Gait classification method, apparatus and electronic device

Abstract

The present application provides a gait classification method, device and electronic equipment, comprising: obtaining a gait silhouette sequence to be classified; processing each silhouette image in the gait silhouette sequence to obtain data; inputting the gait silhouette sequence and the data into a preset 3D convolutional neural network to output gait sequence features of the gait silhouette sequence; the gait sequence features carry time sequence information corresponding to the gait silhouette sequence; determining a gait classification result of the gait silhouette sequence according to the gait sequence features and the time sequence information corresponding to the gait sequence features. The method classifies the gait silhouette features carrying the time sequence information, thereby increasing the accuracy of gait classification.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of gait classification, and in particular to a gait classification method, device and electronic equipment. BACKGROUND

[0002] The existing gait classification method generally classifies based on GEI. Specifically, first, a gait sequence is added and averaged to compress into an energy graph, and then a gait classification is performed based on the energy graph through a preset algorithm. The existing gait classification method cannot capture the spatiotemporal features of each part in the classification process, resulting in low accuracy of gait classification. SUMMARY

[0003] The purpose of the present application is to provide a gait classification method, device and electronic equipment to improve the accuracy of gait classification.

[0004] In a first aspect, the embodiments of the present application provide a gait classification method, comprising: obtaining a gait silhouette sequence to be classified; performing strip processing on each silhouette image in the gait silhouette sequence to obtain strip data; inputting the gait silhouette sequence and the strip data into a 3D convolutional neural network to output a gait sequence feature of the gait silhouette sequence; the gait sequence feature carries time sequence information corresponding to the gait silhouette sequence; the 3D convolutional neural network takes the time dimension as the third dimension, extracts time features and spatial features from the input gait silhouette sequence through a 3D convolution kernel, and models in the time-space dimension to obtain motion information of the gait silhouette sequence; and determining a gait classification result of the gait silhouette sequence according to the gait sequence feature and the time sequence information corresponding to the gait sequence feature.

[0005] In combination with the first aspect, the embodiments of the present application provide a first possible implementation manner of the first aspect, wherein the step of performing strip processing on each silhouette image in the gait silhouette sequence to obtain strip data comprises: cutting each silhouette image in the gait silhouette sequence based on a preset parameter to obtain strip data.

[0006] With the first possible implementation manner of the first aspect, the second possible implementation manner of the first aspect is provided. In the second possible implementation manner, the step of inputting the gait silhouette sequence and the striping data into the preset 3D convolutional neural network and outputting the gait sequence feature of the gait silhouette sequence includes: inputting each silhouette image in the gait silhouette sequence into the preset 3D convolutional neural network and outputting a global feature of the gait silhouette sequence, and inputting the striping data into the 3D convolutional neural network and outputting a local feature of the gait silhouette sequence; determining the gait sequence feature of the gait silhouette sequence according to the global feature and the local feature, and outputting the gait sequence feature of the gait silhouette sequence.

[0007] With the first aspect, the third possible implementation manner of the first aspect is provided. After the step of determining the gait sequence feature of the gait silhouette sequence according to the gait silhouette sequence and the striping data, the method includes: reconstructing the gait sequence feature based on a preset attention module to output a gait sequence feature reaching a preset performance threshold; and the step of determining the gait classification result of the gait silhouette sequence according to the gait sequence feature and the time sequence information corresponding to the gait sequence feature includes: determining the gait classification result of the gait silhouette sequence according to the gait sequence feature reaching the preset performance threshold and the time sequence information corresponding to the gait sequence feature.

[0008] With the third possible implementation manner of the first aspect, the fourth possible implementation manner of the first aspect is provided. The step of reconstructing the gait sequence feature based on the preset attention module to output the gait sequence feature reaching the preset performance threshold includes: calculating the correlation of the gait sequence feature in the time sequence information based on the preset attention module; determining the gait sequence feature reaching the preset performance threshold according to the correlation, and outputting the gait sequence feature reaching the preset performance threshold.

