CN111368635A - Millimeter wave-based multi-person gait recognition method and device - Google Patents
Millimeter wave-based multi-person gait recognition method and device Download PDFInfo
- Publication number
- CN111368635A CN111368635A CN202010080893.0A CN202010080893A CN111368635A CN 111368635 A CN111368635 A CN 111368635A CN 202010080893 A CN202010080893 A CN 202010080893A CN 111368635 A CN111368635 A CN 111368635A
- Authority
- CN
- China
- Prior art keywords
- point cloud
- frame
- gait
- cloud data
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/20—Movements or behaviour, e.g. gesture recognition
- G06V40/23—Recognition of whole body movements, e.g. for sport training
- G06V40/25—Recognition of walking or running movements, e.g. gait recognition
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Biology (AREA)
- General Health & Medical Sciences (AREA)
- Psychiatry (AREA)
- Social Psychology (AREA)
- Human Computer Interaction (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Abstract
The embodiment of the invention provides a multi-person gait recognition method and device based on millimeter waves, which are characterized in that multi-frame point cloud data obtained by detecting a plurality of target pedestrians by millimeter wave equipment are obtained, each frame of point cloud data is composed of a plurality of reflection point data, and each reflection point data comprises a three-dimensional coordinate, a speed and a confidence coefficient of a reflection point; dividing each frame of point cloud data into target pedestrian number group point cloud subdata based on the position distribution of the reflection points in the point cloud; matching each group of point cloud subdata in each frame of point cloud data with each group of point cloud subdata in other frames of point cloud data, and determining a gait point cloud sequence of each target pedestrian based on a matching result; and respectively inputting the gait point cloud sequence of each target pedestrian into a pre-trained gait recognition neural network to obtain the gait recognition result of each target pedestrian. The gait recognition accuracy is improved on the premise of protecting the privacy of the user.
Description
Technical Field
The invention relates to the technical field of wireless communication, in particular to a millimeter wave-based multi-person gait recognition method and device.
Background
Gait recognition is a new biological feature recognition technology, aims to identify the identity through the walking posture of people, and has the advantages of non-contact remote distance and difficulty in camouflage compared with other biological recognition technologies. Among them, gait refers to the way people walk, and is a complex behavior characteristic that varies from person to person.
The existing gait recognition technology is usually based on computer vision, namely a camera is used for capturing a visual image of a pedestrian walking, and then the gait of the pedestrian is analyzed based on the visual image, so that the identity of the pedestrian is recognized.
However, there are some problems with using a camera for identification. Firstly, the camera captures real images of the daily life of the user, so that serious privacy threats are inevitably generated to the user, and the privacy of the user is not protected. Secondly, the camera is easily affected by illumination, and a clear image cannot be obtained under the conditions of darkness, dim light, shielding and the like, so that the recognition fails.
Therefore, the existing gait recognition method based on computer vision has the technical problems of invasion of user privacy and low recognition accuracy caused by easiness in being influenced by the environment.
Disclosure of Invention
The embodiment of the invention aims to provide a multi-person gait recognition method and device based on millimeter waves, which can improve the gait recognition accuracy on the premise of protecting the privacy of users. The specific technical scheme is as follows:
in order to achieve the above object, an embodiment of the present invention provides a millimeter wave-based multi-person gait recognition method, where the method includes:
acquiring multi-frame point cloud data obtained by detecting a plurality of target pedestrians by millimeter wave equipment, wherein each frame of point cloud data is composed of a plurality of reflection point data, and each reflection point data comprises a three-dimensional coordinate, a speed and a confidence coefficient of a reflection point;
dividing each frame of point cloud data into target pedestrian number group point cloud subdata based on the position distribution of the reflection points in the point cloud;
matching each group of point cloud subdata in each frame of point cloud data with each group of point cloud subdata in other frames of point cloud data, and determining a gait point cloud sequence of each target pedestrian based on a matching result;
and respectively inputting the gait point cloud sequence of each target pedestrian into a pre-trained gait recognition neural network to obtain a gait recognition result of each target pedestrian, wherein the gait recognition neural network is pre-trained according to a training set, and the training set comprises real identity identifications of a plurality of sample pedestrians and the sample gait point cloud sequence of each sample pedestrian.
Optionally, before the dividing each frame of point cloud data into the target pedestrian number group point cloud sub-number data based on the position distribution of the reflection points in the point cloud, the method further includes:
converting the three-dimensional coordinates of each reflection point data in multi-frame point cloud data acquired by detection of each millimeter wave device based on the relative position of each millimeter wave device to acquire the three-dimensional coordinates under the same rectangular coordinate system;
and merging the point cloud data of each frame obtained by the detection of each millimeter wave device based on the time stamp of the point cloud data of each frame.
Optionally, the step of matching each group of point cloud sub-data in each frame of point cloud data with each group of point cloud sub-data in other frames of point cloud data, and determining a gait point cloud sequence of each target pedestrian based on a matching result includes:
matching each group of point cloud subdata in each frame of point cloud data by adopting a Hungarian algorithm, and determining the corresponding relation between the target pedestrian and the point cloud subdata in each frame of point cloud data based on the matching result;
and determining the point cloud subdata in each frame of point cloud data corresponding to the target pedestrian as a gait point cloud sequence of the target pedestrian.
Optionally, the gait recognition neural network is trained according to the following steps:
acquiring a preset neural network model and a preset training set;
inputting the sample gait point cloud sequence into the neural network model to obtain an identity recognition result of a sample pedestrian;
determining a loss value based on the identity recognition result and the real identity;
determining whether the neural network model converges based on the loss value;
if not, adjusting parameter values in the neural network model, and returning to the step of inputting the sample gait point cloud sequence into the neural network model to obtain the identity recognition result of the sample pedestrian;
and if so, determining the current neural network model as the gait recognition neural network.
In order to achieve the above object, an embodiment of the present invention further provides a multi-person gait recognition device based on millimeter waves, where the device includes:
the acquisition module is used for acquiring multi-frame point cloud data obtained by detecting a plurality of target pedestrians by the millimeter wave equipment, each frame of point cloud data is composed of a plurality of reflection point data, and each reflection point data comprises a three-dimensional coordinate, a speed and a confidence coefficient of a reflection point;
the dividing module is used for dividing each frame of point cloud data into target pedestrian number group point cloud subdata based on the position distribution of the reflection points in the point cloud;
the matching module is used for matching each group of point cloud data in each frame of point cloud data with each group of point cloud data in other frames of point cloud data, and determining a gait point cloud sequence of each target pedestrian based on a matching result;
the gait recognition neural network is trained in advance according to a training set, and the training set comprises real identification marks of a plurality of sample pedestrians and a sample gait point cloud sequence of each sample pedestrian.
