenly triggered due to the user's movement uncertainty and brief fluctuations in intent are eliminated to determine the user's target interaction event.

[0012] In a preferred embodiment, generating visual guidance information synchronized with the naked-eye 3D display space based on the target interaction event includes: The core operation type, associated object attributes, and expected feedback level in the target interaction event are analyzed to obtain the visual feature set of the target interaction event; The precise boundaries, surface orientation, and relative perspective relationship with the user of the target interaction area are obtained to obtain real-time spatial state data of the target interaction area. By combining the visual feature set and the real-time spatial state data, the user's basic primitives are obtained; Based on the relative perspective relationship, the basic primitives are geometrically transformed to obtain the user's initial visual information frame; Using the user's continuous posture sequence and current action mode as context, the animation easing curve and visual element intensity of the initial visual information frame are adjusted to obtain the user's optimized visual information frame. The optimized visual information frames are arranged in chronological order to obtain the user's initial visual guidance information. The initial visual guidance information is then chronologically aligned with the naked-eye 3D display space to obtain the user's visual guidance information.

[0013] In a preferred embodiment, the step of overlaying and rendering the visual guidance information onto the spatial location of the target interaction area to obtain the user's interaction guidance image includes: Based on the timestamp and spatial coordinate information of the visual guidance information, a virtual overlay layer corresponding to the current frame buffer in the naked-eye 3D display space is created. On the virtual overlay layer, based on the precise boundary and surface orientation of the target interaction area, the visual elements in the visual guidance information are mapped to the target interaction area to obtain the user's initial interactive virtual layer; Based on the relative position of the user's viewpoint and the initial interactive virtual layer, the occlusion, shadow, and transparency changes caused by the visual guidance information on the initial interactive virtual layer are simulated to correct the pixel rendering parameters of the initial interactive virtual layer and obtain the user's optimized interactive virtual layer. The optimized interactive virtual layer is temporally synchronized and synthesized with the main scene layer of the naked-eye 3D display space to obtain the user's interactive guidance image.

[0014] To address the above problems, the present invention also provides a naked-eye 3D interactive guidance system based on posture perception data, the system comprising: The posture data processing module is used to acquire the user's original posture perception data in the naked-eye 3D display space, and analyze the spatiotemporal characteristics of the original posture perception data to obtain the user's continuous posture sequence. The action pattern recognition module is used to identify the user's action pattern based on the motion vector and acceleration changes of key joints in the continuous posture sequence over consecutive time frames. The interaction trajectory prediction module is used to generate the user's predicted interaction trajectory based on the spatial orientation and change trend of the action pattern in the naked-eye 3D display space. The interaction area determination module is used to compare the spatial coordinates of the predicted interaction trajectory with the preset interaction area in real time. When the main path segment of the predicted interaction trajectory is within the trigger range of the interaction area, the interaction area is taken as the user's target interaction area. The interaction event determination module is used to call the interaction logic bound to the target interaction area in the naked-eye 3D display space to determine the user's target interaction event; The guidance image generation module is used to generate visual guidance information synchronized with the naked-eye 3D display space based on the target interaction event, and to overlay and render the visual guidance information onto the spatial position of the target interaction area to obtain the user's interaction guidance image.

[0015] Compared with the prior art, the present invention has the following beneficial effects: 1. This invention achieves accurate identification of user action patterns by precisely acquiring and deeply analyzing the user's original posture perception data, eliminating redundant information and abnormal transition points, and combining motion vectors, acceleration changes, and multi-dimensional similarity measurement calculations of key joints. Simultaneously, based on the construction and trend extrapolation of a three-dimensional reference coordinate system, the generated predicted interaction trajectory undergoes continuous verification and historical data correction, significantly improving the accuracy of target interaction area determination, effectively reducing the probability of false triggering of interaction events, and enhancing the reliability and accuracy of naked-eye 3D interaction.

[0016] 2. This invention generates visual guidance information that is synchronized with the spatial and temporal sequence of naked-eye 3D display by analyzing the core features of the target interactive event and the real-time spatial state of the interactive area. This information is then superimposed and rendered after geometric transformation, animation parameter adjustment, and simulation of realistic lighting and shadow effects. This ensures that the visual guidance not only matches the user's interactive intent but also deeply integrates with the 3D scene. This precise and synchronized guidance method allows users to clearly perceive the direction and goal of the interaction, lowers the operational threshold for naked-eye 3D interaction, significantly improves the efficiency and smoothness of interactive guidance, and brings users a more natural and seamless naked-eye 3D interactive experience. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating a naked-eye 3D interactive guidance method based on posture perception data, provided in an embodiment of the present invention. Figure 2 This is a functional block diagram of a naked-eye 3D interactive guidance system based on posture perception data, provided in an embodiment of the present invention. Figure 3 This is a demonstration diagram of user position tracking for a naked-eye 3D interactive guidance system based on posture perception data, provided in an embodiment of the present invention. Figure 4 This is a demonstration diagram of user motion tracking for a naked-eye 3D interactive guidance system based on posture perception data, provided in an embodiment of the present invention. Figure 5 This is a user-device interaction diagram of a naked-eye 3D interactive guidance system based on posture perception data provided in an embodiment of the present invention; The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0018] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0019] This application provides a glasses-free 3D interactive guidance method based on posture-aware data. The execution subject of this method includes, but is not limited to, at least one of the following electronic devices that can be configured to execute the method provided in this application: a server, a terminal, etc. In other words, the glasses-free 3D interactive guidance method based on posture-aware data can be executed by software or hardware installed on a terminal device or a server device. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster. The server can be an independent server or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0020] Reference Figure 1 The diagram shown is a flowchart illustrating a glasses-free 3D interactive guidance method based on posture-aware data according to an embodiment of the present invention. In this embodiment, the glasses-free 3D interactive guidance method based on posture-aware data includes: S1. Obtain the user's original posture perception data in the naked-eye 3D display space, and analyze the spatiotemporal characteristics of the original posture perception data to obtain the user's continuous posture sequence. In this embodiment of the invention, the step of acquiring the user's original posture perception data in the naked-eye 3D display space and analyzing the spatiotemporal characteristics of the original posture perception data to obtain the user's continuous posture sequence includes: Collect the user's original pose point cloud data in the naked-eye 3D display space, align the timestamps of the original pose point cloud data, and obtain the user's synchronized pose point cloud data. Redundant frames are removed from the synchronized attitude point cloud data to obtain the user's standard attitude data; In the spatiotemporal dimension, the standard attitude data is segmented by a sliding window to obtain attitude data fragments containing spatial coordinate information. Extract the attitude key points from the attitude data segment, and obtain the user's motion trajectory segment based on the motion direction and displacement of the attitude key points; The motion trajectory segments are smoothly spliced ​​together in chronological order, and abnormal jump points in the spliced ​​motion trajectory are removed to obtain the user's continuous posture sequence.

[0021] The demonstration method, which uses depth-sensing devices deployed within a glasses-free 3D display space to continuously capture the three-dimensional spatial position information of various parts of the user's body within that space, can be found here. Figure 3 A demonstration diagram of user position tracking for a naked-eye 3D interactive guidance system based on posture perception data; This information is presented in the form of point clouds to form raw attitude point cloud data. Each set of raw attitude point cloud data is accompanied by a corresponding time stamp. Then, the time stamp of each set of data is checked one by one, and all data are calibrated and adjusted according to a unified time reference to ensure that the point cloud data at different acquisition moments are accurately corresponding on the time axis, and finally, synchronized attitude point cloud data is obtained.

[0022] By comparing the spatial coordinates of all attitude points in two adjacent frames of the synchronized attitude point cloud data, if the spatial positions of all corresponding attitude points in two adjacent frames are completely consistent, or the positional deviation is within a very small range that would not occur under the natural posture of the human body, the frame is determined to be a redundant frame. Such redundant frames are directly removed from the synchronized attitude point cloud data, and the remaining point cloud data that can completely present the user's posture change process is the standard attitude data.