[0009] With the fourth possible implementation manner of the first aspect, the fifth possible implementation manner of the first aspect is provided. The step of determining the gait sequence feature reaching the preset performance threshold according to the correlation includes: calculating the weighted sum of the gait sequence feature according to the correlation; and determining the gait sequence feature reaching the preset performance threshold according to the weighted sum.

[0010] With the fourth possible implementation manner of the first aspect, the sixth possible implementation manner of the first aspect is provided, and the step of determining the gait classification result of the gait silhouette sequence according to the gait sequence feature reaching the preset performance threshold and the time sequence information corresponding to the gait sequence feature comprises: performing pooling on the gait sequence feature reaching the preset performance threshold according to the time sequence information corresponding to the gait sequence feature to obtain an aggregated time sequence feature; and determining the gait classification result of the gait silhouette sequence according to the time sequence feature.

[0011] With the sixth possible implementation manner of the first aspect, the seventh possible implementation manner of the first aspect is provided, and the step of determining the gait classification result of the gait silhouette sequence according to the time sequence feature comprises: fusing the time sequence feature to obtain a multi-scale feature; and determining the gait classification result of the gait silhouette sequence according to the multi-scale feature.

[0012] In the second aspect, the embodiment of the present application provides a gait classification device, which comprises: a data acquisition module configured to acquire a gait silhouette sequence to be classified; a data processing module configured to perform strip processing on each silhouette image in the gait silhouette sequence to obtain strip data; a feature determination module configured to input the gait silhouette sequence and the strip data into a preset 3D convolutional neural network to output a gait sequence feature of the gait silhouette sequence; the gait sequence feature carries time sequence information corresponding to the gait silhouette sequence; the 3D convolutional neural network takes a time dimension as a third dimension, and is configured to extract time features and spatial features from the input gait silhouette sequence through a 3D convolution kernel to model the motion information of the gait silhouette sequence in a time-space dimension; and a gait classification result output module configured to determine a gait classification result of the gait silhouette sequence according to the gait sequence feature and the time sequence information corresponding to the gait sequence feature.

[0013] In the third aspect, the embodiment of the present application provides an electronic device, which comprises a processor and a memory, the memory stores machine executable instructions capable of being executed by the processor, and the processor executes the machine executable instructions to implement the gait classification method of any one of the first aspect to the seventh possible implementation manner of the first aspect.

[0014] The embodiment of the present application brings the following beneficial effects:

[0015] The embodiment of the present application provides a gait classification method, device and electronic equipment, which comprises the following steps: obtaining a gait silhouette sequence to be classified; performing strip processing on each silhouette image in the gait silhouette sequence to obtain strip data; inputting the gait silhouette sequence and the strip data into a preset 3D convolutional neural network to output gait sequence features of the gait silhouette sequence; the gait sequence features carry time sequence information corresponding to the gait silhouette sequence; the 3D convolutional neural network takes a time dimension as a third dimension, is used for extracting time features and space features from the input gait silhouette sequence through a 3D convolution kernel, and models motion information of the gait silhouette sequence in a time-space dimension; and determining a gait classification result of the gait silhouette sequence according to the gait sequence features and the time sequence information corresponding to the gait sequence features. The method classifies the gait silhouette features carrying the time sequence information, and the accuracy of gait classification is improved.

[0016] Other features and advantages of the embodiments disclosed in the present application will be described in the following description, or can be known or determined by those skilled in the art from the description, or can be known or determined by those skilled in the art from the description without any creative effort.

[0017] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the following preferred embodiments are described in detail below, and the accompanying drawings are described as follows. BRIEF DESCRIPTION OF DRAWINGS

[0018] In order to more clearly illustrate the specific embodiments of the present application or the technical solutions in the prior art, the following will briefly introduce the drawings needed to be used in the specific embodiments or the prior art description. Obviously, the drawings in the following description are some embodiments of the present application, and those skilled in the art can obtain other drawings according to these drawings without any creative effort.