Optionally, the number of the millimeter wave devices is multiple, and the apparatus further includes:
the conversion module is used for converting the three-dimensional coordinates of each reflection point data in multi-frame point cloud data acquired by detection of each millimeter wave device based on the relative position of each millimeter wave device to acquire the three-dimensional coordinates under the same rectangular coordinate system;
and the merging module is used for merging each frame of point cloud data obtained by detection of each millimeter wave device based on the timestamp of each frame of point cloud data.
Optionally, the matching module is specifically configured to:
matching each group of point cloud subdata in each frame of point cloud data by adopting a Hungarian algorithm, and determining the corresponding relation between the target pedestrian and the point cloud subdata in each frame of point cloud data based on the matching result;
and determining the point cloud subdata in each frame of point cloud data corresponding to the target pedestrian as a gait point cloud sequence of the target pedestrian.
Optionally, the apparatus further includes a training module, where the training module is configured to train the gait recognition neural network, and the training module is specifically configured to:
acquiring a preset neural network model and a preset training set;
inputting the sample gait point cloud sequence into the neural network model to obtain an identity recognition result of a sample pedestrian;
determining a loss value based on the identity recognition result and the real identity;
determining whether the neural network model converges based on the loss value;
if not, adjusting parameter values in the neural network model, and returning to the step of inputting the sample gait point cloud sequence into the neural network model to obtain the identity recognition result of the sample pedestrian;
and if so, determining the current neural network model as the gait recognition neural network.
In order to achieve the above object, an embodiment of the present invention further provides an electronic device, including a processor, a communication interface, a memory, and a communication bus; the processor, the communication interface and the memory complete mutual communication through a communication bus;
a memory for storing a computer program;
and the processor is used for realizing any method step when executing the program stored in the memory.
To achieve the above object, an embodiment of the present invention further provides a computer-readable storage medium, in which a computer program is stored, and the computer program, when executed by a processor, implements any of the above method steps.
By applying the millimeter wave-based multi-person gait recognition method and device provided by the embodiment of the invention, multi-frame point cloud data obtained by detecting a plurality of target pedestrians by a millimeter wave device can be obtained, each frame of point cloud data is composed of a plurality of reflection point data, and each reflection point data comprises a three-dimensional coordinate, a speed and a confidence coefficient of a reflection point; dividing each frame of point cloud data into target pedestrian number group point cloud subdata based on the position distribution of the reflection points in the point cloud; matching each group of point cloud subdata in each frame of point cloud data with each group of point cloud subdata in other frames of point cloud data, and determining a gait point cloud sequence of each target pedestrian based on a matching result; and respectively inputting the gait point cloud sequence of each target pedestrian into a pre-trained gait recognition neural network to obtain a gait recognition result of each target pedestrian, wherein the gait recognition neural network is pre-trained according to a training set, and the training set comprises the identity identifications of a plurality of sample pedestrians and the sample gait point cloud sequence of each sample pedestrian. Therefore, the millimeter wave equipment is adopted to detect the pedestrians, only point cloud information needs to be acquired, and images do not need to be acquired, so that the privacy of users is effectively protected, and the method is not influenced by factors such as illumination and shielding.
Of course, not all of the advantages described above need to be achieved at the same time in the practice of any one product or method of the invention.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
Fig. 1 is a schematic flow chart of a millimeter wave-based multi-person gait recognition method according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a gait recognition neural network according to an embodiment of the present invention;
FIG. 3 is a schematic flow chart of a training gait recognition neural network according to an embodiment of the invention;
fig. 4 is another schematic flow chart of a millimeter wave-based multi-person gait recognition method according to an embodiment of the present invention;
fig. 5 is a schematic structural diagram of a millimeter wave-based multi-person gait recognition apparatus according to an embodiment of the present invention;
fig. 6 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
In order to solve the technical problems that the existing gait recognition based on computer vision violates user privacy and is easily influenced by environment, so that the recognition accuracy is low, the embodiment of the invention provides a millimeter wave-based multi-person gait recognition method and device, electronic equipment and a computer-readable storage medium.
For ease of understanding, the following description will first describe an application scenario of the embodiment of the present invention.
Because the gait is a behavior characteristic which is different from person to person, the embodiment of the invention can determine the identity of the user by identifying the gait of the user, and can be applied to the fields of health detection, novel human-computer interaction and the like. For example, in the field of smart home, the smart home system can recognize the identity of the current owner through gait, and then can adjust the home environment according to the habit of the current owner.
Referring to fig. 1, fig. 1 is a schematic flow chart of a millimeter wave-based multi-person gait recognition method according to an embodiment of the present invention, where the method may include the following steps:
s101: acquiring multi-frame point cloud data obtained by detecting a plurality of target pedestrians by millimeter wave equipment, wherein each frame of point cloud data is composed of a plurality of reflection point data, and each reflection point data comprises a three-dimensional coordinate, a speed and a confidence coefficient of a reflection point.
In the embodiment of the invention, the target pedestrian is detected by adopting millimeter wave equipment. Specifically, the millimeter wave device transmits a millimeter wave signal, the millimeter wave signal is reflected after contacting a target pedestrian, and the millimeter wave device receives the reflected millimeter wave. The millimeter wave device can generate multi-frame point cloud data according to the reflected millimeter wave signals, wherein a certain time interval exists between two adjacent frames of point cloud data, and the specific time interval can be set according to actual requirements, for example, the time interval of the two adjacent frames of point cloud data can be 0.1 second by adjusting parameters of the millimeter wave device. Each frame of point cloud data is composed of a plurality of reflection point data, each reflection point data comprises a three-dimensional coordinate, a speed and a confidence coefficient of a reflection point, and the confidence coefficient can be represented by a signal-to-noise ratio of the reflection point.
As an example, a millimeter wave device with a model of TI IWR1443 or TI IWR6843 may be used, and the millimeter wave device has 3 transmitting antennas and 4 receiving antennas, and is capable of acquiring a three-dimensional point cloud, where the point cloud includes a plurality of points, and each point includes 5 pieces of attribute information, namely x-axis coordinates, y-axis coordinates, z-axis coordinates, speed, and signal-to-noise ratio.