[0023] A fixed time interval is set as a sliding window. Starting from the start time of the standard attitude data, the window is moved gradually along the time axis at fixed time intervals. After each window moves, all standard attitude data contained within the window range are completely extracted. Since each set of standard attitude data carries spatial coordinate information, the part extracted each time is the attitude data segment containing spatial coordinate information. The window continues to move until it covers all standard attitude data, ensuring that all standard attitude data are divided into the corresponding attitude data segments.

[0024] The demonstration method, which uses depth-sensing devices deployed within a glasses-free 3D display space to continuously capture the posture and movement trajectories of various parts of the user's body within that space, can be found here. Figure 4 A demonstration diagram of user motion tracking for a naked-eye 3D interactive guidance system based on posture perception data; In each posture data segment, based on the characteristics of the human skeletal structure, the point cloud data corresponding to the preset posture key points such as the head, shoulders, elbows, wrists, hips, knees, and ankles are identified. The specific spatial coordinates of each posture key point at the start and end of the data segment are determined. The direction of movement of the posture key point is determined by the direction from the start coordinate to the end coordinate. The displacement of the posture key point is determined by calculating the straight path length between the start coordinate and the end coordinate. The movement directions and displacements of all posture key points are integrated to form a motion trajectory segment that can reflect the changes in the user's posture during the time period.

[0025] According to the chronological order of each motion trajectory segment, the beginning of the next motion trajectory segment is connected to the end of the previous motion trajectory segment. During the connection process, the motion trend of the posture key points in the adjacent trajectory segments is referenced to make the trajectory at the connection point naturally transition, so that the overall trajectory is smooth and continuous. Then, the complete motion trajectory after splicing is fully checked. If the coordinates of a certain posture key point at a certain moment deviate significantly from the coordinates of the same key point at the previous and next moments and do not conform to the normal human motion law, the point is determined to be an abnormal jump point and removed from the motion trajectory. After smooth splicing and removal of abnormal jump points, the trajectory that can continuously and accurately reflect the user's posture changes in the naked-eye 3D display space is the continuous posture sequence.

[0026] The beneficial effects are that, through the above specific implementation process, the user's posture data in the naked-eye 3D display space can be accurately collected and processed, and synchronous posture point cloud data, standard posture data, posture data fragments, and motion trajectory fragments can be obtained in sequence, ultimately forming a continuous posture sequence. Each processing step ensures the accuracy, integrity, and coherence of the data, providing reliable and high-quality data support for subsequent steps such as recognizing user action patterns and generating predicted interaction trajectories, effectively ensuring the accuracy and smoothness of naked-eye 3D interactive guidance.

[0027] S2. Identify the user's action pattern based on the motion vectors and acceleration changes of key joints in the continuous posture sequence across consecutive time frames; In this embodiment of the invention, identifying the user's action pattern based on the motion vector and acceleration changes of key joints in the continuous posture sequence across consecutive time frames includes: Extract the spatial displacement sequence of key joints in consecutive time frames from the continuous attitude sequence; In the time dimension, differential analysis is performed on the spatial displacement sequence to obtain the instantaneous motion vector sequence and acceleration change sequence of the key joint points; The instantaneous motion vector sequence and the acceleration change sequence are normalized to obtain the user's motion state sequence; In terms of time sequence, the motion state sequence is segmented to obtain the user's motion segments, and the motion feature vector direction, amplitude change law and acceleration trend in the motion segments are integrated to obtain the user's structured motion feature set; The motion feature vectors in the structured motion feature set are compared with the baseline motion vectors of the preset action pattern in multiple dimensions to obtain a similarity measure between the motion feature vectors and the baseline motion vectors. Based on the similarity metric, analyze the overall matching degree between the user's motion state and the preset action pattern; The preset motion mode with the highest overall matching degree is selected as the user's current motion mode.

[0028] The formula for calculating the overall matching degree is: ; in, This indicates the overall matching degree. This indicates the number of the motion feature vectors. Indicates the first The aforementioned motion feature vectors Indicates the first The aforementioned reference motion vectors This represents the preset spatial similarity weight. This represents the preset time-series weights, and , Indicates the first The actual timestamps of the motion feature vectors. Indicates the first The reference timestamp of each of the reference motion vectors This represents the preset time decay factor. This refers to the similarity metric. Represents the magnitude of a vector. This represents an exponential function.

[0029] By pre-determining the key joints that the human body needs to focus on in naked-eye 3D interactive scenarios, including the head, shoulders, elbows, wrists, hips, knees, and ankles, the three-dimensional spatial coordinates of these key joints are extracted one by one from the continuous posture sequence for each time frame. Then, for each key joint, the difference between its three-dimensional coordinates in the current time frame and the previous time frame is calculated. This difference represents the spatial displacement of the key joint between the two time frames. The spatial displacements of all key joints in each continuous time frame are arranged in chronological order to form the spatial displacement sequence corresponding to each key joint. Finally, the spatial displacement sequences of all key joints are integrated to obtain the user's spatial displacement sequence.

[0030] In the time dimension, using consecutive time frames as units, for the spatial displacement sequence of each key joint point, the spatial displacement of the current time frame and the corresponding time interval are taken. The motion direction is determined based on the direction of the spatial displacement, and the motion speed is determined based on the ratio of the length of the spatial displacement to the time interval. Combining the motion direction and the motion speed, the instantaneous motion vector of the key joint point in the current time frame is formed. The instantaneous motion vectors of all time frames are arranged in chronological order to obtain the instantaneous motion vector sequence of the key joint point. Then, for the instantaneous motion vector sequence of the key joint point, the instantaneous motion vector of the current time frame is calculated and compared with the previous time frame. The difference in instantaneous motion vectors within a time frame, including changes in direction and magnitude of velocity, is combined with the time interval between two time frames. The direction of acceleration is determined based on the direction of the difference, and the magnitude of acceleration is determined based on the ratio of the magnitude of the difference to the time interval. Combining the acceleration direction and magnitude, the acceleration change value of the key joint point in the current time frame is formed. The acceleration change values ​​of all time frames are arranged in chronological order to obtain the acceleration change sequence of the key joint point. Finally, the instantaneous motion vector sequences and acceleration change sequences of all key joint points are integrated to obtain the user's instantaneous motion vector sequence and acceleration change sequence, respectively.

[0031] First, for the instantaneous motion vector sequence, iterate through all instantaneous motion vectors, measure the length of each vector, and find the vector with the longest length. Divide the length of each instantaneous motion vector in the sequence by this maximum length, so that the length of all processed instantaneous motion vectors is within the range of 0 to 1, while keeping the direction of each vector unchanged. Next, for the acceleration change sequence, iterate through all acceleration change values, find the acceleration change value with the largest value, and divide the acceleration change value in the sequence by this maximum acceleration change value, so that all processed acceleration change values ​​are within the range of 0 to 1, while keeping the positive or negative attribute of each acceleration change value unchanged. Then, integrate the processed instantaneous motion vector sequence and the processed acceleration change sequence according to time frames. Each time frame corresponds to a set of processed instantaneous motion vectors and acceleration change values. Arrange all combinations corresponding to time frames in chronological order to obtain the user's motion state sequence.

[0032] In terms of temporal sequence, the motion state sequence is traversed at fixed time intervals to observe the changes in motion state within adjacent time frames. When the fluctuation amplitude of motion state within multiple consecutive time frames is less than a preset stability threshold, the motion state sequence corresponding to these consecutive time frames is divided into an independent motion segment. This process is repeated to traverse the entire motion state sequence until it is completely divided into multiple non-overlapping motion segments. For each motion segment, the direction of all instantaneous motion vectors within the segment is statistically analyzed, and the direction with the highest frequency is identified as the motion feature vector direction of that segment. The changes in the length of instantaneous motion vectors within the segment are observed to determine whether they gradually increase, gradually decrease, or remain stable, thus forming an amplitude change pattern. The trend of acceleration change values ​​within the segment is analyzed to determine whether they continuously increase, continuously decrease, or remain within a fixed range, thus forming an acceleration trend. The motion feature vector direction, amplitude change pattern, and acceleration trend of each motion segment are compiled. After summarizing all this information from all motion segments, the user's structured motion feature set is obtained.