[0019] Figure 1 A flowchart of a gait classification method provided by the embodiment of the present application is shown in the figure;

[0020] Figure 2 A flowchart of another gait classification method provided by the embodiment of the present application is shown in the figure;

[0021] Figure 3 A structure diagram of a gait classification device provided by the embodiment of the present application is shown in the figure;

[0022] Figure 4 A structure diagram of an electronic device provided by the embodiment of the present application is shown in the figure.

[0023] Icon: 31 - data acquisition module; 32 - data processing module; 33 - feature determination module; 34 - gait classification result output module; 41 - memory; 42 - processor; 43 - bus; 44 - communication interface. DETAILED DESCRIPTION

[0024] To make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are some but not all of the embodiments of the present application. The components of the embodiments of the present application described and shown in the drawings herein can be arranged and designed in various different configurations.

[0025] The existing gait classification method generally classifies based on GEI. Specifically, first, a gait sequence is added and averaged to compress into an energy graph, and then the gait classification is performed based on the energy graph through a preset algorithm. The existing gait classification method cannot capture the spatiotemporal features of each part in the classification process, resulting in a low accuracy of gait classification.

[0026] Based on this, the embodiments of the present application provide a gait classification method, device and electronic equipment. The technology classifies the gait silhouette features carrying time sequence information, thereby increasing the accuracy of gait classification. To facilitate the understanding of the embodiments of the present application, first, a gait classification method disclosed by the embodiments of the present application is described in detail.

[0027] Embodiment 1

[0028] The embodiments of the present application provide a gait classification method. Figure 1 A flowchart of a gait classification method provided by the embodiments of the present application is shown.

[0029] As seen, the method comprises: Figure 1

[0030] Step S101: acquiring a gait silhouette sequence to be classified.

[0031] Step S102: performing strip processing on each silhouette graph in the gait silhouette sequence to obtain strip data.

[0032] In actual operation, the step S102 comprises: based on a preset parameter, cutting each silhouette graph in the gait silhouette sequence to obtain strip data.

[0033] ​Step S103: input the gait silhouette sequence and the strip data into a preset 3D convolutional neural network, and output gait sequence features of the gait silhouette sequence; the gait sequence features carry time sequence information corresponding to the gait silhouette sequence; the 3D convolutional neural network takes the time dimension as the third dimension, and is used for extracting time features and space features from the input gait silhouette sequence by a 3D convolution kernel, so as to model the motion information of the gait silhouette sequence in the time-space dimension.

[0034] In the embodiment, the gait silhouette sequence and the strip data are input into the preset 3D convolutional neural network, the convolutional network weight of the 3D convolutional neural network is shared, and the gait sequence features carrying the time sequence information corresponding to the gait silhouette sequence are output.

[0035] Here, the method utilizes the 3D convolution to better model the time sequence feature information in the sequence, and better utilizes the time sequence information of the gait sequence in the training, so as to establish the feature correlation in the time dimension.

[0036] Step S104: determine a gait classification result of the gait silhouette sequence according to the gait sequence features and the time sequence information corresponding to the gait sequence features.

[0037] The embodiment of the application provides a gait classification method, which comprises the following steps: obtaining a gait silhouette sequence to be classified; performing strip processing on each silhouette image in the gait silhouette sequence to obtain strip data; inputting the gait silhouette sequence and the strip data into a preset 3D convolutional neural network to output gait sequence features of the gait silhouette sequence; the gait sequence features carry time sequence information corresponding to the gait silhouette sequence; the 3D convolutional neural network takes the time dimension as the third dimension, and is used for extracting time features and space features from the input gait silhouette sequence by a 3D convolution kernel, so as to model the motion information of the gait silhouette sequence in the time-space dimension; and determining a gait classification result of the gait silhouette sequence according to the gait sequence features and the time sequence information corresponding to the gait sequence features. The method classifies the gait silhouette features carrying the time sequence information, and the accuracy of gait classification is improved.