In the embodiment of the invention, the millimeter wave equipment can be connected with the electronic equipment, the millimeter wave equipment sends the multi-frame point cloud data obtained by detecting the target pedestrian to the electronic equipment, and then the electronic equipment can analyze the point cloud data to complete the gait recognition of the target pedestrian.
S102: and dividing each frame of point cloud data into a target pedestrian number group of point cloud subdata based on the position distribution of the reflection points in the point cloud.
In the embodiment of the invention, the millimeter wave device detects a plurality of target pedestrians and subsequently performs gait recognition on a single target pedestrian, so that point cloud data needs to be divided. Specifically, the division may be performed based on the position distribution of the reflection points in the point cloud, and each frame of point cloud data is divided into a plurality of groups of point cloud sub-data, where the number of the divided groups is the same as the number of target pedestrians. For example, if the number of target rows is 3, each frame of point cloud data is divided into 3 groups of point cloud sub-data.
In the embodiment of the invention, a clustering algorithm can be adopted to divide each frame of point cloud data. As an example, a DBSCAN (sensitivity-Based Spatial Clustering of Applications with Noise) Clustering algorithm is used to remove Noise points in the point cloud and then divide the points in the frame into different groups, each group representing a target pedestrian.
S103: and matching each group of point cloud subdata in each frame of point cloud data with each group of point cloud subdata in other frames of point cloud data, and determining the gait point cloud sequence of each target pedestrian based on the matching result.
In the embodiment of the invention, after each frame of point cloud data is divided into a plurality of groups of point cloud subdata, each group of point cloud subdata in each frame of point cloud data can be matched with each group of point cloud subdata in other frames of point cloud data, and a gait point cloud sequence of each target pedestrian is determined based on a matching result.
For ease of understanding, the following examples are given. Suppose that 3 frames of point cloud data are currently acquired and are respectively marked as the 1 st frame of point cloud data, the 2 nd frame of point cloud data and the 3 rd frame of point cloud data, two target pedestrians are respectively marked as a pedestrian A and a pedestrian B, and each frame of point cloud data is divided into two groups of point cloud subdata. In order to identify the gait of the pedestrian A, the point cloud sub-data corresponding to the pedestrian A in the 1 st frame of point cloud data, the point cloud sub-data corresponding to the pedestrian A in the 2 nd frame of point cloud data and the point cloud sub-data corresponding to the pedestrian A in the 3 rd frame of point cloud data need to be obtained, and similarly, when identifying the gait of the pedestrian B, the point cloud sub-data corresponding to the pedestrian B in the 1 st frame of point cloud data, the point cloud sub-data corresponding to the pedestrian B in the 2 nd frame of point cloud data and the point cloud sub-data corresponding to the pedestrian B in the 3 rd frame of point cloud data need to be obtained. Therefore, each group of point cloud subdata in each frame of point cloud data is matched to determine the corresponding relation between the target pedestrian and the point cloud subdata in each frame of point cloud data. In the matching process, a bipartite graph matching related algorithm can be adopted.
In the embodiment of the invention, the gait point cloud sequence of each target pedestrian can be determined based on the matching result. In the above example, the point cloud sub-data in the 1 st frame of point cloud data, the point cloud sub-data in the 2 nd frame of point cloud data, and the point cloud sub-data in the 3 rd frame of point cloud data corresponding to the pedestrian a may be determined as the gait point cloud sequence of the pedestrian a.
S104: and respectively inputting the gait point cloud sequence of each target pedestrian into a pre-trained gait recognition neural network to obtain a gait recognition result of each target pedestrian, wherein the gait recognition neural network is pre-trained according to a training set, and the training set comprises the identity identifications of a plurality of sample pedestrians and the sample gait point cloud sequence of each sample pedestrian.
In the embodiment of the invention, the gait point cloud sequence of each target pedestrian can be respectively input into the gait recognition neural network, and because the gait recognition neural network is trained and finished in advance according to the training set, the gait recognition result of the target pedestrian can be output, namely the identity of the target pedestrian can be determined by recognizing the gait of the target pedestrian.
The training set comprises the identity identifications of a plurality of sample pedestrians and a sample gait point cloud sequence of each sample pedestrian.
The network structure and training process of the gait recognition neural network can be seen below.
By applying the millimeter wave-based multi-person gait recognition method provided by the embodiment of the invention, multi-frame point cloud data obtained by detecting a plurality of target pedestrians by a millimeter wave device can be obtained, each frame of point cloud data is composed of a plurality of reflection point data, and each reflection point data comprises a three-dimensional coordinate, a speed and a confidence coefficient of a reflection point; dividing each frame of point cloud data into target pedestrian number group point cloud subdata based on the position distribution of the reflection points in the point cloud; matching each group of point cloud subdata in each frame of point cloud data with each group of point cloud subdata in other frames of point cloud data, and determining a gait point cloud sequence of each target pedestrian based on a matching result; and respectively inputting the gait point cloud sequence of each target pedestrian into a pre-trained gait recognition neural network to obtain a gait recognition result of each target pedestrian, wherein the gait recognition neural network is pre-trained according to a training set, and the training set comprises the identity identifications of a plurality of sample pedestrians and the sample gait point cloud sequence of each sample pedestrian. Therefore, the millimeter wave equipment is adopted to detect the pedestrians, only point cloud information needs to be acquired, and images do not need to be acquired, so that the privacy of users is effectively protected, and the method is not influenced by factors such as illumination and shielding.
In an embodiment of the invention, in order to acquire more point cloud data and further improve the accuracy of gait recognition, a plurality of millimeter wave devices may be used to detect a target pedestrian. Specifically, a plurality of millimeter wave devices can be placed at different positions in advance, target pedestrians are detected respectively, and then point cloud data obtained by detection of the millimeter wave devices are combined.
In an embodiment of the present invention, after S101 and before S102, the following steps may be further included:
step A: and converting the three-dimensional coordinates of each reflection point data in multi-frame point cloud data acquired by detection of each millimeter wave device based on the relative position of each millimeter wave device to acquire the three-dimensional coordinates under the same rectangular coordinate system.
Specifically, because the positions of the millimeter wave devices are different, the rectangular coordinate systems in which the three-dimensional coordinates of the reflection point data obtained by detection are located are also different, and the three-dimensional coordinates of each reflection point data can be converted based on the relative positions of the millimeter wave devices to obtain the three-dimensional coordinates under the same rectangular coordinate system.
