Lower limb motor dysfunction detection method and device based on data fusion and medium

Abstract

The application provides a lower limb motor dysfunction detection method and device based on data fusion and a medium. The method comprises the following steps: determining a first probability and a second probability that a target object has lower limb motor dysfunction; the first probability is determined based on kinematic data of the target object; the second probability is determined based on plantar pressure data of the target object; fusing the first probability and the second probability to determine a third probability; and detecting whether the target object has lower limb motor dysfunction based on the third probability. The lower limb motor dysfunction detection method based on data fusion can detect lower limb motor dysfunction of the target object from multiple dimensions by fusing the kinematic data and the plantar pressure data, and the detection process is simple and the detection accuracy is high.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a method, device and medium for detecting lower limb motor dysfunction based on data fusion. Background Technology

[0002] Incorrect walking posture or various diseases can lead to lower limb motor dysfunction. This dysfunction can manifest through various indicators. For example, during walking, to avoid pain caused by weight-bearing on the affected side, the trunk may tilt, shortening the stance phase of the hip joint. Pain during stepping can result in decreased walking speed and stride length. Abnormalities may also occur in joint flexion, external rotation, and adduction. These abnormalities are also reflected in plantar pressure data. Muscle atrophy, abnormal joint movement, and shifted center of gravity all contribute to different stress distribution on the feet during walking compared to healthy individuals.

[0003] Currently, the commonly used method for evaluating motor function is scale-based, which has the disadvantages of being cumbersome in the testing process and relying on the doctor's subjective judgment. Therefore, objective methods for evaluating motor function are particularly important. Summary of the Invention

[0004] To address the problems existing in the prior art, embodiments of the present invention provide a method, device, and medium for detecting lower limb motor dysfunction based on data fusion.

[0005] This invention provides a method for detecting lower limb motor dysfunction based on data fusion, comprising:

[0006] A first probability and a second probability of the target object having lower limb motor dysfunction are determined; the first probability is determined based on the target object's kinematic data; the second probability is determined based on the target object's plantar pressure data; the first probability and the second probability are fused to determine a third probability;

[0007] Based on the third probability, it is detected whether the target object has lower limb motor dysfunction.

[0008] In some embodiments, the kinematic data includes hip joint angle data;

[0009] The first probability of determining that the target object has lower limb motor dysfunction includes:

[0010] Collect hip joint angle data of the left and right legs of the target object during the gait cycle;

[0011] Determine the maximum and minimum angles from the hip joint angle data;

[0012] The first feature is determined based on the mean, standard deviation, and coefficient of variation of the maximum angle, and the mean, standard deviation, and coefficient of variation of the minimum angle;

[0013] Based on the first feature, the first probability is determined.

[0014] In some embodiments, determining the first probability based on the first feature includes:

[0015] The motion obstacle detection model is trained based on sample kinematic data to determine the first recognition model; the motion obstacle detection model is built based on support vector machine.

[0016] The first feature is input into the first recognition model to obtain the first probability.

[0017] In some embodiments, determining the second probability that the target object has lower limb motor dysfunction includes:

[0018] Collect plantar pressure data of the left and right legs of the target object during the gait cycle;

[0019] The maximum pressure value and target duration are determined based on the plantar pressure data; the target duration is the duration corresponding to the pressure value in the plantar pressure data being greater than the preset pressure value within the gait cycle.

[0020] The second feature is determined based on the average, standard deviation, and coefficient of variation of the maximum pressure value, and the average, standard deviation, and coefficient of variation of the target duration;

[0021] Based on the second feature, the second probability is determined.

[0022] In some embodiments, determining the second probability based on the second feature includes:

[0023] The movement obstacle detection model is trained based on sample plantar pressure data to determine the second recognition model; the movement obstacle detection model is built based on support vector machine.

[0024] The second feature is input into the second recognition model to obtain the second probability.

[0025] In some embodiments, fusing the first probability and the second probability to determine the third probability includes:

[0026] The decision fusion model was validated based on sample kinematic data and sample plantar pressure data to determine the third identification model; the decision fusion model was constructed based on the decision fusion algorithm.

