CN114792444A - Human recognition method and device based on gait characteristics - Google Patents
Human recognition method and device based on gait characteristics Download PDFInfo
- Publication number
- CN114792444A CN114792444A CN202210466060.7A CN202210466060A CN114792444A CN 114792444 A CN114792444 A CN 114792444A CN 202210466060 A CN202210466060 A CN 202210466060A CN 114792444 A CN114792444 A CN 114792444A
- Authority
- CN
- China
- Prior art keywords
- dimensional coordinates
- target person
- coordinates corresponding
- matrix
- joint nodes
- 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.)
- Pending
Links
Images
Landscapes
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
Abstract
The invention discloses a person identification method and a person identification device based on gait characteristics, wherein the method comprises the following steps: acquiring a moving image sequence of a target person; extracting three-dimensional coordinates of each moving joint point of the target person according to the moving image sequence, and constructing a three-dimensional coordinate matrix of the target person according to the three-dimensional coordinates; determining a motion characteristic matrix of the target person according to the three-dimensional coordinate matrix; and extracting the gait characteristics of the target person from the motion characteristic matrix through a first model network, and identifying the identity of the target person according to the gait characteristics. The invention realizes the beneficial effect of identifying the human beings in a longer distance.
Description
Technical Field
The invention relates to the technical field of artificial intelligence, in particular to a person identification method and device based on gait characteristics.
Background
With the continuous progress of technology, the technology of biological recognition (biometric technology) is developed, which applies a computer and identifies the identity of a person according to the physiological characteristics of the person. The identification system firstly samples the personal biological characteristics by means of photographing, scanning and the like, then digitalizes the sampling result, and compares the digitalized result with the information stored in the database to find out the corresponding identity of the person. Compared with the traditional pattern authentication and password authentication, the theoretical foundation of the biometric identification technology is the uniqueness of the human body, so that the biometric password cannot be copied, stolen and forgotten. The physiological features often selected for biometric identification are fingerprints, irises, faces, etc., wherein fingerprint identification and face identification are the most widely used biometric identification methods today.
Although the conventional fingerprint identification and face identification are convenient to operate and high in identification accuracy, the defects that the identification distance is short, the fingerprint identification needs to be performed by touching a fingerprint identifier with a finger, and the face identification can identify the face of about 4 meters farthest. The prior art lacks a scheme for identifying the identity of a person at a longer distance.
Disclosure of Invention
In order to solve at least one technical problem in the background art, the invention provides a person identification method and device based on gait characteristics.
In order to achieve the above object, according to an aspect of the present invention, there is provided a person recognition method based on gait characteristics, the method including:
acquiring a motion image sequence of a target person;
extracting three-dimensional coordinates of each moving joint point of the target person according to the moving image sequence, and constructing a three-dimensional coordinate matrix of the target person according to the three-dimensional coordinates;
determining a motion characteristic matrix of the target person according to the three-dimensional coordinate matrix;
and extracting the gait characteristics of the target person from the motion characteristic matrix through a first model network, and identifying the identity of the target person according to the gait characteristics.
Optionally, the person identification method based on gait characteristics further includes:
acquiring the acquired video data of the target person;
and identifying the target person by adopting a second model network aiming at each frame of image in the video data and extracting the person image to obtain the motion image sequence of the target person.
Optionally, the extracting three-dimensional coordinates of each moving joint point of the target person according to the moving image sequence includes:
extracting two-dimensional coordinates of each moving joint point of the target person through openposition 2D for each image in the moving image sequence respectively;
the two-dimensional coordinates are converted to three-dimensional coordinates with openpos 3D.
Optionally, the three-dimensional coordinate matrix includes: three-dimensional coordinates corresponding to the knee joint nodes, three-dimensional coordinates corresponding to the hip joint nodes and three-dimensional coordinates corresponding to the ankle joint nodes; the motion feature matrix specifically includes: knee joint motion characteristics;
the determining the motion characteristic matrix of the target person according to the three-dimensional coordinate matrix specifically includes:
and determining the knee joint movement characteristics of the target person according to the three-dimensional coordinates corresponding to the knee joint nodes, the three-dimensional coordinates corresponding to the hip joint nodes and the three-dimensional coordinates corresponding to the ankle joint nodes.