The specific coordinate transformation formula may be as follows:
x′＝xcos(θ)-ysin(θ)
y′＝xcos(θ)+ysin(θ)
where θ represents the rotation angle of the coordinate system, (x, y) is the coordinates before conversion of the original coordinate system, and (x ', y') is the coordinates after conversion.
And B: and merging the point cloud data of each frame obtained by the detection of each millimeter wave device based on the time stamp of the point cloud data of each frame.
In the embodiment of the invention, after the three-dimensional coordinates are converted, the point cloud data detected by each millimeter wave device can be combined based on the timestamp of each frame of point cloud data.
Specifically, all the point cloud data can be sorted according to the time stamps of the point cloud data, and the point cloud data obtained by detecting different millimeter wave devices with time differences smaller than a preset threshold value are combined. The preset threshold may be set according to actual requirements, for example, to 50 ms.
The detection result of each millimeter wave device is integrated with each frame of merged point cloud data, so that the number of point clouds can be increased, the influence caused by shielding among pedestrians is effectively reduced, and the gait recognition accuracy is further improved.
In an embodiment of the present invention, the step S103 may specifically include the following steps:
step a: and matching each group of point cloud subdata in each frame of point cloud data by adopting a Hungarian algorithm, and determining the corresponding relation between the target pedestrian and the point cloud subdata in each frame of point cloud data based on the matching result.
The Hungarian algorithm is mainly used for solving the problems related to bipartite graph matching, and the point cloud sub-data in the current frame can be matched with the point cloud sub-data appearing in the previous frame by the Hungarian algorithm. Specifically, matching may be performed by using the position coordinates of the reflection points included in the point cloud sub-data. Because the interval time of two adjacent frames of point cloud data is very small, the position coordinate change of the point cloud subdata in the two adjacent frames of point clouds is very small. If the position coordinates of the group of point cloud sub-data of the current frame and the position coordinates of the group of point cloud sub-data of the previous frame have higher similarity, the group of point cloud sub-data of the current frame and the group of point cloud sub-data of the previous frame can be considered to be matched, namely, the group of point cloud sub-data of the current frame and the group of point cloud sub-data of the previous frame both correspond to the same target pedestrian.
After matching is completed, the corresponding relation between the target pedestrian and the point cloud subdata in each frame of point cloud data can be determined.
Step b: and determining the point cloud subdata in each frame of point cloud data corresponding to the target pedestrian as a gait point cloud sequence of the target pedestrian.
In this step, the point cloud sub-data in each frame of point cloud data corresponding to the target pedestrian can be determined as the gait point cloud sequence of the target pedestrian. For example, the gait point cloud sequence of the pedestrian a can be determined by the point cloud sub-data in the 1 st frame of point cloud data, the point cloud sub-data in the 2 nd frame of point cloud data, and the point cloud sub-data in the 3 rd frame of point cloud data corresponding to the pedestrian a.
In an embodiment of the present invention, a residual neural network resnet that performs well in a picture video may be selected as the deep neural network model to be trained. In addition, the existing residual error neural network can be adjusted according to the characteristics of the point cloud data.
Specifically, each piece of reflection point data in each frame of point cloud data includes 5 attributes including an x-axis coordinate, a y-axis coordinate, a z-axis coordinate, a speed, and a confidence, so that the preset neural network model may include 5 sub-network models having the same structure, and each sub-network model corresponds to one attribute of the reflection point data.
In the embodiment of the present invention, since the number of reflection points in the acquired point cloud may be small, and the number of emission points in each frame of point cloud data may also be different, the point cloud in each frame may be copied before the point cloud data is input to the neural network, so that the number of reflection points included in each frame of point cloud is a preset value, for example, 128.
In addition, the number of frames of the point cloud data included in the gait point cloud sequence may be preset, for example, 30 frames, and the input of each sub-network model is a matrix of 30 × 128.
As an example, referring to fig. 2, fig. 2 is a schematic structural diagram of a gait recognition neural network provided by an embodiment of the present invention, and as shown in fig. 2, a neural network model adopted by the embodiment of the present invention may be a 7-layer structure, and an anterior 5-layer neural network structure is shared, but 5 attributes each perform training of network parameters. That is, the x-axis coordinate, the y-axis coordinate, the z-axis coordinate, the speed and the confidence in the 5 attributes are respectively input into the corresponding network layers. The first layer of the network structure is a convolutional layer with convolutional kernel size 7x7, step size 2. The maximum pooling operation with step size 2 is performed after the convolutional layer. A residual block is then connected, which consists of a 4-tier network, i.e., layers 2-5, each of which can be a convolutional layer with a convolutional kernel of 3 and a step size of 1. And after the layer 3 convolution operation is finished, performing residual operation on the obtained characteristics and the characteristics obtained by the maximum pooling operation. And after the 5 th layer convolution operation is finished, performing residual operation on the obtained features and the features obtained by the 3 rd layer convolution. And after residual operation, performing average pooling operation on the features. And finally, connecting the 5 attribute convolution obtained features together, and extracting the overall features of the point cloud by using a 6 th layer network. And finally, classifying by using a 7 th layer module, namely a full connection layer module, and outputting a final judgment result by using a network model.
In one embodiment of the present invention, referring to fig. 3, the gait recognition neural network can be trained as follows:
s301: and acquiring a preset neural network model and a preset training set.
The preset neural network model may be structured as shown in fig. 2 and the related description. The preset training set may include real ids of a plurality of sample pedestrians, and a sample gait point cloud sequence of each sample pedestrian.
S302: and inputting the sample gait point cloud sequence into a neural network model to obtain an identity recognition result of the sample pedestrian.
As above, the data of 5 attributes in the sample gait point cloud sequence can be output to 5 sub-network models, respectively. And after the operation of the neural network model, outputting the identification result of the pedestrian in each sample.
S303: and determining a loss value based on the identity recognition result and the real identity.
At the beginning of training, the recognition result of the neural network may be different from the real identity, and the loss value of the current iteration may be determined based on the recognition result and the real identity.
In the embodiment of the present invention, the loss value is obtained by using, but not limited to, a cross entropy formula, a Mean Squared Error (MSE) formula, and the like as the loss function.
S304: judging whether the neural network model converges based on the loss value; if not, executing S305; if so, S306 is executed.