[0038] Embodiment 2

[0039] On the basis of the gait classification method provided in Embodiment 1, Embodiment 2 provides another gait classification method. Figure 2 The flowchart of another gait classification method provided in the embodiment of the application is shown in FIG. 2.

[0040] As shown in FIG. 2, the method comprises the following steps: Figure 2

[0041] ​Step S201: Obtain a gait silhouette sequence to be classified.

[0042] Step S202: Perform strip processing on each silhouette image in the gait silhouette sequence to obtain strip data.

[0043] Step S203: Input each silhouette image in the gait silhouette sequence into a preset 3D convolutional neural network, output global features of the gait silhouette sequence, and input the strip data into the 3D convolutional neural network, output local features of the gait silhouette sequence.

[0044] Step S204: Determine gait sequence features of the gait silhouette sequence according to the global features and the local features, and output the gait sequence features of the gait silhouette sequence.

[0045] Step S205: Determine a gait classification result of the gait silhouette sequence according to the gait sequence features and the timing information corresponding to the gait sequence features.

[0046] In one possible implementation, after step S204, the method includes the following step S2041: reconstructing the gait sequence features based on a preset attention module, and outputting gait sequence features reaching a preset performance threshold. Further, step S205 includes: determining a gait classification result of the gait silhouette sequence according to the gait sequence features reaching the preset performance threshold and the timing information corresponding to the gait sequence features reaching the preset performance threshold.

[0047] In actual operation, step S2041 includes the following steps A1-A2:

[0048] Step A1: Calculate the correlation of the gait sequence features in the corresponding timing information based on a preset attention module.

[0049] Here, in specific implementation, step A1 needs to calculate the correlation of each frame of gait sequence features with other frames of gait sequence features in the time dimension.

[0050] Step A2: Determine gait sequence features reaching a preset performance threshold according to the correlation, and output the gait sequence features reaching the preset performance threshold.

[0051] Further, step A2 includes: first, calculating a weighted sum of the gait sequence features according to the correlation. Then, determining gait sequence features reaching a preset performance threshold according to the weighted sum.

[0052] Here, the method obtains the gait sequence features with high performance by presetting the performance threshold in the preset attention module, so as to obtain the gait sequence features corresponding to the time sequence information with better classification effect.

[0053] In actual operation, the above step S205 includes the following steps B1-B2:

[0054] Step B1: Pooling the gait sequence features reaching the preset performance threshold according to the time sequence information corresponding to the gait sequence features, to obtain the aggregated time sequence features.

[0055] Step B2: Determine the gait classification result of the gait silhouette sequence according to the time sequence features.

[0056] Here, the above step B2 includes: first, fuse the above time sequence features to obtain multi-scale features. Then, determine the gait sequence features reaching the preset performance threshold according to the multi-scale features. Further, the step of determining the gait sequence features reaching the preset performance threshold according to the multi-scale features includes: first, reduce the dimension of the multi-scale features by standardization processing in the 3D convolutional neural network and passing through the full connection layer of the 3D convolutional neural network, to obtain the reduced dimension features. Then, according to the reduced dimension features, the gait classification result of the gait silhouette sequence can be obtained.

[0057] Here, the step of fusing the above time sequence features to obtain multi-scale features mainly fuses by element-by-element addition or concatenation, to obtain multi-scale features; the method fuses the feature information of each frame in the time dimension by fusing local and global features and attention mechanism, fully extracts the time sequence features and fine-grained spatial features in the gait sequence.

[0058] In actual application, the above gait classification method is mainly used for detecting abnormal gait.