Optionally, the three-dimensional coordinate matrix includes: three-dimensional coordinates corresponding to the knee joint nodes and three-dimensional coordinates corresponding to the ankle joint nodes; the motion characteristic matrix specifically includes: a foot corner feature;
the determining the motion characteristic matrix of the target person according to the three-dimensional coordinate matrix specifically includes:
and determining the foot angle characteristics of the target character according to the three-dimensional coordinates corresponding to the knee joint nodes and the three-dimensional coordinates corresponding to the ankle joint nodes.
Optionally, the three-dimensional coordinate matrix includes: three-dimensional coordinates corresponding to the knee joint nodes and three-dimensional coordinates corresponding to the hip joint nodes; the motion feature matrix specifically includes: a center of gravity deviation feature;
the determining the motion characteristic matrix of the target person according to the three-dimensional coordinate matrix specifically includes:
and determining the gravity center deviation characteristic of the target person according to the three-dimensional coordinates corresponding to the knee joint nodes and the three-dimensional coordinates corresponding to the hip joint nodes.
Optionally, the three-dimensional coordinate matrix includes: three-dimensional coordinates corresponding to the ankle joint nodes and three-dimensional coordinates corresponding to the head joint nodes; the motion characteristic matrix specifically includes: body vertical height and step length features;
the determining the motion characteristic matrix of the target person according to the three-dimensional coordinate matrix specifically includes:
and determining the body vertical height and the step length characteristic of the target person according to the three-dimensional coordinates corresponding to the ankle joint nodes and the three-dimensional coordinates corresponding to the head joint nodes.
Optionally, the three-dimensional coordinate matrix includes: the three-dimensional coordinates corresponding to the knee joint nodes and the three-dimensional coordinates corresponding to the head joint nodes; the motion characteristic matrix specifically includes: human body structure motion characteristics;
the determining the motion characteristic matrix of the target person according to the three-dimensional coordinate matrix specifically includes:
and determining the human body structure motion characteristics of the target person according to the three-dimensional coordinates corresponding to the knee joint nodes and the three-dimensional coordinates corresponding to the head joint nodes.
Optionally, the first model network includes: a former network; the second model network comprises: YOLOv5 network.
In order to achieve the above object, according to another aspect of the present invention, there is provided a person recognition apparatus based on gait characteristics, the apparatus including:
an image data acquisition unit for acquiring a moving image sequence of a target person;
a three-dimensional coordinate matrix construction unit, configured to extract three-dimensional coordinates of each moving joint point of the target person according to the moving image sequence, and construct a three-dimensional coordinate matrix of the target person according to the three-dimensional coordinates;
the motion characteristic matrix generating unit is used for determining a motion characteristic matrix of the target person according to the three-dimensional coordinate matrix;
and the identity recognition unit is used for extracting the gait features of the target person from the motion feature matrix through a first model network and carrying out identity recognition on the target person according to the gait features.
In order to achieve the above object, according to another aspect of the present invention, there is also provided a computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the steps of the human figure identification method based on gait characteristics when executing the computer program.
In order to achieve the above object, according to another aspect of the present invention, there is also provided a computer readable storage medium having stored thereon a computer program/instructions which, when executed by a processor, implement the steps of the above-described gait feature-based person identification method.
To achieve the above object, according to another aspect of the present invention, there is also provided a computer program product comprising a computer program/instructions for implementing the steps of the human recognition method based on gait characteristics as described above when being executed by a processor.
The invention has the beneficial effects that:
the embodiment of the invention extracts the gait characteristics of the person according to the remotely acquired moving image data of the person, and then identifies the person according to the gait characteristics.
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 introduced below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to these drawings without creative efforts. In the drawings:
fig. 1 is a first flowchart of a human recognition method based on gait characteristics according to an embodiment of the invention;
fig. 2 is a second flowchart of a person identification method based on gait characteristics according to an embodiment of the invention;
FIG. 3 is a third flowchart of a method for identifying a person based on gait characteristics according to an embodiment of the invention;
FIG. 4 is a schematic view of a kinematic joint point according to an embodiment of the present invention;
FIG. 5 is a table of parameters of a Conformer network according to an embodiment of the present invention;
FIG. 6 is a diagram illustrating a structure of a Conformer network according to an embodiment of the present invention;
fig. 7 is a block diagram showing a configuration of a human recognition apparatus based on gait characteristics according to an embodiment of the present invention;
FIG. 8 is a schematic diagram of a computer apparatus according to an embodiment of the present invention.