[0027] The first probability and the second probability are input into the third recognition model to obtain the third probability.

[0028] The present invention also provides a lower limb motor dysfunction detection device based on data fusion, comprising:

[0029] The determination module is used to determine a first probability and a second probability that the target object has lower limb motor dysfunction; the first probability is determined based on the kinematic data of the target object; the second probability is determined based on the plantar pressure data of the target object.

[0030] The fusion module is used to fuse the first probability and the second probability to determine the third probability;

[0031] The detection module is used to detect the lower limb motor function of the target object based on the third probability.

[0032] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the program to implement the lower limb motor dysfunction detection method based on data fusion as described above.

[0033] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the lower limb motor dysfunction detection method based on data fusion as described above.

[0034] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the data fusion-based method for detecting lower limb motor dysfunction as described above.

[0035] The present invention provides a method, device and medium for detecting lower limb motor dysfunction based on data fusion. By integrating kinematic data and plantar pressure data, it can detect lower limb motor dysfunction of target objects from multiple dimensions. The detection process is simple and the detection accuracy is high. Attached Figure Description

[0036] To more clearly illustrate the technical solutions in this 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 some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0037] Figure 1 This is one of the flowcharts of the lower limb motor dysfunction detection method based on data fusion provided in the embodiments of the present invention;

[0038] Figure 2 This is the second flowchart of the method for detecting lower limb motor dysfunction based on data fusion provided in this embodiment of the invention;

[0039] Figure 3 This is a schematic diagram of the structure of the lower limb motor dysfunction detection device based on data fusion provided in an embodiment of the present invention;

[0040] Figure 4 This is a schematic diagram of the structure of the electronic device provided in an embodiment of the present invention. Detailed Implementation

[0041] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0042] The terms "first," "second," etc., used in the specification and claims of this invention are used to distinguish similar objects and not to describe a specific order or sequence. It should be understood that such terms can be used interchangeably where appropriate so that embodiments of the invention can be implemented in orders other than those illustrated or described herein, and the objects distinguished by "first" and "second" are generally of the same class, not limited in number; for example, the first object can be one or more. Furthermore, in the specification and claims, "and / or" indicates at least one of the connected objects, and the character " / " generally indicates that the preceding and following objects are in an "or" relationship.

[0043] Figure 1 This is one of the flowcharts of the lower limb motor dysfunction detection method based on data fusion provided in this embodiment of the invention, such as... Figure 1 As shown, the data fusion-based method for detecting lower limb motor dysfunction provided in this embodiment of the invention includes:

[0044] Step 101: Determine the first probability and the second probability that the target object has lower limb motor dysfunction; the first probability is determined based on the kinematic data of the target object; the second probability is determined based on the plantar pressure data of the target object.

[0045] Step 102: Combine the first probability and the second probability to determine the third probability;

[0046] Step 103: Based on the third probability, detect whether the target object has lower limb motor dysfunction.

[0047] It should be noted that the execution subject of the data fusion-based lower limb motor dysfunction detection method provided by this invention can be an electronic device, a component in an electronic device, an integrated circuit, or a chip. The electronic device can be a mobile electronic device or a non-mobile electronic device. For example, mobile electronic devices can be mobile phones, tablets, laptops, PDAs, in-vehicle electronic devices, wearable devices, ultra-mobile personal computers (UMPCs), netbooks, or personal digital assistants (PDAs), etc., while non-mobile electronic devices can be servers, network-attached storage (NAS), personal computers (PCs), televisions (TVs), ATMs, or self-service machines, etc. This invention does not impose specific limitations.

[0048] In step 101, a first probability and a second probability of the target object having lower limb motor dysfunction are determined; the first probability is determined based on the kinematic data of the target object; the second probability is determined based on the plantar pressure data of the target object.

[0049] Kinesiological data can include: stride speed and stride length, joint flexion, external rotation, adduction, etc.