[0033] A corresponding baseline motion vector sequence is pre-defined for each preset motion mode. Each baseline motion vector sequence contains the baseline motion vectors that the key joints of that motion mode should have in each time frame. The motion feature vector of each motion segment is extracted from the structured motion feature set and compared with the baseline motion vector of the corresponding time frame in the baseline motion vector sequence of the preset motion mode in multiple dimensions. First, the direction is compared to see if they are consistent. If the angle between the directions of the two vectors is less than a preset angle threshold, the direction dimension is determined to be consistent. Second, the length is compared to see if they are similar. If the ratio of the length difference between the two vectors to the length of the baseline motion vector is less than a preset ratio threshold, the length dimension is determined to be consistent. The consistency between each motion feature vector and the corresponding baseline motion vector in each dimension is statistically analyzed. If both dimensions are consistent, the similarity is recorded as high. If only one dimension is consistent, the similarity is recorded as medium. If neither dimension is consistent, the similarity is recorded as low. The similarity is converted into a specific value, which is the similarity measure between the motion feature vector and the baseline motion vector.

[0034] Based on the similarity metric corresponding to each motion feature vector, and considering the importance of that motion feature vector in the entire structured motion feature set, a corresponding weight is assigned to each similarity metric. Each similarity metric is multiplied by its corresponding weight to obtain a weighted similarity value for each motion feature vector. The weighted similarity values ​​of all motion feature vectors are summed to obtain a total value. This total value reflects the overall matching degree between the user's motion state and the preset action pattern. The overall matching degree between the user's motion state and each preset action pattern is calculated in this way.

[0035] The overall matching degree between the user's movement state and all preset action patterns is compared numerically. The preset action pattern with the highest overall matching degree is identified and determined as the user's current action pattern.

[0036] Motion feature vectors are obtained through feature extraction algorithms after data collection of target motion behavior. These algorithms preprocess the raw motion data to remove noise, extract key information characterizing the essential attributes of the motion, and convert it into vector form. Baseline motion vectors are a pre-constructed set of feature vectors corresponding to standard motion behaviors. The construction process involves collecting a large amount of raw data on standard motion behaviors, performing the same feature extraction process as for the target motion feature vectors to obtain multiple standard feature vectors, and then organizing these standard feature vectors into a baseline set that corresponds one-to-one with the motion feature vectors according to classification rules. The actual timestamp is the specific time information synchronously recorded by a time recording device when collecting the raw motion data corresponding to each motion feature vector. The baseline timestamp is used when constructing the baseline motion vector set for each baseline motion vector. The motion vector has pre-defined standard time information, which is consistent with the temporal pattern of standard motion behavior. Spatial similarity weight and temporal weight are set according to the importance of spatial matching accuracy and temporal matching accuracy in specific application scenarios. During the setting process, the ratio between the two is determined by analyzing the core requirements of the application scenario to ensure that the sum of the two is equal. The time decay factor is a value set according to the temporal sensitivity of motion behavior. During the setting process, the impact of different values ​​on the temporal matching results is tested to determine a reasonable value that can accurately reflect the impact of time differences on the overall matching degree. The similarity metric is a pre-defined indicator based on the attributes and matching requirements of the motion feature vector and the reference motion vector. During the definition process, the calculation logic of the similarity between the two vectors is clearly quantified to ensure that the indicator can effectively reflect the similarity relationship between the two.

[0037] First, calculate the dot product between each motion feature vector and its corresponding reference motion vector. The dot product is calculated by multiplying the corresponding elements of the two vectors and then summing all the products. Next, calculate the magnitude of each motion feature vector and the magnitude of its corresponding reference motion vector. The magnitude is calculated by summing the squares of all elements in the vector and then taking the square root of the sum. Divide the dot product by the product of the two magnitudes to obtain the similarity correlation value in the spatial dimension. Then, calculate the difference between the actual timestamp of each motion feature vector and the reference timestamp of its corresponding reference motion vector. Take the absolute value of this difference, divide it by the time decay factor, and then take the negative of the result. Then, using the natural constant as the base, an exponential operation is performed on the negative result to obtain the similarity correlation value in the temporal dimension. Next, the similarity correlation value in the spatial dimension is multiplied by the spatial similarity weight, and the similarity correlation value in the temporal dimension is multiplied by the temporal weight. The two product results are added together to obtain the comprehensive correlation value between a single motion feature vector and its corresponding baseline motion vector. Then, the comprehensive correlation value is multiplied by the similarity metric. Finally, the above product results corresponding to all motion feature vectors are summed to obtain the comprehensive matching degree. This comprehensive matching degree can comprehensively and quantitatively reflect the overall fit between the target motion behavior and the standard motion behavior in the spatial and temporal dimensions.

[0038] When the directions of a motion feature vector and its corresponding reference motion vector are closer, the dot product of the two is larger. With the magnitudes of the two vectors remaining constant, the spatial similarity correlation value is larger, and the result multiplied by the spatial similarity weight is larger, leading to a larger overall correlation value for a single motion feature vector, ultimately resulting in an upward trend in the overall matching degree. Conversely, when the absolute value of the difference between the actual timestamp of the motion feature vector and the reference timestamp of the corresponding reference motion vector is smaller, the temporal similarity correlation value after division by the time decay factor, negation, and exponentiation is larger, and the result multiplied by the temporal weight is larger, leading to a larger overall correlation value for a single motion feature vector, ultimately resulting in an upward trend in the overall matching degree. When the spatial similarity weight increases while the temporal weight decreases accordingly, the influence of the spatial similarity correlation value on the overall correlation value of a single motion feature vector strengthens, while the influence of the temporal similarity correlation value weakens, and the trend in the overall matching degree becomes more dependent on the spatial matching situation. Similarly, when the temporal weight increases while the spatial similarity weight decreases accordingly, the influence of the temporal similarity correlation value on the overall correlation value of a single motion feature vector strengthens, while the influence of the spatial similarity correlation value weakens. The overall matching degree is weaker, and its changing trend depends more on the matching situation in the time dimension. When the time decay factor decreases, the result of dividing the absolute value of the difference between the same timestamps by the decay factor increases, and the negative value is smaller. The similarity correlation value in the time dimension after exponential operation decreases more significantly, resulting in a decrease in the overall correlation value corresponding to a single motion feature vector. The overall matching degree shows a downward trend and its sensitivity to time differences increases. When the time decay factor increases, the result of dividing the absolute value of the difference between the same timestamps by the decay factor decreases, and the negative value is larger. The similarity correlation value in the time dimension after exponential operation decreases more significantly. The gradual decrease in the overall correlation value corresponding to a single motion feature vector reduces the sensitivity of the overall matching degree to time differences. When the similarity metric increases, the product of the overall correlation value corresponding to a single motion feature vector and the similarity metric increases, and the sum of all such products, i.e., the overall matching degree, shows an upward trend. When the degree of fit between the motion feature vector and the reference motion vector in the spatial direction decreases or the degree of match in the timestamp decreases, the similarity correlation value in the spatial or temporal dimensions will decrease, and the overall correlation value corresponding to a single motion feature vector will decrease accordingly, ultimately leading to a downward trend in the overall matching degree.

[0039] The beneficial effects are as follows: by starting from the above steps, the spatial displacement sequence of key joints can be gradually extracted from the continuous posture sequence. Then, through differential analysis, the instantaneous motion vector sequence and acceleration change sequence are obtained. Subsequently, through normalization processing, temporal segmentation, feature integration, multi-dimensional comparison, and comprehensive matching degree analysis, the user's current action pattern can be accurately identified. Each step has a clear operation method and a clear output, ensuring the accuracy and reliability of action pattern recognition and avoiding recognition errors caused by ambiguous operation or insufficient processing. This provides accurate action basis for subsequent naked-eye 3D interactive guidance links such as generating predicted interaction trajectories based on action patterns and determining target interaction areas, effectively improving the overall efficiency of naked-eye 3D interactive guidance and the user's interactive experience.