S305: and adjusting parameter values in the neural network model, and returning to the step S302.
S306: and determining the current neural network model as the gait recognition neural network.
Specifically, a loss threshold may be preset, and if the loss value is greater than the loss threshold, it is determined that the neural network model is not converged, parameters in the neural network model need to be adjusted, and the step S302 is returned, that is, the next round of iterative training is performed.
If the loss value is within the preset range, the neural network model can be considered to be converged, and the converged neural network model is determined as the gait recognition neural network.
For convenience of understanding, the millimeter wave based multi-person gait recognition method provided by the embodiment of the invention is further described below with reference to fig. 4 of the accompanying drawings, as shown in fig. 4, a plurality of millimeter wave devices detect target pedestrians, and send point cloud data obtained by detection to an electronic device, the electronic device synthesizes point cloud data obtained by detection of each millimeter wave device to obtain synthesized point cloud data, and then performs data segmentation on each frame of synthesized point cloud data to respectively obtain gait point cloud sequences of each target pedestrian. And respectively inputting the gait point cloud sequence of each target pedestrian into the trained neural network model to obtain the gait recognition result of each target pedestrian.
Based on the same inventive concept, according to the above millimeter wave based multi-user gait recognition method embodiment, an embodiment of the present invention further provides a millimeter wave based multi-user gait recognition apparatus, referring to fig. 5, which may include the following modules:
the acquiring module 501 is configured to acquire multi-frame point cloud data obtained by detecting multiple target pedestrians by a millimeter wave device, where each frame of point cloud data is composed of multiple reflection point data, and each reflection point data includes a three-dimensional coordinate, a speed, and a confidence of a reflection point;
a dividing module 502, configured to divide each frame of point cloud data into target pedestrian number group point cloud sub-data based on the position distribution of the reflection points in the point cloud;
the matching module 503 is configured to match each set of point cloud sub-data in each frame of point cloud data with each set of point cloud sub-data in other frames of point cloud data, and determine a gait point cloud sequence of each target pedestrian based on a matching result;
the input module 504 is configured to input the gait point cloud sequence of each target pedestrian into a pre-trained gait recognition neural network to obtain a gait recognition result of each target pedestrian, where the gait recognition neural network is pre-trained according to a training set, and the training set includes real identification of multiple sample pedestrians and a sample gait point cloud sequence of each sample pedestrian.
In an embodiment of the present invention, on the basis of the apparatus shown in fig. 5, the following modules may be further included:
the conversion module is used for converting the three-dimensional coordinates of each reflection point data in multi-frame point cloud data acquired by detection of each millimeter wave device based on the relative position of each millimeter wave device to acquire the three-dimensional coordinates under the same rectangular coordinate system;
and the merging module is used for merging each frame of point cloud data obtained by detection of each millimeter wave device based on the timestamp of each frame of point cloud data.
In an embodiment of the present invention, the matching module 503 may be specifically configured to:
matching each group of point cloud subdata in each frame of point cloud data by adopting a Hungarian algorithm, and determining the corresponding relation between a target pedestrian and the point cloud subdata in each frame of point cloud data based on a matching result;
and determining the point cloud subdata in each frame of point cloud data corresponding to the target pedestrian as a gait point cloud sequence of the target pedestrian.
In an embodiment of the present invention, on the basis of the apparatus shown in fig. 5, the apparatus may further include a training module, where the training model is used to train a gait recognition neural network, and specifically is used to:
acquiring a preset neural network model and a preset training set;
inputting the sample gait point cloud sequence into a neural network model to obtain an identity recognition result of a sample pedestrian;
determining a loss value based on the identity recognition result and the real identity;
judging whether the neural network model converges based on the loss value;
if not, adjusting parameter values in the neural network model, and returning to the step of inputting the sample gait point cloud sequence into the neural network model to obtain the identity recognition result of the sample pedestrian;
and if so, determining the current neural network model as the gait recognition neural network.
By applying the millimeter wave-based multi-person gait recognition device provided by the embodiment of the invention, multi-frame point cloud data obtained by detecting a plurality of target pedestrians by a millimeter wave device can be obtained, each frame of point cloud data is composed of a plurality of reflection point data, and each reflection point data comprises a three-dimensional coordinate, a speed and a confidence coefficient of a reflection point; dividing each frame of point cloud data into target pedestrian number group point cloud subdata based on the position distribution of the reflection points in the point cloud; matching each group of point cloud subdata in each frame of point cloud data with each group of point cloud subdata in other frames of point cloud data, and determining a gait point cloud sequence of each target pedestrian based on a matching result; and respectively inputting the gait point cloud sequence of each target pedestrian into a pre-trained gait recognition neural network to obtain a gait recognition result of each target pedestrian, wherein the gait recognition neural network is pre-trained according to a training set, and the training set comprises the identity identifications of a plurality of sample pedestrians and the sample gait point cloud sequence of each sample pedestrian. Therefore, the millimeter wave equipment is adopted to detect the pedestrians, only point cloud information needs to be acquired, and images do not need to be acquired, so that the privacy of users is effectively protected, and the method is not influenced by factors such as illumination and shielding.
Based on the same inventive concept, according to the above-mentioned embodiment of the millimeter wave-based multi-person gait recognition method, an embodiment of the present invention further provides an electronic device, as shown in fig. 6, comprising a processor 601, a communication interface 602, a memory 603 and a communication bus 604, wherein the processor 601, the communication interface 602 and the memory 603 complete mutual communication via the communication bus 604,
a memory 603 for storing a computer program;
the processor 601 is configured to implement the following steps when executing the program stored in the memory 603:
acquiring multi-frame point cloud data obtained by detecting a plurality of target pedestrians by millimeter wave equipment, wherein each frame of point cloud data is composed of a plurality of reflection point data, and each reflection point data comprises a three-dimensional coordinate, a speed and a confidence coefficient of a reflection point;
dividing each frame of point cloud data into target pedestrian number group point cloud subdata based on the position distribution of the reflection points in the point cloud;
matching each group of point cloud subdata in each frame of point cloud data with each group of point cloud subdata in other frames of point cloud data, and determining a gait point cloud sequence of each target pedestrian based on a matching result;
and respectively inputting the gait point cloud sequence of each target pedestrian into a pre-trained gait recognition neural network to obtain a gait recognition result of each target pedestrian, wherein the gait recognition neural network is pre-trained according to a training set, and the training set comprises real identity identifications of a plurality of sample pedestrians and the sample gait point cloud sequence of each sample pedestrian.