[0050] In this embodiment of the invention, the kinetic data uses hip joint angle data, which can be acquired through acquisition devices such as motion capture systems and inertial measurement units.

[0051] In some embodiments, determining the first probability that the target object has lower limb motor dysfunction includes:

[0052] Collect hip joint angle data of the left and right legs of the target object during the gait cycle;

[0053] Determine the maximum and minimum angles from the hip joint angle data;

[0054] The first feature is determined based on the mean, standard deviation, and coefficient of variation of the maximum angle, and the mean, standard deviation, and coefficient of variation of the minimum angle;

[0055] Based on the first feature, the first probability is determined.

[0056] Feature extraction was performed on the collected kinematic data. Motor dysfunction can be reflected in a decrease in joint range of motion. Therefore, angle data in the x, y, and z directions of the left and right legs were selected. The maximum and minimum angles of each gait cycle in a single walking data set were calculated. Then, the mean, standard deviation, and coefficient of variation of these angle data were calculated as the first feature.

[0057] Specifically, a gait cycle refers to the process of walking from the moment the heel leaves the ground until it lands again. A single walking data point can include multiple gait cycles.

[0058] During each gait cycle, the angle data of the left hip joint in the x-axis, y-axis, and z-axis directions are collected to determine the maximum and minimum angles in the x-axis direction. The mean, standard deviation, and coefficient of variation of the maximum angle of the left hip joint in the x-axis direction, as well as the mean, standard deviation, and coefficient of variation of the minimum angle of the left hip joint in the x-axis direction, are calculated.

[0059] And determine the maximum and minimum angles in the y-axis direction. Calculate the mean, standard deviation, and coefficient of variation of the maximum angle of the left hip joint in the y-axis direction, and the mean, standard deviation, and coefficient of variation of the minimum angle of the left hip joint in the y-axis direction.

[0060] In addition, determine the maximum and minimum angles in the z-axis direction. Calculate the mean, standard deviation, and coefficient of variation of the maximum angle of the left hip joint in the z-axis direction, as well as the mean, standard deviation, and coefficient of variation of the minimum angle of the left hip joint in the z-axis direction.

[0061] Similarly, during each gait cycle, the angle data of the right hip joint in the x-axis, y-axis, and z-axis directions are collected to determine the maximum and minimum angles in the x-axis direction. The mean, standard deviation, and coefficient of variation of the maximum angle of the right hip joint in the x-axis direction, as well as the mean, standard deviation, and coefficient of variation of the minimum angle of the right hip joint in the x-axis direction, are calculated.

[0062] Determine the maximum and minimum angles along the y-axis. Calculate the mean, standard deviation, and coefficient of variation of the maximum angle of the right hip joint along the y-axis, and the mean, standard deviation, and coefficient of variation of the minimum angle of the right hip joint along the y-axis.

[0063] Determine the maximum and minimum angles along the z-axis. Calculate the mean, standard deviation, and coefficient of variation of the maximum angle of the right hip joint along the z-axis, and the mean, standard deviation, and coefficient of variation of the minimum angle of the right hip joint along the z-axis.

[0064] That is, 12 sets of data can be obtained from a single walking data set. Each set of data includes the mean, standard deviation, and coefficient of variation, which serve as the first feature.

[0065] By extracting the mean, standard deviation, and coefficient of variation of the maximum and minimum hip joint angle data as features, motor dysfunction of the target object can be detected from multiple angles, improving the accuracy of identification.

[0066] In some embodiments, determining the first probability that the target object has lower limb motor dysfunction includes:

[0067] The motion obstacle detection model is trained based on sample kinematic data to determine the first recognition model; the motion obstacle detection model is built based on support vector machine.

[0068] The first feature is input into the first recognition model to obtain the first probability.

[0069] Specifically, after extracting the first feature corresponding to the kinematic data, a recognition model based on the kinematic features can be trained to detect movement disorders.