[0059] The embodiment of the present application provides a gait classification method, which comprises the following steps: obtaining a gait silhouette sequence to be classified; performing strip processing on each silhouette image in the gait silhouette sequence to obtain strip data; inputting each silhouette image in the gait silhouette sequence into a preset 3D convolutional neural network to output global features of the gait silhouette sequence, and inputting the strip data into the 3D convolutional neural network to output local features of the gait silhouette sequence; inputting each silhouette image in the gait silhouette sequence into a preset 3D convolutional neural network to output global features of the gait silhouette sequence, and inputting the strip data into the 3D convolutional neural network to output local features of the gait silhouette sequence; determining gait sequence features of the gait silhouette sequence according to the global features and the local features, and outputting the gait sequence features of the gait silhouette sequence; and determining a gait classification result of the gait silhouette sequence according to the gait sequence features and the time sequence information corresponding to the gait sequence features. The method classifies the global features and the local features of the gait silhouette sequence carrying time sequence information, and further increases the accuracy of gait classification.

[0060] Embodiment 3

[0061] On the basis of the above-mentioned embodiment, the present embodiment provides a gait classification device. Figure 3 A structure diagram of a gait classification device provided by the embodiment of the present application.

[0062] As seen from Figure 3 The device comprises:

[0063] A data acquisition module 31 is configured to acquire a gait silhouette sequence to be classified.

[0064] A data processing module 32 is configured to perform strip processing on each silhouette image in the gait silhouette sequence to obtain strip data.

[0065] A feature determination module 33 is configured to input the gait silhouette sequence and the strip data into a preset 3D convolutional neural network to output gait sequence features of the gait silhouette sequence; the gait sequence features carry time sequence information corresponding to the gait silhouette sequence; the 3D convolutional neural network takes a time dimension as a third dimension, and is configured to extract time features and space features from the input gait silhouette sequence through a 3D convolution kernel to model the motion information of the gait silhouette sequence in a time-space dimension.

[0066] A gait classification result output module 34 is configured to determine a gait classification result of the gait silhouette sequence according to the gait sequence features and the time sequence information corresponding to the gait sequence features.

[0067] The data acquisition module 31, the data processing module 32, the feature determination module 33 and the gait classification result output module 34 are sequentially connected.

[0068] In one of the embodiments, the data processing module 32 is further configured to cut each silhouette image in the gait silhouette sequence based on a preset parameter to obtain strip data.

[0069] In one of the embodiments, the feature determination module 33 is further configured to input each silhouette image in the gait silhouette sequence into a preset 3D convolutional neural network, output global features of the gait silhouette sequence, and input the strip data into the 3D convolutional neural network, output local features of the gait silhouette sequence; determine gait sequence features of the gait silhouette sequence according to the global features and the local features, and output the gait sequence features of the gait silhouette sequence.

[0070] In one of the embodiments, the feature determination module 33 is further configured to reconstruct the gait sequence features based on a preset attention module, output gait sequence features reaching a preset performance threshold; and the gait classification result output module 34 is further configured to determine a gait classification result of the gait silhouette sequence according to the gait sequence features reaching the preset performance threshold and the time sequence information corresponding to the gait sequence features.

[0071] In one of the embodiments, the feature determination module 33 is further configured to calculate a correlation of the gait sequence features in the time sequence information based on a preset attention module; determine gait sequence features reaching a preset performance threshold according to the correlation, and output the gait sequence features reaching the preset performance threshold.

[0072] In one of the embodiments, the feature determination module 33 is further configured to calculate a weighted sum of the gait sequence features according to the correlation; and determine gait sequence features reaching a preset performance threshold according to the weighted sum.

[0073] In one of the embodiments, the gait classification result output module 34 is further configured to pool the gait sequence features reaching the preset performance threshold according to the time sequence information corresponding to the gait sequence features to obtain aggregated time sequence features; and determine a gait classification result of the gait silhouette sequence according to the time sequence features.

[0074] In one of the embodiments, the feature determination module 33 is further configured to fuse the time sequence features to obtain multi-scale features; and determine a gait classification result of the gait silhouette sequence according to the multi-scale features.