The communication bus mentioned in the electronic device may be a Peripheral Component Interconnect (PCI) bus, an Extended Industry Standard Architecture (EISA) bus, or the like. The communication bus may be divided into an address bus, a data bus, a control bus, etc. For ease of illustration, only one thick line is shown, but this does not mean that there is only one bus or one type of bus.
The communication interface is used for communication between the electronic equipment and other equipment.
The Memory may include a Random Access Memory (RAM) or a Non-Volatile Memory (NVM), such as at least one disk Memory. Optionally, the memory may also be at least one memory device located remotely from the processor.
The Processor may be a general-purpose Processor, including a Central Processing Unit (CPU), a Network Processor (NP), and the like; but may also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other Programmable logic device, discrete Gate or transistor logic device, discrete hardware component.
By applying the electronic equipment provided by the embodiment of the invention, multi-frame point cloud data obtained by detecting a plurality of target pedestrians by millimeter wave equipment can be obtained, each frame of point cloud data is composed of a plurality of reflection point data, and each reflection point data comprises a three-dimensional coordinate, a speed and a confidence coefficient of a reflection point; dividing each frame of point cloud data into target pedestrian number group point cloud subdata based on the position distribution of the reflection points in the point cloud; matching each group of point cloud subdata in each frame of point cloud data with each group of point cloud subdata in other frames of point cloud data, and determining a gait point cloud sequence of each target pedestrian based on a matching result; and respectively inputting the gait point cloud sequence of each target pedestrian into a pre-trained gait recognition neural network to obtain a gait recognition result of each target pedestrian, wherein the gait recognition neural network is pre-trained according to a training set, and the training set comprises the identity identifications of a plurality of sample pedestrians and the sample gait point cloud sequence of each sample pedestrian. Therefore, the millimeter wave equipment is adopted to detect the pedestrians, only point cloud information needs to be acquired, and images do not need to be acquired, so that the privacy of users is effectively protected, and the method is not influenced by factors such as illumination and shielding.
In yet another embodiment of the present invention, a computer-readable storage medium is further provided, in which a computer program is stored, and the computer program, when executed by a processor, implements any of the above steps of the millimeter wave based multi-person gait recognition method.
In yet another embodiment of the present invention, there is also provided a computer program product containing instructions which, when run on a computer, cause the computer to perform any of the above-described millimeter wave based multi-person gait recognition methods.
In the above embodiments, the implementation may be wholly or partially realized by software, hardware, firmware, or any combination thereof. When implemented in software, may be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. The procedures or functions according to the embodiments of the invention are brought about in whole or in part when the computer program instructions are loaded and executed on a computer. The computer may be a general purpose computer, a special purpose computer, a network of computers, or other programmable device. The computer instructions may be stored in a computer readable storage medium or transmitted from one computer readable storage medium to another, for example, the computer instructions may be transmitted from one website, computer, server, or data center to another website, computer, server, or data center by wire (e.g., coaxial cable, fiber optic, Digital Subscriber Line (DSL)) or wirelessly (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device, such as a server, a data center, etc., that incorporates one or more of the available media. The usable medium may be a magnetic medium (e.g., floppy Disk, hard Disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., Solid State Disk (SSD)), among others.
It is noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.
All the embodiments in the present specification are described in a related manner, and the same and similar parts among the embodiments may be referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, for the embodiments of the millimeter wave-based multi-person gait recognition device, the electronic device and the computer-readable storage medium, since they are substantially similar to the embodiment of the millimeter wave-based multi-person gait recognition method, the description is relatively simple, and the relevant points can be found in the partial description of the embodiment of the millimeter wave-based multi-person gait recognition method.
The above description is only for the preferred embodiment of the present invention, and is not intended to limit the scope of the present invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention shall fall within the protection scope of the present invention.
Claims (10)
1. A multi-person gait recognition method based on millimeter waves is characterized by comprising the following steps:
acquiring multi-frame point cloud data obtained by detecting a plurality of target pedestrians by millimeter wave equipment, wherein each frame of point cloud data is composed of a plurality of reflection point data, and each reflection point data comprises a three-dimensional coordinate, a speed and a confidence coefficient of a reflection point;
dividing each frame of point cloud data into target pedestrian number group point cloud subdata based on the position distribution of the reflection points in the point cloud;
matching each group of point cloud subdata in each frame of point cloud data with each group of point cloud subdata in other frames of point cloud data, and determining a gait point cloud sequence of each target pedestrian based on a matching result;
and respectively inputting the gait point cloud sequence of each target pedestrian into a pre-trained gait recognition neural network to obtain a gait recognition result of each target pedestrian, wherein the gait recognition neural network is pre-trained according to a training set, and the training set comprises real identity identifications of a plurality of sample pedestrians and the sample gait point cloud sequence of each sample pedestrian.
2. The method of claim 1, wherein the millimeter wave device is plural, and before the dividing of each frame of point cloud data into the target pedestrian number group point cloud sub-number data based on the position distribution of the reflection points in the point cloud, the method further comprises:
converting the three-dimensional coordinates of each reflection point data in multi-frame point cloud data acquired by detection of each millimeter wave device based on the relative position of each millimeter wave device to acquire the three-dimensional coordinates under the same rectangular coordinate system;
and merging the point cloud data of each frame obtained by the detection of each millimeter wave device based on the time stamp of the point cloud data of each frame.
3. The method of claim 1, wherein the step of matching each set of point cloud data in each frame of point cloud data with each set of point cloud data in other frames of point cloud data and determining a gait point cloud sequence of each target pedestrian based on the matching result comprises:
matching each group of point cloud subdata in each frame of point cloud data by adopting a Hungarian algorithm, and determining the corresponding relation between the target pedestrian and the point cloud subdata in each frame of point cloud data based on the matching result;
and determining the point cloud subdata in each frame of point cloud data corresponding to the target pedestrian as a gait point cloud sequence of the target pedestrian.
4. The method of claim 1, wherein the gait recognition neural network is trained by:
acquiring a preset neural network model and a preset training set;
inputting the sample gait point cloud sequence into the neural network model to obtain an identity recognition result of a sample pedestrian;
determining a loss value based on the identity recognition result and the real identity;
determining whether the neural network model converges based on the loss value;
if not, adjusting parameter values in the neural network model, and returning to the step of inputting the sample gait point cloud sequence into the neural network model to obtain the identity recognition result of the sample pedestrian;
and if so, determining the current neural network model as the gait recognition neural network.