[0070] Alternatively, a Support Vector Machine (SVM) can be used to build a motion obstacle detection model Mdl based on single kinematic data. 1 The motion obstacle detection model is trained using pre-collected sample kinematic data and the features corresponding to the sample kinematic data to obtain the first trained recognition model.

[0071] For the i-th set of kinematic data collected, its corresponding first feature is: Inputting this into the first recognition model yields the probability that the target object is classified as having a motor dysfunction. The expression is as follows:

[0072]

[0073] By combining artificial intelligence algorithms for motor function evaluation, the method of motor function evaluation becomes more objective and accurate.

[0074] In some embodiments, determining the second probability that the target object has lower limb motor dysfunction includes:

[0075] Collect plantar pressure data of the left and right legs of the target object during the gait cycle;

[0076] The maximum pressure value and target duration are determined based on the plantar pressure data; the target duration is the duration corresponding to the pressure value in the plantar pressure data being greater than the preset pressure value within the gait cycle.

[0077] The second feature is determined based on the average, standard deviation, and coefficient of variation of the maximum pressure value, and the average, standard deviation, and coefficient of variation of the target duration;

[0078] Based on the second feature, the second probability is determined.

[0079] Plantar pressure data can be obtained through plantar pressure insoles, force plates, and other data collection devices.

[0080] Feature extraction was performed on the collected plantar pressure data. Motor dysfunction is also reflected in changes in the body's weight-bearing pattern; the magnitude and direction of the force when the foot contacts the ground change to some extent during walking.

[0081] Therefore, plantar pressure data in the x, y, and z directions of the left and right legs can be selected to calculate the maximum pressure value and target duration for each gait cycle in a single walking data set. The target duration is the duration during which the pressure value in the plantar pressure data is greater than a preset pressure value within the gait cycle. The preset pressure value can be set to 20% of the maximum pressure value. Then, the mean, standard deviation, and coefficient of variation of these data are calculated as the second feature.

[0082] Specifically, plantar pressure data of the left leg in the x-axis, y-axis, and z-axis directions are collected during each gait cycle to determine the maximum pressure value and target duration in the x-axis direction. The mean, standard deviation, and coefficient of variation of the maximum pressure value in the x-axis direction, as well as the mean, standard deviation, and coefficient of variation of the target duration in the x-axis direction, are calculated.

[0083] In addition, determine the maximum pressure value and target duration in the y-axis direction. Calculate the mean, standard deviation, and coefficient of variation of the maximum pressure value in the y-axis direction, as well as the mean, standard deviation, and coefficient of variation of the target duration in the y-axis direction.

[0084] In addition, determine the maximum pressure value and target duration in the z-axis direction. Calculate the mean, standard deviation, and coefficient of variation of the maximum pressure value in the z-axis direction, as well as the mean, standard deviation, and coefficient of variation of the target duration in the z-axis direction.

[0085] Similarly, during each gait cycle, plantar pressure data of the right leg in the x, y, and z axes are collected to determine the maximum pressure value and target duration in the x-axis direction. The mean, standard deviation, and coefficient of variation of the maximum pressure value in the x-axis direction, as well as the mean, standard deviation, and coefficient of variation of the target duration in the x-axis direction, are calculated.

[0086] In addition, determine the maximum pressure value and target duration in the y-axis direction. Calculate the mean, standard deviation, and coefficient of variation of the maximum pressure value in the y-axis direction, as well as the mean, standard deviation, and coefficient of variation of the target duration in the y-axis direction.

[0087] In addition, determine the maximum pressure value and target duration in the z-axis direction. Calculate the mean, standard deviation, and coefficient of variation of the maximum pressure value in the z-axis direction, as well as the mean, standard deviation, and coefficient of variation of the target duration in the z-axis direction.

[0088] That is, 12 sets of data can be obtained from a single walking data set. Each set of data includes the mean, standard deviation, and coefficient of variation, which serve as the second feature.

[0089] By extracting the maximum pressure value of plantar pressure data, as well as the average, standard deviation, and coefficient of variation of the target duration, as features, motor dysfunction of the target object can be detected from multiple dimensions, improving the accuracy of identification.

[0090] In some embodiments, determining the second probability based on the second feature includes:

[0091] The movement obstacle detection model is trained based on sample plantar pressure data to determine the second recognition model; the movement obstacle detection model is built based on support vector machine.

[0092] The second feature is input into the second recognition model to obtain the second probability.

[0093] Specifically, after extracting the second feature corresponding to the plantar pressure data, a recognition model based on the plantar pressure feature can be trained to detect movement disorders.

[0094] Similarly, a movement obstacle detection model Mdl based on plantar pressure data can be built using support vector machines. 2 The movement obstacle detection model is trained using pre-collected plantar pressure data and the features corresponding to the plantar pressure data to obtain a trained second recognition model.

[0095] For the i-th set of plantar pressure data collected, its corresponding second feature is: Inputting this into the second recognition model yields the probability that the target object is identified as having a motor dysfunction. The expression is as follows:

[0096]

[0097] By combining artificial intelligence algorithms for motor function evaluation, the method of motor function evaluation becomes more objective and accurate.

[0098] In step 102, the first probability and the second probability are fused together to determine the third probability.

[0099] After obtaining the probabilities output by two movement dysfunction detection models based on single-modal data, a decision fusion algorithm can be used to establish a movement dysfunction detection model based on decision fusion.

[0100] The output probabilities of two detection models based on single-modal data are used as input features and fed into a data fusion-based motor dysfunction detection model. A weighted average rule of probabilities is used as the fusion algorithm, and the calculation formula is as follows:

[0101]

[0102] in, and Let ω1 and ω2 represent the probabilities of two recognition models based on single-modal data identifying a motor dysfunction, respectively. ω1 and ω2 represent the weights of the output probabilities of the two recognition models, and they have the following relationship:

[0103] ω1+ω2=1

[0104] A decision fusion model Mdl is established based on the probability weighted average rule. f That is, ω1 and ω2 are the model parameters of the decision fusion model.

[0105] Unlike machine learning algorithms, the weighted average rule for probabilities used in this embodiment of the invention only contains two parameters, ω1 and ω2, and the two parameters have a relationship that the sum is 1. Therefore, all parameters can be directly traversed to obtain all models, and then the model with the highest accuracy can be selected using the validation set.

[0106] The validation set is determined based on the pre-collected sample kinematic data and sample plantar pressure data, the features corresponding to the sample kinematic data, the features corresponding to the sample plantar pressure data, and the probabilities output by the two recognition models respectively.

[0107] By iterating through all weight values ​​within the range [0,1] of ω1 and ω2 with a step size of 0.001, a decision fusion model can be obtained. Specifically, 1000 decision fusion models can be obtained.

[0108] The model with the highest recognition accuracy is selected from 1000 decision fusion models and used as the final trained third recognition model.

[0109] The third recognition model fuses the outputs of the detection models based on kinematic data and plantar pressure data, resulting in a third probability of D. i The expression is as follows:

[0110]

[0111] That is, for the collected data, after passing through two detection models and a decision fusion model based on single-modal data, the probability D of detecting motor dysfunction can be obtained. i .

[0112] In step 103, based on the third probability, it is detected whether the target object has lower limb motor dysfunction.

[0113] Therefore, based on the fused third probability, it is possible to detect whether the target object has lower limb motor dysfunction.

[0114] Optionally, the third probability can be compared with the preset probability. If the third probability is greater than or equal to the preset probability, it is determined that the target object has lower limb motor dysfunction. If the third probability is less than the preset probability, it is determined that the target object does not have lower limb motor dysfunction.

[0115] The data fusion-based method for detecting lower limb motor dysfunction provided in this invention integrates kinematic data and plantar pressure data to detect lower limb motor dysfunction in target subjects from multiple dimensions. The detection process is simple and has a high accuracy rate.