[0040] S3. Based on the spatial orientation and change trend of the action pattern in the naked-eye 3D display space, generate the user's predicted interaction trajectory; In this embodiment of the invention, generating the user's predicted interaction trajectory based on the spatial orientation and change trend of the action pattern within the naked-eye 3D display space includes: A three-dimensional reference coordinate system for the naked-eye 3D display space is established with the midpoint of the naked-eye 3D display space as the origin and the boundary of the naked-eye 3D display space as the constraint condition. The action pattern is mapped to the three-dimensional reference coordinate system to obtain the user's spatial pointing vector set; Based on the changes in the pointing angle sequence of the spatial pointing vector set, identify the current main spatial pointing and angle change trend of the action pattern; By combining the user's historical behavior parameters, the current main space orientation and the angle change trend are extrapolated along the time axis of the three-dimensional reference coordinate system to obtain the user's continuous action evolution sequence; The user's initial predicted trajectory is obtained by smoothly connecting the continuous evolution sequence of the actions in chronological order. Verify the continuity and motion intent of the initial predicted trajectory. Figure 1 Consistency is achieved to obtain the optimized predicted trajectory of the user; Based on the historical interaction trajectory within the naked-eye 3D display space, the optimized prediction trajectory is corrected to obtain the user's predicted interaction trajectory.

[0041] Using a ranging tool, the length, width, and height of the naked-eye 3D display space are precisely measured. By calculating half of the length, half of the width, and half of the height, the geometric center of the space is determined and set as the origin of the three-dimensional reference coordinate system. The directions of the three coordinate axes are defined as follows: the x-axis is parallel to the left and right direction of the front of the naked-eye 3D display space; the y-axis is parallel to the up and down direction of the front; and the z-axis is perpendicular to the front and back direction of the front. At the same time, the actual boundary of the naked-eye 3D display space is used as the coordinate range constraint. The x-axis coordinate range is from the x-coordinate of the origin minus half of the length to the x-coordinate of the origin plus half of the length. The y-axis and z-axis coordinate ranges are set according to the same logic. Finally, the three-dimensional reference coordinate system of the naked-eye 3D display space is established.

[0042] A thorough analysis of the identified motion patterns reveals the direction and complete path of movement for key user limbs such as arms and fingers. For example, in the "wave selection" motion pattern, the key limb is the arm, and the movement direction is a swing from the left to the right of the body. In the established three-dimensional reference coordinate system, the three-dimensional coordinates of the starting and ending points of each motion phase of the motion pattern are determined one by one. The vector corresponding to each motion phase is calculated by subtracting the starting point coordinate from the ending point coordinate. All the vectors corresponding to the motion phases are arranged in chronological order of the action to form a set of spatial pointing vectors that reflect changes in the user's motion direction.

[0043] The angle between each vector in the spatial pointing vector set and the x-axis, y-axis, and z-axis of the three-dimensional reference coordinate system is calculated. These angles are arranged sequentially according to the time order of the vectors in the set, forming a pointing angle sequence. The angle value with the highest frequency in this pointing angle sequence is counted, and the vector direction corresponding to this angle value is identified. This direction is determined as the current main spatial pointing of the action mode. The numerical difference between two adjacent angles in the pointing angle sequence is compared one by one. If the value of the latter angle is consistently greater than that of the former, the angle change trend is "increasing"; if the latter angle is consistently less than that of the former, it is "decreasing"; if the difference between adjacent angle values ​​is always within a very small range, it is "stable". This method accurately identifies the current main spatial pointing of the action mode and the angle change trend.

[0044] The system retrieves the user's historical behavior parameters, which include key data such as movement speed, duration of movement, and amplitude of limb movement when the user performed the same or similar action patterns in the past. Using the time axis of the three-dimensional reference coordinate system as a reference, the system determines the basic direction of the action based on the current main spatial direction and the dynamic adjustment law of the action direction by combining the angle change trend. For example, when the current main spatial direction is the positive z-axis and the angle change trend is "increasing", the action direction will gradually shift from the positive z-axis to the positive x-axis. Referring to the movement speed in the historical behavior parameters, the system calculates the spatial coordinates of key limbs and the corresponding pointing vectors at multiple consecutive time points in the future. These pointing vectors are arranged sequentially according to the future time order to obtain the continuous evolution sequence of the user's actions.

[0045] Following the chronological order on the timeline, the spatial coordinates of key limbs corresponding to each time point in the continuous evolution sequence of actions are extracted, and the spatial coordinates of adjacent time points are connected sequentially with straight lines. During the connection process, if the angle between two adjacent line segments is too large and exceeds the reasonable range of natural human movement, the coordinates of the intermediate transition node are adjusted to reduce the angle between the adjacent line segments and make it conform to the laws of human movement. For example, two line segments with an angle of 130 degrees can be adjusted to a 90-degree angle after adjusting the transition node. In this way, the coordinates of all time points are connected to form a smooth trajectory line, which yields the user's initial predicted trajectory.

[0046] When verifying the continuity of the initial predicted trajectory, the straight-line distance between two adjacent coordinate points on the trajectory is calculated. This distance is compared with the maximum distance that the same limb of the human body can move per unit time. If the distance between all adjacent points is less than or equal to the maximum distance, the trajectory is determined to be continuous. If there are adjacent points that exceed the maximum distance, the coordinates of those points are adjusted to meet the distance requirement. The movement intention is then verified. Figure 1 During alignment, the overall direction of the initial predicted trajectory is compared with the conventional intention direction of the current action mode. For example, the conventional intention direction of the "click interaction" action mode is pointing towards the interaction area within the naked-eye 3D display space. If the overall direction of the initial predicted trajectory deviates from this area, the trajectory direction is adjusted to conform to the conventional intention; thus achieving continuity and action intent. Figure 1 After consistency verification, the user's optimized predicted trajectory is obtained.

[0047] Collect all historical interaction trajectories generated when users perform the same action patterns in the naked-eye 3D display space. Analyze the common path ranges and key stop coordinates of these historical interaction trajectories in a 3D reference coordinate system to form historical trajectory reference data. Compare the optimized predicted trajectory with the common path ranges of the historical interaction trajectories segment by segment. If a segment of the optimized predicted trajectory exceeds the common range, adjust that segment to fall within the common path range. For example, if the common range of historical interaction trajectories in the y-axis direction is 0.3 meters to 1.0 meters, and the y-axis coordinate of a segment of the optimized predicted trajectory is 1.2 meters, adjust it to 1.0 meters. Simultaneously, refer to the key stop coordinates of the historical interaction trajectories. If the coordinates of the corresponding position in the optimized predicted trajectory deviate significantly from the key stop coordinates, fine-tune them to reduce the deviation to a reasonable range. Through the above correction based on historical data, the user's predicted interaction trajectory is obtained.

[0048] The beneficial effects are as follows: First, a unified and accurate three-dimensional reference coordinate system is established through the above steps to provide a reliable spatial benchmark for subsequent trajectory calculation and avoid spatial positioning confusion; then, the action pattern is converted into a set of spatial pointing vectors to realize the quantitative expression of the action direction; the core direction and dynamic trend of the action are accurately captured through pointing angle sequence analysis; trend extrapolation is performed by combining historical behavior parameters to make the trajectory prediction fit the user's movement habits; unreasonable fluctuations are eliminated through smooth connection and double verification to ensure that the trajectory is coherent and conforms to the action intention; finally, the prediction deviation is further reduced by correcting and optimizing based on historical interaction trajectories. The final generated predicted interaction trajectory is accurate, coherent and conforms to the user's actual interaction intention, providing a high-quality trajectory basis for naked-eye 3D interaction guidance and effectively improving the timeliness and accuracy of interaction guidance.

[0049] S4. Compare the spatial coordinates of the predicted interaction trajectory with the preset interaction area in real time. When the main path segment of the predicted interaction trajectory is within the trigger range of the interaction area, the interaction area is taken as the user's target interaction area. In this embodiment of the invention, the step of comparing the spatial coordinates of the predicted interaction trajectory with a preset interaction area in real time, and when the main path segment of the predicted interaction trajectory is within the trigger range of the interaction area, designating the interaction area as the user's target interaction area, includes: In the three-dimensional reference coordinate system of the naked-eye 3D display space, the regional center coordinates, spatial orientation, and half-width information along the coordinate axes of the interactive area are integrated to obtain the boundary space description of the interactive area. According to the time sequence, the predicted interaction trajectory is discretized into trajectory sampling points, and the spatial connections between the trajectory sampling points are summarized to obtain a set of continuous path segments of the predicted interaction trajectory. The minimum spatial distance between the continuous path segment set and the boundary space description is calculated. When the minimum spatial distance is less than the trigger range of the interaction area, it indicates that the path segment is within the trigger range. The ratio of the number of path segments within the trigger range to the total number of path segments in the continuous path segment set is used as the trajectory region overlap of the predicted interaction trajectory. When the overlap of the trajectory regions exceeds a preset ratio threshold, the interaction region is used as the user's candidate interaction region. The stability of the candidate interaction regions is verified to obtain the user's target interaction region.

[0050] In the three-dimensional reference coordinate system of the naked-eye 3D display space, the geometric center of the preset interactive area is accurately located using coordinate measuring tools, and the specific value of the center in the coordinate system is recorded as the coordinates of the area center. The placement angle of the interactive area in the display space is observed, and the spatial orientation is determined by judging the direction of its surface normal. For example, when the surface normal is parallel to the positive z-axis of the three-dimensional reference coordinate system, the spatial orientation is directly forward. The straight-line distance from the center of the interactive area to the edge along the x-axis, y-axis, and z-axis is measured using a distance measuring tool, and these three distances are recorded as the half-width information along each coordinate axis. The area center coordinates, spatial orientation description, and half-width values ​​along the coordinate axes are integrated together to clearly define the specific position, orientation, and range boundary of the interactive area in three-dimensional space, and finally form the boundary space description of the interactive area.

[0051] The predicted interactive trajectory is sampled at fixed time intervals in chronological order. Each time a sample is taken, the spatial coordinates of the trajectory at that time point in the three-dimensional reference coordinate system are accurately recorded. These sequentially collected spatial coordinates are arranged in an orderly manner to form trajectory sampling points. After all sampling is completed, adjacent trajectory sampling points are connected by straight lines in chronological order to form spatial lines. All spatial lines between adjacent trajectory sampling points are summarized to form a set containing multiple continuous line segments. This set is the continuous path segment set of the predicted interactive trajectory.

[0052] For each path segment in the continuous path segment set, multiple evenly distributed points are selected on the path segment. The straight-line distance from each selected point to the edge of the region defined by the boundary space description of the interaction area is calculated using a distance measurement method. Among all the calculated straight-line distances, the distance with the smallest value is selected and determined as the minimum spatial distance between the current path segment and the boundary space description. This minimum spatial distance is directly compared with the preset interaction area trigger range value. If the value of the minimum spatial distance is less than the trigger range value, it indicates that part or all of the path segment is within the trigger range of the interaction area.

[0053] First, count all path segments contained in the continuous path segment set to obtain the total number of path segments; then check each path segment one by one to determine whether it is within the trigger range of the interaction area, count the path segments within the trigger range separately to obtain the number of path segments within the trigger range; divide the number of path segments within the trigger range by the total number of path segments in the continuous path segment set, and the calculated ratio is the overlap of the trajectory area of ​​the predicted interaction trajectory.

[0054] Based on the actual application requirements of naked-eye 3D interaction scenarios, a fixed ratio threshold is preset. The overlap of the calculated predicted interaction trajectory area is compared with the preset ratio threshold. If the overlap value of the trajectory area exceeds the preset ratio threshold, it indicates that the interaction area is highly correlated with the user's predicted interaction trajectory. At this time, the interaction area is determined as the user's candidate interaction area.

[0055] When verifying the stability of the candidate interaction region, the overlap of the trajectory region corresponding to the candidate interaction region is continuously monitored in multiple consecutive time frames. If the overlap of the trajectory region remains above the preset ratio threshold in these consecutive time frames, and there are no significant fluctuations or sudden drops below the threshold, it indicates that the association between the candidate interaction region and the user's predicted interaction trajectory is stable. After the stability verification, the candidate interaction region is officially determined as the user's target interaction region.

[0056] The beneficial effects are as follows: First, by integrating multi-dimensional spatial information of the interactive area to form an accurate boundary space description, a clear and unified judgment standard is provided for the comparison of trajectory and area. Then, by sampling at fixed intervals and connecting lines, the predicted interactive trajectory is transformed into a set of continuous path segments that can be analyzed one by one, making the comparison operation more operable. By calculating the minimum spatial distance to determine whether the path segment is within the trigger range, the attribution of each trajectory segment is ensured to be accurate. By combining the overlap of trajectory areas to screen candidate interactive areas, the problem of incorrect area selection caused by misjudgment of a single path segment is effectively avoided. Finally, temporary associated areas are excluded after stability verification. The final determined target interactive area has both accuracy and stability, providing a reliable regional basis for subsequent invocation of interactive logic and generation of visual guidance information. This significantly improves the accuracy and reliability of naked-eye 3D interactive guidance and reduces the occurrence of interaction mis-triggers.

[0057] S5. Invoke the interaction logic bound to the target interaction area in the naked-eye 3D display space to determine the user's target interaction event; In this embodiment of the invention, the step of invoking the interaction logic bound to the target interaction area in the naked-eye 3D display space to determine the user's target interaction event includes: Analyze the motion convergence characteristics of the predicted interaction trajectory before it enters the target interaction area, and evaluate the intensity of intent persistence of the user's action pattern. Based on the motion convergence characteristics and the intensity of intent persistence, a dynamic adaptive condition for event triggering within the target interaction area is constructed, and the original static triggering condition within the target interaction area is replaced by the dynamic adaptive condition. Based on the dynamic adaptability conditions, the candidate event types bound to the target interaction area are analyzed, and event types that are mistakenly triggered due to the user's movement uncertainty and brief fluctuations in intent are eliminated to determine the user's target interaction event.

[0058] Observe the path changes of the predicted interaction trajectory over multiple consecutive time frames before it enters the target interaction area, and calculate the distance changes between each point on the trajectory and the center of the target interaction area. If the distance gradually decreases and the distance difference between adjacent time frames continues to decrease, it indicates that the motion trajectory is gradually converging towards the target area, and the motion convergence characteristic is judged to be strong. If the distance fluctuates greatly and the difference does not show a significant decreasing trend, the motion convergence characteristic is judged to be weak. At the same time, the duration of the action pattern before entering the target interaction area is statistically analyzed. If the core features of the action pattern, such as the direction and amplitude of movement of key joints, remain consistent over multiple consecutive time frames without deviation or abrupt change, it indicates that the user's intention to perform the action is stable, and the intensity of the intention persistence is evaluated as high. If the core features of the action pattern change frequently or other action features appear midway, the intensity of the intention persistence is evaluated as low.

[0059] By combining the motion convergence characteristics and intent persistence strength obtained from the analysis, dynamic adaptive conditions are constructed. If the motion convergence characteristics are strong and the intent persistence strength is high, the trigger condition is set as "the overlap between the predicted interaction trajectory and the target interaction area reaches 60%"; if the motion convergence characteristics are strong but the intent persistence strength is low, the trigger condition is set as "the overlap reaches 80% and lasts for 2 time frames"; if the motion convergence characteristics are weak but the intent persistence strength is high, the trigger condition is set as "the overlap reaches 70% and the trajectory fluctuation amplitude is less than a preset fixed range"; if the motion convergence characteristics are weak and the intent persistence strength is low, the trigger condition is set as "the overlap reaches 90% and lasts for 3 time frames". This set of conditions, which are dynamically adjusted according to the actual motion state and intent, completely replaces the original static trigger condition of the target interaction area, such as requiring an overlap of 80% in all cases.

[0060] The system retrieves all pre-bound candidate event types for the target interaction area, such as "click to open," "long press to select," and "swipe to switch." It then compares the triggering requirements for each candidate event type with the constructed dynamic adaptive conditions, analyzing whether the candidate event meets the current motion convergence characteristics and intent persistence strength. If a candidate event type is "click to open," its triggering logic requires strong motion convergence and high intent persistence, and the current dynamic adaptive conditions match these requirements, then the candidate event is retained. If a candidate event type is "swipe to switch," its triggering logic requires weak convergence characteristics, but the current dynamic conditions show strong motion convergence, and the candidate event may be falsely triggered due to motion uncertainties such as slight hand tremors, then it is eliminated. After filtering all candidate event types, the only event type that ultimately meets the dynamic adaptive conditions is the user's target interaction event.

[0061] The beneficial effects are as follows: by first analyzing the motion convergence characteristics and evaluating the intensity of intent persistence, the event triggering conditions are provided with a basis that fits the user's actual operation state. Then, dynamic adaptive conditions are constructed to replace static conditions, making the triggering judgment more flexible and targeted. Finally, candidate event types are filtered based on dynamic conditions, effectively eliminating false triggering events caused by motion uncertainty and intent fluctuations. This greatly improves the accuracy and reliability of target interaction event determination, making naked-eye 3D interaction more in line with the user's real intent, and improving the smoothness of interaction and user experience.

[0062] S6. Based on the target interaction event, generate visual guidance information synchronized with the naked-eye 3D display space, and overlay and render the visual guidance information onto the spatial position of the target interaction area to obtain the user's interaction guidance image.

[0063] In this embodiment of the invention, generating visual guidance information synchronized with the naked-eye 3D display space based on the target interaction event includes: The core operation type, associated object attributes, and expected feedback level in the target interaction event are analyzed to obtain the visual feature set of the target interaction event; The precise boundaries, surface orientation, and relative perspective relationship with the user of the target interaction area are obtained to obtain real-time spatial state data of the target interaction area. By combining the visual feature set and the real-time spatial state data, the user's basic primitives are obtained; Based on the relative perspective relationship, the basic primitives are geometrically transformed to obtain the user's initial visual information frame; Using the user's continuous posture sequence and current action mode as context, the animation easing curve and visual element intensity of the initial visual information frame are adjusted to obtain the user's optimized visual information frame. The optimized visual information frames are arranged in chronological order to obtain the user's initial visual guidance information. The initial visual guidance information is then chronologically aligned with the naked-eye 3D display space to obtain the user's visual guidance information.

[0064] The step of overlaying and rendering the visual guidance information onto the spatial location of the target interaction area to obtain the user's interaction guidance image includes: Based on the timestamp and spatial coordinate information of the visual guidance information, a virtual overlay layer corresponding to the current frame buffer in the naked-eye 3D display space is created. On the virtual overlay layer, based on the precise boundary and surface orientation of the target interaction area, the visual elements in the visual guidance information are mapped to the target interaction area to obtain the user's initial interactive virtual layer; Based on the relative position of the user's viewpoint and the initial interactive virtual layer, the occlusion, shadow, and transparency changes caused by the visual guidance information on the initial interactive virtual layer are simulated to correct the pixel rendering parameters of the initial interactive virtual layer and obtain the user's optimized interactive virtual layer. The optimized interactive virtual layer is temporally synchronized and synthesized with the main scene layer of the naked-eye 3D display space to obtain the user's interactive guidance image.

[0065] To make the user-device interaction process more observable, see the detailed interaction process below. Figure 5 User-device interaction diagram of a naked-eye 3D interactive guidance system based on posture perception data; When analyzing a target interaction event, examine the preset interaction logic corresponding to the event to determine whether the user needs to perform a click, swipe, or select, thus identifying the core operation type. Retrieve the preset attribute data of the naked-eye 3D display elements involved in the interaction event to determine their color, shape, and size, and obtain the attributes of related objects. Classify the expected feedback level according to the importance of the interaction event: events triggered by important functions are strong feedback, regular functions are medium feedback, and auxiliary prompts are weak feedback. Strong feedback corresponds to a highlighting and flashing effect, medium feedback corresponds to a normal glowing effect, and weak feedback corresponds to a slight color change. Integrate the visual performance corresponding to the core operation type, the attribute parameters of related objects, and the effect parameters corresponding to the expected feedback level to form the visual feature set of the target interaction event.

[0066] When obtaining the precise boundary of the target interactive area, multiple feature points on the edge of the area are selected in the three-dimensional reference coordinate system of the naked-eye 3D display space. The three-dimensional coordinates of each feature point are recorded one by one. By connecting these coordinate points, a closed contour is formed, thereby clarifying the precise boundary. When determining the surface orientation, the normal direction of the surface of the area is calculated. The direction pointed to by the normal direction is the surface orientation of the target interactive area. When obtaining the relative perspective relationship with the user, the user's current viewpoint coordinates are first recorded. Then, the direction of the line connecting the viewpoint coordinates and the center coordinates of the target interactive area is calculated. Combined with the surface orientation of the area, the specific angle at which the user observes the area is determined. The coordinate set of the precise boundary, the surface orientation parameters, and the angle information of the relative perspective relationship are integrated to obtain the real-time spatial state data of the target interactive area.

[0067] When combining visual feature sets and real-time spatial state data, the precise boundaries in the real-time spatial state data are used as the basic framework. The framework is filled with the corresponding color based on the color parameters of the associated objects in the visual feature set. According to the core operation type in the visual feature set, corresponding guiding elements are added within the framework; for example, a circular prompt is added to the center of the framework when the core operation type is click. A basic display mode is set for the framework based on the expected feedback level in the visual feature set; for example, the framework is preset to a blinking mode under a strong feedback level. Through the above methods, the key information of the two types of data is deeply integrated to generate graphic elements that possess both spatial location attributes and visual performance attributes. These graphic elements are the user's basic primitives.

[0068] When performing geometric transformations on basic primitives based on relative perspective, if the relative perspective is a side view, such as when the user observes the target interaction area from the left, the rectangular frame in the basic primitive is adjusted to a trapezoid that is narrower on the left and wider on the right to ensure consistency with the user's actual viewing angle. If the relative perspective is a top view, the circular tooltip in the basic primitive is adjusted to an ellipse, with the major axis of the ellipse matching the user's top-down view. By adjusting the shape, proportion, and spatial position of the basic primitives to fully match the visual effect observed by the user, the single-frame graphic obtained after this geometric transformation becomes the user's initial visual information frame.

[0069] Using the user's continuous posture sequence as a reference, the rate of change of adjacent postures in the sequence is analyzed. If the posture changes rapidly, it indicates that the user's movements are abrupt. In this case, the animation easing curve of the initial visual information frame is adjusted to a steep shape to make the appearance and changes of visual elements more rapid. If the posture changes slowly, it indicates that the user's movements are gentle. In this case, the animation easing curve is adjusted to a gentle shape to make the changes of visual elements smoother. Combined with the current action mode, if the action mode is "rapid click", the intensity parameters such as brightness and contrast of visual elements in the initial visual information frame are increased. If the action mode is "slow swipe", the intensity of visual elements is appropriately reduced to ensure that the prompts are clear and not dazzling. By adjusting the shape of the animation easing curve and the intensity parameters of visual elements, the optimized visual information frame for the user is obtained.

[0070] All optimized visual information frames are arranged sequentially according to time sequence to ensure a smooth transition between frames. For example, if the brightness of the prompt point in the previous frame is 50%, it will naturally increase to 70% in the next frame according to the easing curve. After the arrangement is completed, the user's initial visual guidance information is formed. The timestamp of each frame in the naked-eye 3D display space is extracted, and the timestamps of each optimized visual information frame in the initial visual guidance information are precisely matched with the corresponding timestamps in the naked-eye 3D display space to ensure that the display rhythm of the visual guidance information is completely synchronized with the update rhythm of the naked-eye 3D display space. After time alignment, the user's visual guidance information is obtained.

[0071] Based on the timestamp of each segment of guidance content in the visual guidance information, locate the current frame buffer in the naked-eye 3D display space corresponding to that timestamp, i.e., the area storing the display data at that time point; based on the spatial coordinate information of the visual guidance information, determine the specific location in the frame buffer where the guidance content needs to be added; based on the frame buffer specifications and spatial coordinates corresponding to that timestamp, create an independent virtual overlay layer. The resolution and frame rate of this layer are completely consistent with the current frame buffer, and the spatial position of the layer corresponds precisely to the spatial coordinates of the guidance content, thus obtaining the virtual overlay layer corresponding to the current frame buffer in the naked-eye 3D display space.

[0072] On the virtual overlay layer, referencing the precise boundary data of the target interaction area, determine the placement range of each visual element in the visual guidance information. For example, when the visual element is a prompt arrow, place the arrow at the outer edge of the precise boundary to avoid exceeding the area range. Adjust the orientation of the visual elements according to the surface orientation of the target interaction area so that the front of the visual element is consistent with the surface orientation of the area. For example, if the surface orientation of the area is diagonally upward, then the prompt arrow is also adjusted to face diagonally upward. By precisely placing and adjusting the orientation of the visual elements according to the boundary position and surface orientation, the mapping of the visual elements to the target interaction area is completed, and the initial interactive virtual layer for the user is obtained after the mapping is completed.

[0073] Obtain the user's viewpoint coordinates, calculate the spatial distance and relative angle between these coordinates and each visual element in the initial interactive virtual layer, and determine the user's specific viewing angle for observing the initial interactive virtual layer. If there are other scene elements in the naked-eye 3D display space in front of the initial interactive virtual layer, determine the occlusion area of ​​the visual elements in the initial interactive virtual layer based on the spatial coordinates and outlines of these scene elements. Within the occlusion area, adjust the transparency of the visual elements to 80%, while maintaining 100% transparency in the unoccluded areas. Based on the preset light source direction of the naked-eye 3D display space, add a shadow with a grayscale value of 60% to the backlight side of the visual elements in the initial interactive virtual layer, with the length and direction of the shadow matching the light source direction. Based on the occlusion, shadow, and transparency adjustment requirements, correct the color values, transparency parameters, and grayscale parameters of the corresponding pixels in the initial interactive virtual layer. After correction, the optimized interactive virtual layer for the user is obtained.

[0074] The timestamps and spatial location information of each frame in the optimized interactive virtual layer are extracted, and the corresponding timestamp image data in the main scene layer of the naked-eye 3D display space is also extracted to ensure that the two are completely synchronized in the time dimension. The images of the optimized interactive virtual layer are superimposed onto the corresponding positions in the main scene layer according to their spatial location information. During the superposition process, it is ensured that the visual elements of the optimized interactive virtual layer do not obscure the core content of the target interactive area in the main scene layer, and that the display effect of the visual elements is coordinated and unified with the 3D effect of the main scene. By accurately compositing the two layers of images in time and space, a user interaction guidance image containing the naked-eye 3D main scene and clear visual guidance information is obtained.

[0075] The beneficial effects are as follows: by analyzing the key information of the target interaction event and the real-time spatial state of the target area step by step, the generated visual guidance information is ensured to not only meet the user's interaction needs but also accurately match the spatial observation perspective; after geometric transformation and adjustment of animation and intensity parameters, the guidance information is more adapted to the user's action rhythm and observation habits; during the overlay rendering process, by accurately creating virtual overlay layers, mapping visual elements, and simulating real light and shadow effects, the clear visibility of the guidance information is ensured without destroying the integrity of the naked-eye 3D main scene. The final generated interactive guidance image can provide users with accurate, synchronous, and natural visual guidance, helping users quickly locate the target interaction area and smoothly complete the operation, effectively reducing the operation threshold of naked-eye 3D interaction and improving the smoothness, accuracy, and user experience of the interaction.

[0076] like Figure 2 The diagram shown is a functional block diagram of a naked-eye 3D interactive guidance system based on posture perception data provided in an embodiment of the present invention.

[0077] The naked-eye 3D interactive guidance system 100 based on posture perception data described in this invention can be installed in an electronic device. Depending on the functions implemented, the naked-eye 3D interactive guidance system 100 may include a posture data processing module 101, a motion pattern recognition module 102, an interaction trajectory prediction module 103, an interaction area determination module 104, an interaction event determination module 105, and a guidance image generation module 106. The modules described in this invention can also be referred to as units, which are a series of computer program segments that can be executed by the processor of an electronic device and can perform a fixed function, and are stored in the memory of the electronic device.

[0078] In this embodiment, the functions of each module / unit are as follows: The posture data processing module 101 is used to acquire the user's original posture perception data in the naked-eye 3D display space, and analyze the spatiotemporal characteristics of the original posture perception data to obtain the user's continuous posture sequence. The action pattern recognition module 102 is used to identify the user's action pattern based on the motion vector and acceleration changes of key joints in the continuous posture sequence in continuous time frames. The interaction trajectory prediction module 103 is used to generate the user's predicted interaction trajectory based on the spatial orientation and change trend of the action mode in the naked-eye 3D display space. The interaction area determination module 104 is used to compare the spatial coordinates of the predicted interaction trajectory with the preset interaction area in real time. When the main path segment of the predicted interaction trajectory is within the trigger range of the interaction area, the interaction area is taken as the user's target interaction area. The interaction event determination module 105 is used to call the interaction logic bound to the target interaction area in the naked-eye 3D display space to determine the user's target interaction event; The guidance image generation module 106 is used to generate visual guidance information synchronized with the naked-eye 3D display space according to the target interaction event, and to overlay and render the visual guidance information onto the spatial position of the target interaction area to obtain the user's interaction guidance image.

[0079] In the several embodiments provided by this invention, it should be understood that the disclosed methods and systems can be implemented in other ways. For example, the system embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and other division methods may be used in actual implementation.

[0080] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0081] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0082] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0083] This application embodiment can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0084] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A naked-eye 3D interactive guidance method based on posture perception data, characterized in that, The method includes: S1. Obtain the user's original posture perception data in the naked-eye 3D display space, and analyze the spatiotemporal characteristics of the original posture perception data to obtain the user's continuous posture sequence. S2. Identify the user's action pattern based on the motion vectors and acceleration changes of key joints in the continuous posture sequence across consecutive time frames; S3. Based on the spatial orientation and change trend of the action pattern in the naked-eye 3D display space, generate the user's predicted interaction trajectory; S4. The spatial coordinates of the predicted interaction trajectory are compared with the preset interaction area in real time. When the main path segment of the predicted interaction trajectory is within the trigger range of the interaction area, the interaction area is taken as the target interaction area of ​​the user. S5. Invoke the interaction logic bound to the target interaction area in the naked-eye 3D display space to determine the user's target interaction event; S6. Based on the target interaction event, generate visual guidance information synchronized with the naked-eye 3D display space, and overlay and render the visual guidance information onto the spatial position of the target interaction area to obtain the user's interaction guidance image.

2. The naked-eye 3D interactive guidance method based on posture perception data as described in claim 1, characterized in that, The process of acquiring the user's original posture perception data within the naked-eye 3D display space and analyzing the spatiotemporal characteristics of the original posture perception data to obtain the user's continuous posture sequence includes: Collect the user's original pose point cloud data in the naked-eye 3D display space, align the timestamps of the original pose point cloud data, and obtain the user's synchronized pose point cloud data. Redundant frames are removed from the synchronized attitude point cloud data to obtain the user's standard attitude data; In the spatiotemporal dimension, the standard attitude data is segmented by a sliding window to obtain attitude data fragments containing spatial coordinate information. Extract the attitude key points from the attitude data segment, and obtain the user's motion trajectory segment based on the motion direction and displacement of the attitude key points; The motion trajectory segments are smoothly spliced ​​together in chronological order, and abnormal jump points in the spliced ​​motion trajectory are removed to obtain the user's continuous posture sequence.

3. The naked-eye 3D interactive guidance method based on posture perception data as described in claim 1, characterized in that, The step of identifying the user's action pattern based on the motion vectors and acceleration changes of key joints in the continuous posture sequence across consecutive time frames includes: Extract the spatial displacement sequence of key joints in consecutive time frames from the continuous attitude sequence; In the time dimension, differential analysis is performed on the spatial displacement sequence to obtain the instantaneous motion vector sequence and acceleration change sequence of the key joint points; The instantaneous motion vector sequence and the acceleration change sequence are normalized to obtain the user's motion state sequence; In terms of time sequence, the motion state sequence is segmented to obtain the user's motion segments, and the motion feature vector direction, amplitude change law and acceleration trend in the motion segments are integrated to obtain the user's structured motion feature set; The motion feature vectors in the structured motion feature set are compared with the baseline motion vectors of the preset action pattern in multiple dimensions to obtain a similarity measure between the motion feature vectors and the baseline motion vectors. Based on the similarity metric, analyze the overall matching degree between the user's motion state and the preset action pattern; The preset motion mode with the highest overall matching degree is selected as the user's current motion mode.

4. The naked-eye 3D interactive guidance method based on posture perception data as described in claim 3, characterized in that, The formula for calculating the overall matching degree is: ； in, This indicates the overall matching degree. This indicates the number of the motion feature vectors. Indicates the first The aforementioned motion feature vectors Indicates the first The aforementioned reference motion vectors This represents the preset spatial similarity weight. This represents the preset time-series weights, and , Indicates the first The actual timestamps of the motion feature vectors. Indicates the first The reference timestamp of each of the reference motion vectors This represents the preset time decay factor. This refers to the similarity metric. Represents the magnitude of a vector. This represents an exponential function.

5. The naked-eye 3D interactive guidance method based on posture perception data as described in claim 1, characterized in that, The step of generating the user's predicted interaction trajectory based on the spatial orientation and change trend of the action pattern within the naked-eye 3D display space includes: A three-dimensional reference coordinate system for the naked-eye 3D display space is established with the midpoint of the naked-eye 3D display space as the origin and the boundary of the naked-eye 3D display space as the constraint condition. The action pattern is mapped to the three-dimensional reference coordinate system to obtain the user's spatial pointing vector set; Based on the changes in the pointing angle sequence of the spatial pointing vector set, identify the current main spatial pointing and angle change trend of the action pattern; By combining the user's historical behavior parameters, the current main space orientation and the angle change trend are extrapolated along the time axis of the three-dimensional reference coordinate system to obtain the user's continuous action evolution sequence; The user's initial predicted trajectory is obtained by smoothly connecting the continuous evolution sequence of the actions in chronological order. Verify the continuity and consistency of the initial predicted trajectory with the user's action intent to obtain the user's optimized predicted trajectory; Based on the historical interaction trajectory within the naked-eye 3D display space, the optimized prediction trajectory is corrected to obtain the user's predicted interaction trajectory.

6. The naked-eye 3D interactive guidance method based on posture perception data as described in claim 1, characterized in that, The step of comparing the spatial coordinates of the predicted interaction trajectory with a preset interaction area in real time, and when the main path segment of the predicted interaction trajectory is within the trigger range of the interaction area, designating the interaction area as the user's target interaction area, includes: In the three-dimensional reference coordinate system of the naked-eye 3D display space, the regional center coordinates, spatial orientation, and half-width information along the coordinate axes of the interactive area are integrated to obtain the boundary space description of the interactive area. According to the time sequence, the predicted interaction trajectory is discretized into trajectory sampling points, and the spatial connections between the trajectory sampling points are summarized to obtain a set of continuous path segments of the predicted interaction trajectory. The minimum spatial distance between the continuous path segment set and the boundary space description is calculated. When the minimum spatial distance is less than the trigger range of the interaction area, it indicates that the path segment is within the trigger range. The ratio of the number of path segments within the trigger range to the total number of path segments in the continuous path segment set is used as the trajectory region overlap of the predicted interaction trajectory. When the overlap of the trajectory regions exceeds a preset ratio threshold, the interaction region is used as the user's candidate interaction region. The stability of the candidate interaction regions is verified to obtain the user's target interaction region.

7. The naked-eye 3D interactive guidance method based on posture perception data as described in claim 1, characterized in that, The step of invoking the interaction logic bound to the target interaction area in the naked-eye 3D display space to determine the user's target interaction event includes: Analyze the motion convergence characteristics of the predicted interaction trajectory before it enters the target interaction area, and evaluate the intensity of intent persistence of the user's action pattern. Based on the motion convergence characteristics and the intensity of intent persistence, a dynamic adaptive condition for event triggering within the target interaction area is constructed, and the original static triggering condition within the target interaction area is replaced by the dynamic adaptive condition. Based on the dynamic adaptability conditions, the candidate event types bound to the target interaction area are analyzed, and event types that are mistakenly triggered due to the user's movement uncertainty and brief fluctuations in intent are eliminated to determine the user's target interaction event.

8. The naked-eye 3D interactive guidance method based on posture perception data as described in claim 1, characterized in that, The step of generating visual guidance information synchronized with the naked-eye 3D display space based on the target interaction event includes: The core operation type, associated object attributes, and expected feedback level in the target interaction event are analyzed to obtain the visual feature set of the target interaction event; The precise boundaries, surface orientation, and relative perspective relationship with the user of the target interaction area are obtained to obtain real-time spatial state data of the target interaction area. By combining the visual feature set and the real-time spatial state data, the user's basic primitives are obtained; Based on the relative perspective relationship, the basic primitives are geometrically transformed to obtain the user's initial visual information frame; Using the user's continuous posture sequence and current action mode as context, the animation easing curve and visual element intensity of the initial visual information frame are adjusted to obtain the user's optimized visual information frame. The optimized visual information frames are arranged in chronological order to obtain the user's initial visual guidance information. The initial visual guidance information is then chronologically aligned with the naked-eye 3D display space to obtain the user's visual guidance information.

9. The naked-eye 3D interactive guidance method based on posture perception data as described in claim 1, characterized in that, The step of overlaying and rendering the visual guidance information onto the spatial location of the target interaction area to obtain the user's interaction guidance image includes: Based on the timestamp and spatial coordinate information of the visual guidance information, a virtual overlay layer corresponding to the current frame buffer in the naked-eye 3D display space is created. On the virtual overlay layer, based on the precise boundary and surface orientation of the target interaction area, the visual elements in the visual guidance information are mapped to the target interaction area to obtain the user's initial interactive virtual layer; Based on the relative position of the user's viewpoint and the initial interactive virtual layer, the occlusion, shadow, and transparency changes caused by the visual guidance information on the initial interactive virtual layer are simulated to correct the pixel rendering parameters of the initial interactive virtual layer and obtain the user's optimized interactive virtual layer. The optimized interactive virtual layer is temporally synchronized and synthesized with the main scene layer of the naked-eye 3D display space to obtain the user's interactive guidance image.

10. A naked-eye 3D interactive guidance system based on posture perception data, characterized in that, The system for implementing the glasses-free 3D interactive guidance method based on posture perception data as described in claim 1 includes: The posture data processing module is used to acquire the user's original posture perception data in the naked-eye 3D display space, and analyze the spatiotemporal characteristics of the original posture perception data to obtain the user's continuous posture sequence. The action pattern recognition module is used to identify the user's action pattern based on the motion vector and acceleration changes of key joints in the continuous posture sequence over consecutive time frames. The interaction trajectory prediction module is used to generate the user's predicted interaction trajectory based on the spatial orientation and change trend of the action pattern in the naked-eye 3D display space. The interaction area determination module is used to compare the spatial coordinates of the predicted interaction trajectory with the preset interaction area in real time. When the main path segment of the predicted interaction trajectory is within the trigger range of the interaction area, the interaction area is taken as the user's target interaction area. The interaction event determination module is used to call the interaction logic bound to the target interaction area in the naked-eye 3D display space to determine the user's target interaction event; The guidance image generation module is used to generate visual guidance information synchronized with the naked-eye 3D display space based on the target interaction event, and to overlay and render the visual guidance information onto the spatial position of the target interaction area to obtain the user's interaction guidance image.

Images