5. A millimeter wave based multi-person gait recognition device, the device comprising:
the acquisition module is used for acquiring multi-frame point cloud data obtained by detecting a plurality of target pedestrians by the millimeter wave equipment, each frame of point cloud data is composed of a plurality of reflection point data, and each reflection point data comprises a three-dimensional coordinate, a speed and a confidence coefficient of a reflection point;
the dividing module is used for dividing each frame of point cloud data into target pedestrian number group point cloud subdata based on the position distribution of the reflection points in the point cloud;
the matching module is used for matching each group of point cloud data in each frame of point cloud data with each group of point cloud data in other frames of point cloud data, and determining a gait point cloud sequence of each target pedestrian based on a matching result;
the gait recognition neural network is trained in advance according to a training set, and the training set comprises real identification marks of a plurality of sample pedestrians and a sample gait point cloud sequence of each sample pedestrian.
6. The apparatus of claim 5, wherein the millimeter wave device is in plurality, the apparatus further comprising:
the conversion module is used for converting the three-dimensional coordinates of each reflection point data in multi-frame point cloud data acquired by detection of each millimeter wave device based on the relative position of each millimeter wave device to acquire the three-dimensional coordinates under the same rectangular coordinate system;
and the merging module is used for merging each frame of point cloud data obtained by detection of each millimeter wave device based on the timestamp of each frame of point cloud data.
7. The apparatus of claim 5, wherein the matching module is specifically configured to:
matching each group of point cloud subdata in each frame of point cloud data by adopting a Hungarian algorithm, and determining the corresponding relation between the target pedestrian and the point cloud subdata in each frame of point cloud data based on the matching result;
and determining the point cloud subdata in each frame of point cloud data corresponding to the target pedestrian as a gait point cloud sequence of the target pedestrian.
8. The apparatus according to claim 5, further comprising a training module, the training model being configured to train the gait recognition neural network, the training module being specifically configured to:
acquiring a preset neural network model and a preset training set;
inputting the sample gait point cloud sequence into the neural network model to obtain an identity recognition result of a sample pedestrian;
determining a loss value based on the identity recognition result and the real identity;
determining whether the neural network model converges based on the loss value;
if not, adjusting parameter values in the neural network model, and returning to the step of inputting the sample gait point cloud sequence into the neural network model to obtain the identity recognition result of the sample pedestrian;
and if so, determining the current neural network model as the gait recognition neural network.
9. An electronic device is characterized by comprising a processor, a communication interface, a memory and a communication bus, wherein the processor and the communication interface are used for realizing mutual communication by the memory through the communication bus;
a memory for storing a computer program;
a processor for implementing the method steps of any of claims 1 to 4 when executing a program stored in the memory.
10. A computer-readable storage medium, characterized in that a computer program is stored in the computer-readable storage medium, which computer program, when being executed by a processor, carries out the method steps of any one of claims 1 to 4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010080893.0A CN111368635B (en) | 2020-02-05 | 2020-02-05 | Millimeter wave-based multi-person gait recognition method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010080893.0A CN111368635B (en) | 2020-02-05 | 2020-02-05 | Millimeter wave-based multi-person gait recognition method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111368635A true CN111368635A (en) | 2020-07-03 |
| CN111368635B CN111368635B (en) | 2021-05-25 |
Family
ID=71208060
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010080893.0A Active CN111368635B (en) | 2020-02-05 | 2020-02-05 | Millimeter wave-based multi-person gait recognition method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111368635B (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111914762A (en) * | 2020-08-04 | 2020-11-10 | 浙江大华技术股份有限公司 | Gait information-based identity recognition method and device |
| CN112666553A (en) * | 2020-12-16 | 2021-04-16 | 动联(山东)电子科技有限公司 | Road ponding identification method and equipment based on millimeter wave radar |
| CN112966780A (en) * | 2021-03-31 | 2021-06-15 | 动联(山东)电子科技有限公司 | Animal behavior identification method and system |
| CN115049039A (en) * | 2021-03-08 | 2022-09-13 | 北京金茂绿建科技有限公司 | State recognition method based on neural network, neural network training method and device |
| CN115661935A (en) * | 2022-10-31 | 2023-01-31 | 海信集团控股股份有限公司 | Method and equipment for determining human body action accuracy |
| WO2023080018A1 (en) * | 2021-11-04 | 2023-05-11 | オムロン株式会社 | Biological information processing device, biological information processing method, and program |
| WO2024088143A1 (en) * | 2022-10-28 | 2024-05-02 | ***股份有限公司 | Passive keyless vehicle entry method and passive keyless vehicle entry system |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105574510A (en) * | 2015-12-18 | 2016-05-11 | 北京邮电大学 | Gait identification method and device |
| US9471989B2 (en) * | 2013-06-03 | 2016-10-18 | University Of Florida Research Foundation, Inc. | Vascular anatomy modeling derived from 3-dimensional medical image processing |
| CN108036793A (en) * | 2017-12-11 | 2018-05-15 | 北京奇虎科技有限公司 | Localization method, device and electronic equipment based on a cloud |
| CN108152831A (en) * | 2017-12-06 | 2018-06-12 | 中国农业大学 | A kind of laser radar obstacle recognition method and system |
| CN108600202A (en) * | 2018-04-11 | 2018-09-28 | Oppo广东移动通信有限公司 | A kind of information processing method and device, computer readable storage medium |
| CN108985171A (en) * | 2018-06-15 | 2018-12-11 | 上海仙途智能科技有限公司 | Estimation method of motion state and state estimation device |
| CN108986450A (en) * | 2018-07-25 | 2018-12-11 | 北京万集科技股份有限公司 | Vehicle environmental cognitive method, terminal and system |