[0075] The gait classification device provided by the embodiments of the present application has the same technical features as the gait classification method provided by the above embodiments, and can solve the same technical problems and achieve the same technical effects. The skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the device described above can refer to the corresponding process in the foregoing method embodiments, which will not be described here.

[0076] Embodiment 4

[0077] The embodiment provides an electronic device, including a processor and a memory, the memory stores computer executable instructions capable of being executed by the processor, and the processor executes the computer executable instructions to implement the steps of the gait classification method.

[0078] The embodiment provides a computer readable storage medium, wherein the computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement the steps of the gait classification method.

[0079] Referring to Figure 4 The electronic device includes a memory 41 and a processor 42, the memory 41 stores a computer program capable of running on the processor 42, and the processor executes the computer program to implement the steps provided by the above gait classification method.

[0080] As Figure 4 The device further includes a bus 43 and a communication interface 44, the processor 42, the communication interface 44 and the memory 41 are connected through the bus 43; and the processor 42 is configured to execute executable modules stored in the memory 41, such as computer programs.

[0081] The memory 41 can include a high-speed random access memory (RAM) and can also include a non-volatile memory, such as at least one disk memory. The communication between the system network element and at least one other network element is realized through at least one communication interface 44 (which can be wired or wireless), and the Internet, a wide area network, a local area network, a metropolitan area network, etc. can be used.

[0082] The bus 43 can be an ISA bus, a PCI bus or an EISA bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For the convenience of representation, Figure 4 Only one bidirectional arrow is used to represent the bus in the figure, but it does not mean that there is only one bus or only one type of bus.

[0083] The memory 41 is configured to store a program, and the processor 42 is configured to execute the program after receiving an execution instruction. The method performed by the gait classification device according to any of the embodiments of the present application can be applied to the processor 42 or implemented by the processor 42. The processor 42 can be an integrated circuit chip with a processing capability of signals. In the implementation process, each step of the above method can be completed by the integrated logic circuit of hardware in the processor 42 or the instruction in the form of software. The processor 42 can be a general processor, including a central processing unit (CPU), a network processor (NP), etc.; or a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components. Each method, step and logic block disclosed in the embodiments of the present application can be implemented or executed. The general processor can be a microprocessor or any conventional processor. The steps of the method disclosed in combination with the embodiments of the present application can be directly embodied as a hardware decoding processor for execution, or a combination of hardware and software modules in the decoding processor for execution. The software module can be located in a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register or other mature storage medium in the art. The storage medium is located in the memory 41, and the processor 42 reads the information in the memory 41 and combines the hardware to complete the steps of the above method.

[0084] Further, the embodiments of the present application also provide a machine readable storage medium, which stores machine executable instructions. When the machine executable instructions are called and executed by the processor 42, the machine executable instructions cause the processor 42 to implement the above gait classification method.

[0085] The electronic device and the computer readable storage medium provided by the embodiments of the present application have the same technical features, so they can also solve the same technical problems and achieve the same technical effects.

[0086] In addition, in the description of the embodiments of the present application, unless specifically defined and limited, the terms "mounting", "connection", "connecting" should be understood in a broad sense, for example, can be fixedly connected, can also be detachably connected, or integrally connected; can be mechanically connected, can also be electrically connected; can be directly connected, can also be indirectly connected through an intermediate medium, can be internal communication of two elements. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.

[0087] In the description of the present application, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer" and the like indicate the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the devices or elements referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as a limitation on the present application. In addition, the terms "first", "second", "third" are only for the purpose of description, and cannot be understood as indicating or implying relative importance.

Claims

1. A gait classification method, characterized in that, include: Obtain the gait silhouette sequence to be classified; Each silhouette image in the gait silhouette sequence is segmented into strips to obtain striped data; The gait silhouette sequence and the segmented data are input into a 3D convolutional neural network, which outputs the gait sequence features of the gait silhouette sequence. The gait sequence features carry temporal information corresponding to the gait silhouette sequence. The 3D convolutional neural network uses time as the third dimension and extracts temporal and spatial features from the input gait silhouette sequence through 3D convolutional kernels to model the motion information of the gait silhouette sequence in the spatiotemporal dimension. Based on the gait sequence features and the temporal information corresponding to the gait sequence features, the gait classification result of the gait silhouette sequence is determined; The step of inputting the gait silhouette sequence and the striped data into a 3D convolutional neural network and outputting the gait sequence features of the gait silhouette sequence includes: Each silhouette image in the gait silhouette sequence is input into a 3D convolutional neural network to output the global features of the gait silhouette sequence, and the strip data is input into the 3D convolutional neural network to output the local features of the gait silhouette sequence. Based on the global features and the local features, determine the gait sequence features of the gait silhouette sequence, and output the gait sequence features of the gait silhouette sequence; After determining the gait sequence features of the gait silhouette sequence based on the gait silhouette sequence and the segmented data, the method includes: The gait sequence features are reconstructed based on a preset attention module, and the gait sequence features that reach a preset performance threshold are output. The step of determining the gait classification result of the gait silhouette sequence based on the gait sequence features and the corresponding temporal information includes: Based on the gait sequence features that reach the preset performance threshold and the temporal information corresponding to the gait sequence features, the gait classification result of the gait silhouette sequence is determined.

2. The gait classification method according to claim 1, characterized in that, The step of dividing each silhouette image in the gait silhouette sequence into strips to obtain striped data includes: Based on preset parameters, each silhouette image in the gait silhouette sequence is cut to obtain segmented data.

3. The gait classification method according to claim 1, characterized in that, The steps of reconstructing the gait sequence features based on a preset attention module and outputting gait sequence features that meet a preset performance threshold include: Based on a preset attention module, the correlation between the gait sequence features and the corresponding temporal information is calculated; Based on the correlation, gait sequence features that reach a preset performance threshold are determined and output.

4. The gait classification method according to claim 3, characterized in that, The step of determining gait sequence features that reach a preset performance threshold based on the correlation includes: Based on the correlation, calculate the weighted sum of the gait sequence features; Based on the weighted sum, gait sequence features that reach a preset performance threshold are determined.

5. The gait classification method according to claim 3, characterized in that, The step of determining the gait classification result of the gait silhouette sequence based on the gait sequence features that reach the preset performance threshold and the temporal information corresponding to the gait sequence features includes: The gait sequence features that have reached the preset performance threshold are pooled according to the temporal information corresponding to the gait sequence features to obtain aggregated temporal features; Based on the temporal features, the gait classification result of the gait silhouette sequence is determined.

6. The gait classification method according to claim 5, characterized in that, The step of determining the gait classification result of the gait silhouette sequence based on the temporal features includes: The temporal features are fused to obtain multi-scale features; Based on the multi-scale features, the gait classification result of the gait silhouette sequence is determined.

7. A gait classification device, characterized in that, A method for implementing the gait classification method according to any one of claims 1 to 6 includes: The data acquisition module is used to acquire gait silhouette sequences to be classified. The data processing module is used to divide each silhouette image in the gait silhouette sequence into strips to obtain strip data; The feature determination module is used to input the gait silhouette sequence and the segmented data into a 3D convolutional neural network and output the gait sequence features of the gait silhouette sequence; the gait sequence features carry the temporal information corresponding to the gait silhouette sequence; the 3D convolutional neural network uses the time dimension as the third dimension and extracts temporal and spatial features from the input gait silhouette sequence through 3D convolutional kernels to model the motion information of the gait silhouette sequence in the spatiotemporal dimension; The gait classification result output module is used to determine the gait classification result of the gait silhouette sequence based on the gait sequence features and the temporal information corresponding to the gait sequence features.

8. An electronic device, characterized in that, The electronic device includes a processor and a memory, the memory storing computer-executable instructions that can be executed by the processor, the processor executing the computer-executable instructions to implement the gait classification method according to any one of claims 1 to 6.

Images