[0116] Figure 2 This is the second flowchart of the lower limb motor dysfunction detection method based on data fusion provided in this embodiment of the invention, as shown below. Figure 2 As shown, the data fusion-based method for detecting lower limb motor dysfunction provided in this embodiment of the invention includes:

[0117] Feature extraction was performed on the collected kinematic data. Angle data in the x, y, and z directions of the left and right legs were selected. The maximum and minimum angles of each gait cycle in a single walking sequence were calculated. Then, the mean, standard deviation, and coefficient of variation of these angle data were calculated as features.

[0118] Feature extraction was performed on the collected plantar pressure data. Plantar pressure data in the x, y, and z directions of both legs were selected. The maximum pressure value and target duration for each gait cycle in a single walking session were calculated. Then, the mean, standard deviation, and coefficient of variation of these data were calculated as features.

[0119] After obtaining the features of the two modalities, a kinematic feature-based recognition model is first trained to detect movement obstacles. A support vector machine is then used to establish a movement obstacle detection model Mdl based on single kinematic data. 1 .

[0120] For the i-th set of data, the first feature is The first feature is input into the first recognition model, and the probability of judging it as motor dysfunction is:

[0121]

[0122] Then, a recognition model based on plantar pressure features is trained to detect movement disorders. Similarly, a support vector machine is used to build a movement disorder detection model Mdl based on single plantar pressure data. 2 .

[0123] For the i-th set of data, the second feature is The second feature is input into the second recognition model, and the probability of judging it as motor dysfunction is:

[0124]

[0125] After obtaining two motor function disorder (MFD) detection models based on single-modal data, a decision fusion algorithm was further used to establish a decision fusion-based MFD detection model. The output probabilities of the two single-modal data-based detection models were used as input features and fed into the decision fusion-based MFD detection model. A weighted average rule of probabilities was used as the fusion algorithm, and the specific calculation formula is as follows:

[0126]

[0127] in, and Let ω1 and ω2 represent the probabilities of two recognition models based on single-modal data identifying a motor dysfunction, respectively. ω1 and ω2 represent the weights of the output probabilities of these two recognition models, and they have the following relationship:

[0128] ω1+ω2=1

[0129] A fusion model Mdl is established based on the weighted average rule of probability. f This model fuses the outputs of detection models based on kinematic data and plantar pressure data, namely:

[0130]

[0131] like Figure 2 As shown, for the collected data, after passing through two detection models and a decision fusion model based on single-modal data, the probability D of detecting motor dysfunction can be obtained. i .

[0132] The data fusion-based method for detecting lower limb motor dysfunction provided by this invention effectively integrates kinematic data and plantar pressure data to analyze the motor function of the target object from multiple perspectives. Combined with artificial intelligence algorithms, it will promote a more objective and accurate method for evaluating motor function.

[0133] The following describes the lower limb motor dysfunction detection device based on data fusion provided by the present invention. The lower limb motor dysfunction detection device based on data fusion described below and the lower limb motor dysfunction detection method based on data fusion described above can be referred to and correspond to each other.

[0134] Figure 3 This is a schematic diagram of the structure of the lower limb motor dysfunction detection device based on data fusion provided in an embodiment of the present invention, as shown below. Figure 3 As shown, the lower limb motor dysfunction detection device based on data fusion provided in this embodiment of the invention includes:

[0135] The determining module 310 is used to determine a first probability and a second probability that the target object has lower limb motor dysfunction; the first probability is determined based on the kinematic data of the target object; the second probability is determined based on the plantar pressure data of the target object.

[0136] The fusion module 320 is used to fuse the first probability and the second probability to determine the third probability;

[0137] The detection module 330 is used to detect whether the target object has lower limb motor dysfunction based on the third probability.

[0138] It should be noted that the data fusion-based lower limb motor dysfunction detection device provided in this embodiment of the invention can realize all the method steps implemented in the above-mentioned data fusion-based lower limb motor dysfunction detection method embodiment, and can achieve the same technical effect. Here, the parts that are the same as those in the method embodiment and the beneficial effects will not be described in detail.