| CN109145677A (en) * | 2017-06-15 | 2019-01-04 | 百度在线网络技术（北京）有限公司 | Obstacle detection method, device, equipment and storage medium |
| CN109459759A (en) * | 2018-11-13 | 2019-03-12 | 中国科学院合肥物质科学研究院 | City Terrain three-dimensional rebuilding method based on quadrotor drone laser radar system |
| CN109800689A (en) * | 2019-01-04 | 2019-05-24 | 西南交通大学 | A kind of method for tracking target based on space-time characteristic fusion study |
| CN110084156A (en) * | 2019-04-12 | 2019-08-02 | 中南大学 | A kind of gait feature abstracting method and pedestrian's personal identification method based on gait feature |
| CN110298376A (en) * | 2019-05-16 | 2019-10-01 | 西安电子科技大学 | A kind of bank money image classification method based on improvement B-CNN |
-
2020
- 2020-02-05 CN CN202010080893.0A patent/CN111368635B/en active Active
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9471989B2 (en) * | 2013-06-03 | 2016-10-18 | University Of Florida Research Foundation, Inc. | Vascular anatomy modeling derived from 3-dimensional medical image processing |
| CN105574510A (en) * | 2015-12-18 | 2016-05-11 | 北京邮电大学 | Gait identification method and device |
| CN109145677A (en) * | 2017-06-15 | 2019-01-04 | 百度在线网络技术（北京）有限公司 | Obstacle detection method, device, equipment and storage medium |
| CN108152831A (en) * | 2017-12-06 | 2018-06-12 | 中国农业大学 | A kind of laser radar obstacle recognition method and system |
| CN108036793A (en) * | 2017-12-11 | 2018-05-15 | 北京奇虎科技有限公司 | Localization method, device and electronic equipment based on a cloud |
| CN108600202A (en) * | 2018-04-11 | 2018-09-28 | Oppo广东移动通信有限公司 | A kind of information processing method and device, computer readable storage medium |
| CN108985171A (en) * | 2018-06-15 | 2018-12-11 | 上海仙途智能科技有限公司 | Estimation method of motion state and state estimation device |
| CN108986450A (en) * | 2018-07-25 | 2018-12-11 | 北京万集科技股份有限公司 | Vehicle environmental cognitive method, terminal and system |
| CN109459759A (en) * | 2018-11-13 | 2019-03-12 | 中国科学院合肥物质科学研究院 | City Terrain three-dimensional rebuilding method based on quadrotor drone laser radar system |
| CN109800689A (en) * | 2019-01-04 | 2019-05-24 | 西南交通大学 | A kind of method for tracking target based on space-time characteristic fusion study |
| CN110084156A (en) * | 2019-04-12 | 2019-08-02 | 中南大学 | A kind of gait feature abstracting method and pedestrian's personal identification method based on gait feature |
| CN110298376A (en) * | 2019-05-16 | 2019-10-01 | 西安电子科技大学 | A kind of bank money image classification method based on improvement B-CNN |
Non-Patent Citations (3)
| Title |
|---|
| PEIJUN ZHAO ET AL.: "mID：Tracking and Identifying People with Millimeter Wave Radar", 《2019 15TH INTERNATIONAL CONFERENCE ON DISTRIBUTED COMPUTING IN SENSOR SYSTEMS (DCOSS)》 * |
| 李炯 等: "基于改进进化匈牙利的多目标跟踪算法研究", 《军事交通学院学报》 * |
| 池凌云: "基于深度学习的多视角步态识别算法研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111914762A (en) * | 2020-08-04 | 2020-11-10 | 浙江大华技术股份有限公司 | Gait information-based identity recognition method and device |
| CN112666553A (en) * | 2020-12-16 | 2021-04-16 | 动联(山东)电子科技有限公司 | Road ponding identification method and equipment based on millimeter wave radar |
| CN115049039A (en) * | 2021-03-08 | 2022-09-13 | 北京金茂绿建科技有限公司 | State recognition method based on neural network, neural network training method and device |
| CN115049039B (en) * | 2021-03-08 | 2023-11-14 | 北京金茂绿建科技有限公司 | Neural network-based state identification method, neural network training method and device |
| CN112966780A (en) * | 2021-03-31 | 2021-06-15 | 动联(山东)电子科技有限公司 | Animal behavior identification method and system |
| WO2023080018A1 (en) * | 2021-11-04 | 2023-05-11 | オムロン株式会社 | Biological information processing device, biological information processing method, and program |
| WO2024088143A1 (en) * | 2022-10-28 | 2024-05-02 | ***股份有限公司 | Passive keyless vehicle entry method and passive keyless vehicle entry system |
| CN115661935A (en) * | 2022-10-31 | 2023-01-31 | 海信集团控股股份有限公司 | Method and equipment for determining human body action accuracy |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111368635B (en) | 2021-05-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111368635B (en) | Millimeter wave-based multi-person gait recognition method and device | |
| CN108898086B (en) | Video image processing method and device, computer readable medium and electronic equipment | |
| JP7062878B2 (en) | Information processing method and information processing equipment | |
| US8750573B2 (en) | Hand gesture detection | |
| CN109325456B (en) | Target identification method, target identification device, target identification equipment and storage medium | |
| CN112528831B (en) | Multi-target attitude estimation method, multi-target attitude estimation device and terminal equipment | |
| KR101433472B1 (en) | Apparatus, method and computer readable recording medium for detecting, recognizing and tracking an object based on a situation recognition | |
| CN111401265A (en) | Pedestrian re-identification method and device, electronic equipment and computer-readable storage medium | |
| CN113095370B (en) | Image recognition method, device, electronic equipment and storage medium | |
| CN109116129B (en) | Terminal detection method, detection device, system and storage medium | |
| CN111505632A (en) | Ultra-wideband radar action attitude identification method based on power spectrum and Doppler characteristics | |
| Liu et al. | Hand Gesture Recognition Based on Single‐Shot Multibox Detector Deep Learning | |
| CN111723773A (en) | Remnant detection method, device, electronic equipment and readable storage medium | |
| CN111144284A (en) | Method and device for generating depth face image, electronic equipment and medium | |
| CN114509785A (en) | Three-dimensional object detection method, device, storage medium, processor and system | |
| KR20210157194A (en) | Crop growth measurement device using image processing and method thereof | |
| CN107832598B (en) | Unlocking control method and related product | |
| CN110348434A (en) | Camera source discrimination method, system, storage medium and calculating equipment | |
| CN111353325A (en) | Key point detection model training method and device | |
| Lecca et al. | Comprehensive evaluation of image enhancement for unsupervised image description and matching | |
| CN115273208A (en) | Track generation method, system and device and electronic equipment | |
| CN109919164B (en) | User interface object identification method and device | |
| CN112070035A (en) | Target tracking method and device based on video stream and storage medium | |
| CN111680670A (en) | Cross-mode human head detection method and device | |
| CN114863337A (en) | Novel screen anti-photographing recognition method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |