Wearable device-based sensor data labeling method, tool, and storage medium

Abstract

The application discloses a wearable device-based sensor data labeling method, tool and storage medium, and belongs to the technical field of sensor data labeling. The method comprises the following steps: acquiring to-be-labeled file data, pre-processing the received to-be-labeled file data, and marking the pre-processed to-be-labeled file data as target data; setting a corresponding display mode according to user requirements, and visually displaying the read data according to the display mode; establishing a labeling mode library and a feature library, acquiring user labeling requirements, and matching corresponding labeling modes and labeling models from the labeling mode library according to the labeling requirements; labeling the target data according to the matched labeling modes and labeling models; obtaining target labeling data; verifying the target labeling data, and adjusting the target labeling data according to the verification result; exporting the verified target labeling data; and continuously optimizing, analyzing and adjusting the labeling mode library.

Description

Technical Field

[0001] This invention belongs to the field of sensor data annotation technology, specifically a sensor data annotation method, tool, and storage medium based on wearable devices. Background Technology

[0002] With the development of sensor technology, more and more industries rely on sensor data for their work. Sensor data is usually in digital form and needs to be labeled before it can be further analyzed and applied. However, for the processing and analysis of large amounts of sensor data, especially data labeling, traditional manual labeling methods suffer from low efficiency, high cost, and insufficient accuracy, and are prone to subjective errors, leading to inaccurate labeling results. Therefore, a more intelligent and efficient data labeling tool and method are needed.

[0003] Based on this, the present invention provides a sensor data annotation method, tool and storage medium based on wearable devices. By combining artificial intelligence technology and visualization, sensor data is automatically annotated and combined with manual verification, providing training data for model building efficiently and accurately. Summary of the Invention

[0004] To address the problems of the above solutions, this invention provides a sensor data annotation method, tool, and storage medium based on wearable devices.

[0005] The objective of this invention can be achieved through the following technical solutions:

[0006] A method for labeling sensor data based on wearable devices, including:

[0007] Step 1: Obtain the data of the file to be labeled, preprocess the received data of the file to be labeled, and mark the preprocessed data of the file to be labeled as the target data;

[0008] Step 2: Set the appropriate display method according to user needs, and then visualize the read data according to the display method;

[0009] Step 3: Establish a labeling pattern library and a feature library. The labeling pattern library is used to store various labeling patterns and the labeling models corresponding to each labeling pattern; the feature library is used to store various labeling features; obtain the user's labeling requirements, and match the corresponding labeling patterns and labeling models from the labeling pattern library according to the labeling requirements.

[0010] Furthermore, methods for matching corresponding annotation patterns and annotation models from the annotation pattern library according to annotation requirements include:

[0011] Analyze the annotation requirements and determine the requirements characteristics; generate a test set based on the requirements characteristics, and set up standard cases based on the test set;

[0012] Input the requirement features into the annotation pattern library for matching to obtain each candidate annotation method that meets the requirement features;

[0013] Conduct simulation tests on each candidate annotation method through a test set to obtain the annotation accuracy rate and annotation consistency rate corresponding to each candidate annotation method; and set the annotation cases corresponding to each candidate annotation method;

[0014] Evaluate each annotation case according to the standard case to obtain the case evaluation value corresponding to each candidate annotation method;

[0015] Calculate the corresponding annotation value according to the formula PY = b1×PA(s)×(b2×QA + B3×QB);

[0016] In the formula: PY is the annotation value; b1, b2, and b3 are all proportionality coefficients, and the value range is 0 < b1 ≤ 1, 0 < b2 ≤ 1, 0 < b3 ≤ 1; PA(s) is the case evaluation value output by the case evaluation model; QA is the annotation accuracy rate; QB is the annotation consistency rate;

[0017] Sort each candidate annotation method in descending order according to the annotation value to obtain the first sequence, and insert the corresponding annotation cases at each candidate annotation method in the first sequence;

[0018] Determine the target annotation method according to the first sequence, and identify the annotation pattern and annotation model corresponding to the target annotation method.

[0019] Furthermore, the method for evaluating each annotation case according to the standard case includes:

[0020] Establish a case evaluation model, and the expression of the case evaluation model is:

[0021] ;

[0022] In the formula: s is the input data, and the input data includes the standard case and the annotation case; the output data is the case evaluation value; BA is the optimization value, and the value range of the optimization value is [0, 1);

[0023] Input the standard case and each annotation case into the case evaluation model for analysis to obtain the corresponding case evaluation value.

[0024] Furthermore, when the corresponding annotation pattern and annotation model cannot be matched from the annotation pattern library according to the annotation requirements, assist the user in establishing the corresponding annotation pattern and annotation model.

[0025] Furthermore, the method for assisting the user in establishing the corresponding annotation pattern and annotation model includes:

[0026] Step SA1: Calculate the similarity between the requirement features and each annotation pattern, mark the annotation patterns with a similarity of not less than the threshold X1 as reference patterns, identify the annotation models corresponding to the reference patterns, and obtain the annotation principle of the annotation models; determine the corresponding adjustment items according to the annotation principle, and adjust the annotation principle according to the adjustment items and the requirement features to obtain the correction principle;

[0027] Step SA2: Perform annotation simulations according to the correction principle to generate multiple simulation data; display the simulation data to the user, and the user makes adjustments based on the displayed simulation data. Based on the adjustment results of the simulation data, the correction principle is adjusted to obtain a new correction principle.

[0028] Step SA3: Repeat step SA2 until no new correction principle is found;

[0029] Step SA4: Establish a training and adjustment model, obtain the training set corresponding to the labeled model, analyze the labeling principle, training set and correction principle through the training and adjustment model, and obtain training and adjustment data; when no training and adjustment data is obtained, remove the corresponding reference mode.

[0030] Step SA5: Based on the training adjustment data and annotation model, set up a new annotation model and annotation mode that meet the required features.

[0031] Step 4: Label the target data according to the matching annotation pattern and annotation model; obtain the target labeled data;

[0032] Step 5: Verify the target annotation data, adjust the target annotation data accordingly based on the verification results, and export the verified target annotation data;

[0033] Step Six: Continuously optimize, analyze, and adjust the annotation pattern library.

[0034] Furthermore, methods for continuous optimization and adjustment of the annotation pattern library include:

[0035] Real-time acquisition of material data, establishment of material evaluation models, analysis of each material data using the material evaluation models, and determination of the corresponding classification category for each material data;

[0036] The data is categorized according to its corresponding classification category to obtain the first, second, and third category data. The first, second, and third category data are then further categorized to obtain the optimized data for each unit. Finally, each optimized data unit is labeled with its corresponding classification category.

[0037] The identification unit optimizes the amount of data corresponding to the data, and performs corresponding optimization processing when the amount of data reaches the preset standard.

[0038] Furthermore, the expression for the material evaluation model is:

[0039] ;

[0040] In the formula: q represents the input data, which is the material data; the output data is the category value.

[0041] A tool for implementing the above data annotation method includes a data import and file indexing module, a visualization display module, a feature library management module, a pattern library management module, a statistics module, and a label processing module;

[0042] The data import and file indexing module is used to acquire the data to be labeled and to preprocess the data to be labeled.

[0043] The visualization display module is used to visualize the read sensor data according to a preset hierarchical management method;

[0044] The feature library management module is used to manage the feature library;

[0045] The pattern library management module is used to manage the pattern library;

[0046] The statistics module is used to display the statistical information of the tags in real time;

[0047] The label processing module is used to process multiple types of labels, set up a manual verification interface, and allow users to process the label results based on the manual verification interface; and export the corresponding data according to a preset format.

[0048] A storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described data annotation method.

[0049] Compared with the prior art, the beneficial effects of the present invention are:

[0050] A hierarchical management mechanism is introduced, allowing each layer of data to be loaded and displayed individually or simultaneously. Users can selectively load different data layers as needed. Customization options are also provided, such as adjusting curve colors, line styles, and label colors and shapes, to meet individual user needs and increase the flexibility and convenience of the user experience. Interactive functions, such as drag-and-drop and zoom operations, further enhance users' ability to manipulate and control data, making the annotation process more flexible and intuitive. Users can select existing annotation patterns or create new ones based on specific data annotation needs, providing great flexibility. Users can fine-tune the system according to data characteristics and annotation objectives to ensure accuracy and effectiveness. This flexibility makes the system more suitable for various annotation tasks.