[0139] Optionally, the kinematic data includes hip joint angle data;

[0140] The determining module 310 is specifically used for:

[0141] Collect hip joint angle data of the left and right legs of the target object during the gait cycle;

[0142] Determine the maximum and minimum angles from the hip joint angle data;

[0143] The first feature is determined based on the mean, standard deviation, and coefficient of variation of the maximum angle, and the mean, standard deviation, and coefficient of variation of the minimum angle;

[0144] Based on the first feature, the first probability is determined.

[0145] Optionally, the determining module 310 is specifically used for:

[0146] The motion obstacle detection model is trained based on sample kinematic data to determine the first recognition model; the motion obstacle detection model is built based on support vector machine.

[0147] The first feature is input into the first recognition model to obtain the first probability.

[0148] Optionally, the determining module 310 is specifically used for:

[0149] Collect plantar pressure data of the left and right legs of the target object during the gait cycle;

[0150] The maximum pressure value and target duration are determined based on the plantar pressure data; the target duration is the duration corresponding to the pressure value in the plantar pressure data being greater than the preset pressure value within the gait cycle.

[0151] The second feature is determined based on the average, standard deviation, and coefficient of variation of the maximum pressure value, and the average, standard deviation, and coefficient of variation of the target duration;

[0152] Based on the second feature, the second probability is determined.

[0153] Optionally, the determining module 310 is specifically used for:

[0154] The movement obstacle detection model is trained based on sample plantar pressure data to determine the second recognition model; the movement obstacle detection model is built based on support vector machine.

[0155] The second feature is input into the second recognition model to obtain the second probability.

[0156] Optionally, the fusion module 320 is specifically used for:

[0157] The decision fusion model was validated based on sample kinematic data and sample plantar pressure data to determine the third identification model; the decision fusion model was constructed based on the decision fusion algorithm.

[0158] The first probability and the second probability are input into the third recognition model to obtain the third probability.

[0159] Figure 4 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 4As shown, the electronic device may include a processor 410, a communications interface 420, a memory 430, and a communication bus 440, wherein the processor 410, communications interface 420, and memory 430 communicate with each other via the communication bus 440. The processor 410 can call logical instructions in the memory 430 to execute a data fusion-based method for detecting lower limb motor dysfunction. This method includes: determining a first probability and a second probability that a target object has lower limb motor dysfunction; the first probability is determined based on the target object's kinematic data; the second probability is determined based on the target object's plantar pressure data; fusing the first probability and the second probability to determine a third probability; and detecting whether the target object has lower limb motor dysfunction based on the third probability.

[0160] Furthermore, the logical instructions in the aforementioned memory 430 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0161] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the data fusion-based method for detecting lower limb motor dysfunction provided by the above methods. The method includes: determining a first probability and a second probability that a target object has lower limb motor dysfunction; the first probability is determined based on the kinematic data of the target object; the second probability is determined based on the plantar pressure data of the target object; fusing the first probability and the second probability to determine a third probability; and detecting whether the target object has lower limb motor dysfunction based on the third probability.

[0162] In another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the data fusion-based method for detecting lower limb motor dysfunction provided by the methods described above. The method includes: determining a first probability and a second probability that a target object has lower limb motor dysfunction; the first probability being determined based on the kinematic data of the target object; the second probability being determined based on the plantar pressure data of the target object; fusing the first probability and the second probability to determine a third probability; and detecting whether the target object has lower limb motor dysfunction based on the third probability.

[0163] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units 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. Those skilled in the art can understand and implement this without any creative effort.