[0051] By using preset annotation patterns, automated data annotation is achieved, significantly reducing the time and effort required for manual annotation and thus improving annotation efficiency. Simultaneously, data visualization capabilities allow users to intuitively observe data characteristics and annotation results, facilitating faster pattern identification, the development of annotation rules, and the verification of annotation accuracy. Users can easily manage multiple annotation patterns, including adding, deleting, modifying, and recalling patterns. New features and patterns can be stored, and existing features and annotation patterns can be reused. As data annotation tasks continue, new labeled data and optimized models can continuously update and maintain the pattern library, providing stronger support for future annotation tasks. Attached Figure Description

[0052] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0053] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0054] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0055] like Figure 1 As shown, a sensor data annotation method based on wearable devices includes:

[0056] Step 1: Import the data to be labeled from an external source. Perform preprocessing operations such as cleaning, noise reduction, and format conversion on the received data to improve data quality. Mark the preprocessed data as the target data.

[0057] Step 2: Select the data to be visualized. You can choose three axes, resultant acceleration, acceleration vector, and other feature data. Adjust the curve color, line type, and label color and shape as needed. Set the corresponding display method according to user requirements. The data will then be displayed according to the set display method.

[0058] Step 3: Establish a labeling pattern library and a feature library. The labeling pattern library is used to store various labeling patterns and the corresponding labeling models for each labeling pattern. One labeling pattern can correspond to multiple labeling models. The feature library is used to store various labeling features. Obtain the user's labeling requirements and match the corresponding labeling patterns and labeling models from the labeling pattern library according to the labeling requirements.

[0059] In the data annotation process, annotation rules or patterns are pre-defined. Users can choose a single pattern or a combination of multiple patterns. If a matching pattern exists in the annotation pattern library, the preset pattern is directly invoked, and the corresponding annotation model is selected. If no matching pattern is available, the user proceeds to the custom annotation pattern process. First, users can fine-tune and optimize existing annotation patterns to suit specific annotation needs. Second, if existing patterns cannot meet the requirements, users need to deeply analyze data characteristics and select relevant annotation features from the feature library, modify existing features, or create new ones. These annotation features should closely align with the core requirements of data annotation to ensure accuracy and effectiveness. Next, users need to associate the selected annotation features with the corresponding annotation models or rules. This involves configuring parameters and adjusting rule weights to ensure the newly defined annotation pattern accurately captures key information in the data. After configuration, users can choose to store the newly created annotation features and annotation patterns in the feature library and annotation pattern library for direct use in subsequent tasks.

[0060] Methods for matching corresponding annotation patterns and annotation models from the annotation pattern library according to annotation requirements include:

[0061] The annotation requirements are analyzed to determine their characteristics. This involves identifying and analyzing the annotation requirements to determine what types of data need to be annotated and what data needs to be annotated. Based on the data types and characteristics of the annotated data, the requirements characteristics are then determined. A test set is generated based on the requirements characteristics. The test set includes each piece of data to be labeled that meets the requirements characteristics, as well as the corresponding correctly annotated results. The user selects an annotation format that meets their requirements based on the test set or the requirements characteristics, which is considered a standard case.

[0062] The requirement features are input into the annotation pattern library for matching to obtain the candidate annotation methods that meet the requirement features, that is, the annotation methods that realize the data annotation of the requirement features, specifically the corresponding annotation patterns, annotation models or non-single combinations.

[0063] The candidate annotation methods are simulated and tested using a test set. The corresponding annotation accuracy and annotation consistency rates are statistically analyzed based on the simulation test results. The statistical analysis can be performed directly based on the test results. A correctly annotated annotation style is selected as an annotation case.

[0064] Evaluate each annotated case according to the standard case to determine whether the annotated case meets the requirements of the standard case, that is, whether the annotation forms are the same. If they are the same, it is considered to meet the requirements of the standard case; otherwise, it does not meet the requirements of the standard case. Based on this, establish a case evaluation model by combining existing identification and judgment technologies. The expression of the case evaluation model is ; where: s is the input data, and the input data includes the standard case and the annotated case; the output data is the case evaluation value; BA is the optimization value, which is set according to the difficulty of adjusting from the annotated case to the standard case, and the value range is [0, 1). Obtain the possible adjustment work, sort the adjustment work in ascending order of difficulty, and the sorting can be based on the adjustment workload, adjustment duration, etc. In the case where no adjustment can be made, the optimization value is 0, and the others are distributed between 0 and 1. Set the corresponding optimization values according to the difficulty differences between them, and then perform matching later. It can be set or verified by professionals;

[0065] Input the standard case and each annotated case into the case evaluation model for analysis to obtain the corresponding case evaluation value;

[0066] Calculate the corresponding annotation value according to the formula PY = b1 × PA(s) × (b2 × QA + B3 × QB);

[0067] where: PY is the annotation value; b1, b2, and b3 are all proportionality coefficients, and the value range is 0 < b1 ≤ 1, 0 < b2 ≤ 1, 0 < b3 ≤ 1; PA(s) is the case evaluation value output by the case evaluation model; QA is the annotation accuracy rate; QB is the annotation consistency rate.

[0068] Sort each candidate annotation method in descending order of the annotation value to obtain the first sequence, and insert the corresponding annotated case at each candidate annotation method in the first sequence;

[0069] According to the candidate annotation method selected by the user in the first sequence, mark the selected candidate annotation method as the target annotation method, and identify the annotation mode and annotation model corresponding to the target annotation method.

[0070] In one embodiment, if there is no matching annotation mode available in the annotation mode library, the user can only perform custom processing manually, which requires the user to have certain professional knowledge and spend a lot of effort, and the efficiency is not high. Based on this, when the corresponding annotation mode and annotation model cannot be matched from the annotation mode library according to the annotation requirements, assist the user to establish the corresponding annotation mode and annotation model; the detailed process is as follows:

[0071] The similarity between the requirement features and each annotation pattern is calculated. Annotation patterns with a similarity of at least X1 are marked as reference patterns. The annotation model corresponding to the reference pattern is identified, and the annotation principle of the annotation model is obtained, such as how input data is transformed into output data. Based on the annotation principle, the corresponding adjustment items are determined, that is, the parts that need to be adjusted when the annotation principle is adjusted to meet the requirement features. The annotation principle is adjusted according to the adjustment items and requirement features to obtain the correction principle. Simulated annotation is performed according to the correction principle to obtain simulated data. That is, the annotation principle is adjusted according to the requirement features to obtain the correction principle, and then annotation simulation is performed according to the correction principle to form simulated data. There are multiple simulated data sets. The simulated data is displayed to the user, and the user adjusts according to the displayed simulated data, that is, adjusts to meet the requirements. The annotation effect meets the user's needs; the correction principle is adjusted according to the adjustment process; simulated data is generated again according to the adjusted correction principle, and so on, until the user no longer adjusts the simulated data; a training adjustment model is established based on a neural network, and a corresponding training set is established manually for training. The training set includes input data and output data. The input data is the original training set of the annotation model, the annotation principle, and the correction principle; the output data is the training adjustment data after adjusting the original training set according to the correction principle; the corresponding training adjustment data is obtained by analyzing the successfully trained training adjustment model; in other embodiments, the training adjustment data can also be determined based on other existing technologies; reference patterns for which training adjustment data cannot be obtained are eliminated.

[0072] The labeled model is trained and adjusted using training data to obtain a new labeled model, which is then marked as the initial model. The original labeled model is still saved. The initial model is tested using a test set, and the corresponding labeled values ​​are calculated. A new labeled model and labeled mode are set based on the labeled values. That is, the labeled model with the largest labeled value is selected, and a new labeled mode is formed based on the standard model.

[0073] Step 4: Label the target data according to the matching annotation pattern and annotation model; obtain the target labeled data;

[0074] Step 5: Verify the target annotation data and adjust it accordingly based on the verification results. This is generally done manually to ensure accuracy and consistency. If necessary, adjust or correct the annotation results. Export the verified target annotation data.

[0075] When generating multiple labels, you can choose to save them as separate columns or as a single column. In this case, the tool will trigger conflict resolution and result integration to obtain the final label result. Users can select the data range and the exported data format, and save the generated labels to the storage medium.

[0076] Step Six: Continuously optimize, analyze, and adjust the annotation pattern library.

[0077] As data annotation tasks continue, new annotation data is collected, including raw data and modified labels, and labels are checked and updated regularly to optimize existing patterns, create new models, or adjust annotation strategies.

[0078] Once a sufficient amount of new labeled data has been collected, this data can be used to train new models or to fine-tune existing models, and then undergo thorough training and validation. When a model is validated as effective, and consistency and compatibility with the existing model library are ensured, it can be added to the newly created or existing model library. New models must include information such as the model's name, description, parameters, and associated model.

[0079] In one embodiment, the annotation pattern library can be continuously optimized, analyzed, and adjusted in the following manner:

[0080] The adjusted target annotation data after verification is integrated with the original data to form source data. The source data is then evaluated to determine whether the adjustment is based on the original baseline or a complete re-annotation; that is, whether a new annotation model or annotation pattern needs to be established to annotate the data. This can be identified based on the differences before and after the adjustment. Therefore, the source data is divided into three categories: Category 1, Category 2, and Category 3. Category 1 indicates that only the existing annotation pattern or annotation model needs to be optimized and adjusted; Category 2 indicates that a new annotation pattern needs to be established; and Category 3 indicates that a new annotation model needs to be established.

[0081] Based on this, a corresponding material evaluation model is established, and the expression of the material evaluation model is as follows: In the formula: q represents the input data, which is the source material data;

[0082] The material evaluation model is used to analyze the data of each material and determine the corresponding classification category of each material.

[0083] The data is categorized according to its corresponding classification category to obtain the first, second, and third category data, respectively. Within each category, the data is further categorized according to a labeling model or labeling mode to obtain optimized data for each unit. This involves classifying the data based on the corresponding labeling model or mode and assigning a corresponding category label to each optimized unit. The data volume corresponding to each optimized unit is identified; when the data volume reaches a preset standard (i.e., exceeds a preset quantity), appropriate optimization is performed, such as optimizing the original labeling mode or model or establishing a new original labeling mode or model. Specific quantity standards are set based on the required data volume.

[0084] A tool for implementing the above-described data annotation method, the apparatus comprising:

[0085] Data import and file indexing module:

[0086] This module can import and read data to be annotated via file import, network interface, etc. It features a file type filtering function to ensure that only user-defined file types can be selected, and verifies the correct data format of the selected file to reduce problems caused by selecting the wrong file type. Additionally, it adds an index function for files within a folder, allowing users to quickly switch files within the same folder, reducing manual navigation time. When a user selects a file outside the specified folder, the previous / next buttons are disabled to prevent accidental operation in the wrong folder.

[0087] Visualization module:

[0088] The system visualizes the acquired sensor data. A hierarchical management mechanism is introduced, making data visualization clearer and more flexible. Two-dimensional coordinate axes and grid lines serve as the bottom layer, while the data layer is the top layer. Each data layer can be loaded and displayed individually or simultaneously; users can selectively load different data layers as needed. Customization options are provided, such as allowing users to adjust curve colors, line styles, and label colors and shapes to meet individual user requirements. Interactive features, such as dragging and zooming, are also included.

[0089] Feature library management module:

[0090] This module allows users to create, store, and retrieve features to support data analysis and model building. Users can edit existing features or add new features to the feature library, and features are stored in modules.

[0091] Pattern Library Management Module:

[0092] This module allows users to define and manage multiple annotation patterns, each with multiple associated models or rules. Each pattern includes information such as name, description, parameters, and associated models. Users can add, delete, and modify models. Configured patterns are available for direct use.

[0093] Statistics module:

[0094] It displays real-time statistics on tags, including the number of tags, frequency distribution, and related statistical data.

[0095] Tag processing module:

[0096] For multi-type annotation, it features conflict resolution and result integration capabilities. It provides a manual verification interface, allowing users to modify the label results. It also offers one-click cancellation of all labels, undo, and redo functions to help users quickly revert to a previous state in case of operational errors. Finally, it generates labels corresponding to the three-axis positions and saves them to storage media, supporting data export in various formats.

[0097] A storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described data annotation method.

[0098] A hierarchical management mechanism is introduced, allowing each layer of data to be loaded and displayed individually or simultaneously. Users can selectively load different data layers as needed. Customization options are also provided, such as adjusting curve colors, line styles, and label colors and shapes, to meet individual user needs and increase the flexibility and convenience of the user experience. Interactive functions, such as drag-and-drop and zoom operations, further enhance users' ability to manipulate and control data, making the annotation process more flexible and intuitive. Users can select existing annotation patterns or create new ones based on specific data annotation needs, providing great flexibility. Users can fine-tune the system according to data characteristics and annotation objectives to ensure accuracy and effectiveness. This flexibility makes the system more suitable for various annotation tasks.

[0099] By using preset annotation patterns, automated data annotation is achieved, significantly reducing the time and effort required for manual annotation and thus improving annotation efficiency. Simultaneously, data visualization capabilities allow users to intuitively observe data characteristics and annotation results, facilitating faster pattern identification, the development of annotation rules, and the verification of annotation accuracy. Users can easily manage multiple annotation patterns, including adding, deleting, modifying, and recalling patterns. New features and patterns can be stored, and existing features and annotation patterns can be reused. As data annotation tasks continue, new labeled data and optimized models can continuously update and maintain the pattern library, providing stronger support for future annotation tasks.

[0100] The above formulas are all numerical calculations after removing dimensions. The formulas are obtained by software simulation based on a large amount of data and are closest to the real situation. The preset parameters and preset thresholds in the formulas are set by those skilled in the art according to the actual situation or obtained by simulation based on a large amount of data.

[0101] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, apparatus, article, or method that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, apparatus, article, or method. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, apparatus, article, or method that includes that element.

[0102] The above description is merely a preferred embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the description and drawings of the present invention, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

[0103] 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 invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

[0104] For ease of description, the above apparatus is described in terms of function, divided into various units. Of course, in implementing this application, the functions of each unit can be implemented in one or more software and / or hardware components. Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0105] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, as well as combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart. Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0106] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 The function specified in one or more boxes.