[0164] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0165] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for detecting lower limb motor dysfunction based on data fusion, applied to electronic devices, characterized in that, include: Determine the first and second probabilities that the target subject has lower limb motor dysfunction; The first probability is determined based on the kinematic data of the target object; The second probability is determined based on the plantar pressure data of the target object; The process of fusing the first probability and the second probability to determine the third probability includes: constructing a decision fusion model based on a weighted average rule of probabilities; iterating through the weights of the first probability and the second probability with a preset step size of 0.001, using a validation set to determine the weight with the highest recognition accuracy as the optimal model parameter, and determining the third recognition model, wherein the sum of the weights of the first probability and the second probability is 1; and inputting the first probability and the second probability into the third recognition model to obtain the third probability. Based on the third probability, detect whether the target object has lower limb motor dysfunction; The kinematic data includes hip joint angle data; the determination of the first probability that the target object has lower limb motor dysfunction includes: Collect hip joint angle data of the left and right legs of the target object during the gait cycle; Determine the maximum and minimum angles from the hip joint angle data; The first feature is determined based on the mean, standard deviation, and coefficient of variation of the maximum angle, and the mean, standard deviation, and coefficient of variation of the minimum angle; Based on the first feature, the first probability is determined; The second probability of determining that the target object has lower limb motor dysfunction includes: Collect plantar pressure data of the left and right legs of the target object during the gait cycle; The maximum pressure value and target duration are determined based on the plantar pressure data; the target duration is the duration during which the pressure value in the plantar pressure data is greater than a preset pressure value within the gait cycle, and the preset pressure value is 20% of the maximum pressure value; The second feature is determined based on the average, standard deviation, and coefficient of variation of the maximum pressure value, and the average, standard deviation, and coefficient of variation of the target duration; Based on the second feature, the second probability is determined.

2. The method for detecting lower limb motor dysfunction based on data fusion according to claim 1, characterized in that, Determining the first probability based on the first feature includes: The motion obstacle detection model is trained based on sample kinematic data to determine the first recognition model; the motion obstacle detection model is built based on support vector machine. The first feature is input into the first recognition model to obtain the first probability.

3. The method for detecting lower limb motor dysfunction based on data fusion according to claim 1, characterized in that, Determining the second probability based on the second feature includes: The movement obstacle detection model is trained based on sample plantar pressure data to determine the second recognition model; the movement obstacle detection model is built based on support vector machine. The second feature is input into the second recognition model to obtain the second probability.

4. A device for detecting lower limb motor dysfunction based on data fusion, characterized in that, include: The determination module is used to determine the first probability and the second probability that the target object has lower limb motor dysfunction; The first probability is determined based on the kinematic data of the target object; the second probability is determined based on the plantar pressure data of the target object. The fusion module is used to fuse the first probability and the second probability to determine a third probability, including: constructing a decision fusion model based on a weighted average rule of probabilities; traversing the weights of the first probability and the second probability with a preset step size of 0.001, using a validation set to determine the weight with the highest recognition accuracy as the optimal model parameter, and determining a third recognition model, wherein the sum of the weights of the first probability and the second probability is 1; and inputting the first probability and the second probability into the third recognition model to obtain the third probability. The detection module is used to detect whether the target object has lower limb motor dysfunction based on the third probability; The kinematic data includes hip joint angle data; the determining module is used for: Collect hip joint angle data of the left and right legs of the target object during the gait cycle; Determine the maximum and minimum angles from the hip joint angle data; The first feature is determined based on the mean, standard deviation, and coefficient of variation of the maximum angle, and the mean, standard deviation, and coefficient of variation of the minimum angle; Based on the first feature, the first probability is determined; The determining module is used for: Collect plantar pressure data of the left and right legs of the target object during the gait cycle; The maximum pressure value and target duration are determined based on the plantar pressure data; the target duration is the duration during which the pressure value in the plantar pressure data is greater than a preset pressure value within the gait cycle, and the preset pressure value is 20% of the maximum pressure value; The second feature is determined based on the average, standard deviation, and coefficient of variation of the maximum pressure value, and the average, standard deviation, and coefficient of variation of the target duration; Based on the second feature, the second probability is determined.

5. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the program, it implements the lower limb motor dysfunction detection method based on data fusion as described in any one of claims 1 to 3.

6. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the data fusion-based method for detecting lower limb motor dysfunction as described in any one of claims 1 to 3.

7. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the data fusion-based method for detecting lower limb motor dysfunction as described in any one of claims 1 to 3.

Images