[0107] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

Claims

1. A sensor data annotation method based on wearable devices, characterized in that, The method includes: Obtain the data of the file to be annotated, preprocess the received data of the file to be annotated, and mark the preprocessed data of the file to be annotated as target data; Set the corresponding display mode according to the user's requirements, and visually display the read data according to the display mode; Establish an annotation mode library and a feature library. The annotation mode library is used to store various annotation modes and the annotation models corresponding to each annotation mode; the feature library is used to store various annotation features; obtain the user's annotation requirements, and match the corresponding annotation mode and annotation model from the annotation mode library according to the annotation requirements; Annotate the target data according to the matched annotation mode and annotation model; obtain the target annotation data; Verify the target annotation data, and make corresponding adjustments to the target annotation data according to the verification result; export the verified target annotation data; Continuously optimize, analyze and adjust the annotation mode library; The method for matching the corresponding annotation mode and annotation model from the annotation mode library according to the annotation requirements includes: Analyze the annotation requirements to determine the requirement features; generate a test set according to the requirement features, and set standard cases according to the test set; Input the requirement features into the annotation mode library for matching to obtain various candidate annotation methods that meet the requirement features; Perform simulation tests on various candidate annotation methods through the test set to obtain the annotation accuracy rate and annotation consistency rate corresponding to each candidate annotation method; and set the annotation cases corresponding to each candidate annotation method; Evaluate each annotation case according to the standard case to obtain the case evaluation value corresponding to each candidate annotation method; Calculate the corresponding annotation value according to the formula PY = b1 × PA(s) × (b2 × QA + B3 × QB); In the formula: PY is the annotation value; b1, b2, and b3 are all proportionality coefficients, and the value range is 0 < b1 ≤ 1, 0 < b2 ≤ 1, 0 < b3 ≤ 1; PA(s) is the case evaluation value output by the case evaluation model; QA is the annotation accuracy rate; QB is the annotation consistency rate; Sort the various candidate annotation methods in descending order according to the annotation value to obtain the first sequence, and insert the corresponding annotation cases at each candidate annotation method in the first sequence; Determine the target annotation method according to the first sequence, and identify the annotation mode and annotation model corresponding to the target annotation method; When the corresponding annotation mode and annotation model cannot be matched from the annotation mode library according to the annotation requirements, assist the user in establishing the corresponding annotation mode and annotation model; The method for assisting the user in establishing the corresponding annotation mode and annotation model includes: Step SA1: Calculate the similarity between the requirement features and each annotation mode, mark the annotation modes with similarity not lower than the threshold X1 as reference modes, identify the annotation models corresponding to the reference modes, and obtain the annotation principles of the annotation models; determine the corresponding adjustment items according to the annotation principles, and adjust the annotation principles according to the adjustment items and requirement features to obtain the corrected principle; Step SA2: Perform annotation simulation according to the corrected principle to form multiple simulation data; display the simulation data to the user, the user makes adjustments according to the displayed simulation data, and adjust the corrected principle according to the adjustment results of the simulation data to obtain a new corrected principle; Step SA3: Repeat step SA2 until no new correction principle is found; Step SA4: Establish a training and adjustment model, obtain the training set corresponding to the labeled model, analyze the labeling principle, training set and correction principle through the training and adjustment model, and obtain training and adjustment data; when no training and adjustment data is obtained, remove the corresponding reference mode. Step SA5: Based on the training adjustment data and annotation model, set up a new annotation model and annotation mode that meet the required features.

2. The sensor data annotation method based on wearable devices according to claim 1, characterized in that, The methods for evaluating each labeled case based on standard cases include: Establish a case evaluation model, the expression of which is: ； In the formula: s represents the input data, which includes standard cases and labeled cases; the output data is the case evaluation value; BA represents the optimization value, which ranges from [0, 1); The standard cases and each labeled case are input into the case evaluation model for analysis to obtain the corresponding case evaluation values.

3. The sensor data annotation method based on wearable devices according to claim 1, characterized in that, Methods for continuous optimization, analysis, and adjustment of the annotation pattern library include: Real-time acquisition of material data, establishment of material evaluation models, analysis of each material data using the material evaluation models, and determination of the corresponding classification category for each material data; The data is categorized according to its corresponding classification category to obtain the first, second, and third category data. The first, second, and third category data are then further categorized to obtain the optimized data for each unit. Finally, each optimized data unit is labeled with its corresponding classification category. The identification unit optimizes the amount of data corresponding to the data, and performs corresponding optimization processing when the amount of data reaches the preset standard.

4. The sensor data annotation method based on wearable devices according to claim 3, characterized in that, The expression for the material evaluation model is: ； In the formula: q represents the input data, which is the material data; the output data is the category value.

5. A sensor data annotation tool based on wearable devices, characterized in that, The sensor data annotation method based on wearable devices according to any one of claims 1 to 4 includes a data import and file indexing module, a visualization display module, a feature library management module, a pattern library management module, a statistics module, and a tag processing module; The data import and file indexing module is used to acquire the data to be labeled and to preprocess the data to be labeled. The visualization display module is used to visualize the read sensor data according to a preset hierarchical management method; The feature library management module is used to manage the feature library; The pattern library management module is used to manage the pattern library; The statistics module is used to display the statistical information of the tags in real time; The label processing module is used to process multiple types of labels, set up a manual verification interface, and allow users to process the label results based on the manual verification interface; and export the corresponding data according to a preset format.

6. A storage medium, characterized in that, The storage medium stores a computer program, which, when executed by a processor, implements the sensor data annotation method based on a wearable device as described in any one of claims 1 to 4.

Images