Detailed Description
In order to make the technical solutions of the present invention better understood, 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 obtained by a person skilled in the art without making any creative effort based on the embodiments in the present invention, shall fall within the protection scope of the present invention.
As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
It should be noted that the terms "comprises" and "comprising," and any variations thereof, in the description and claims of this invention and the above-described drawings are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements explicitly listed, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
It should be noted that the embodiments and features of the embodiments may be combined with each other without conflict. The present invention will be described in detail below with reference to the embodiments with reference to the attached drawings.
It should be noted that, in the technical solution of the present application, permission of the target person is obtained for data acquisition, storage, use, processing, and the like, that is, all of the permission meet relevant regulations of national laws and regulations.
Fig. 1 is a first flowchart of a human figure recognition method based on gait characteristics according to an embodiment of the present invention, and as shown in fig. 1, in an embodiment of the present invention, the human figure recognition method based on gait characteristics according to the present invention includes steps S101 to S104.
In step S101, a moving image sequence of a target person is acquired.
And S102, extracting three-dimensional coordinates of each moving joint point of the target person according to the moving image sequence, and constructing a three-dimensional coordinate matrix of the target person according to the three-dimensional coordinates.
And S103, determining a motion characteristic matrix of the target person according to the three-dimensional coordinate matrix.
In the embodiment of the present invention, the motion feature matrix specifically has five features, which are respectively: characteristics of knee joint movement
Foot corner feature
Center of gravity deviation feature
Body vertical height and step size features
Structural movement characteristics of the human body
Because the lower limbs are the main moving parts of the human body during walking, the selected characteristics are mostly related to the lower limbs. Finally, a client motion characteristic matrix M can be constructed r 。
And step S104, extracting the gait characteristics of the target person from the motion characteristic matrix through a first model network, and identifying the identity of the target person according to the gait characteristics.
Although the present fingerprint identification and face identification are convenient to operate and high in identification accuracy, the present fingerprint identification and face identification have two defects: firstly, the identification distance is short, the fingerprint identification needs to touch a fingerprint identifier by a finger, and the face identification is farthest so as to identify the face of about 4 meters; and secondly, the cooperation of recognized personnel is needed, and the operation is not convenient enough. Gait recognition is an emerging biometric technology, and perfectly solves the problems. The walking posture of each person is influenced by various factors such as muscle growth conditions, bone density, weight, gravity center and the like of each person, so that the gait of each person is unique, is difficult to imitate by other people, can be used as a biological password for biological identification, has the identification distance of over 20 meters for the gait identification, does not need personal cooperation, and is suitable for being applied to a bank scene to identify the identity of a client.
In an embodiment of the invention, the first model network adopts a former network, and the invention applies the former network to extract gait features in a motion feature matrix and identify the identity of a person.
Specific parameters of the Conformer network in an embodiment of the present invention are shown in fig. 5, and these parameters and structures may be fine-tuned according to different application scenarios and data characteristics.
The former network combines the features of a convolutional neural network and a transformer network. The convolutional neural network is good at capturing local characteristic information, but is difficult to acquire global characteristic information; and the attention mechanism in the Transformer network can capture the characteristic information of the long production distance and weaken the local characteristic information. The former network combines the advantages of the two, and relies on a characteristic Feature Coupling Unit (Feature Coupling Unit) to integrate convolution extraction with local features and transform extraction with global features in an interactive manner. The former network also adopts a parallel network, so that two characteristics can be obtained to the maximum extent, and the calculation speed can be greatly increased, wherein the network structure is shown in fig. 6.
As is apparent from fig. 6, the convolutional neural network acquires local features through multilayer convolution, and the Vision Transformer can acquire global features through cascaded attention modules, in the FCU of the network, the local features acquired by the convolutional neural network branches are fed into the Vision Transformer network to enhance the local perceptibility of the network, and similarly, the global features in the Vision Transformer network are fed into the convolutional neural network branches to enhance the global perceptibility.
In the network overall architecture diagram of fig. 6, (a) is how to transform in spatial dimension during the convolution branch and transform data interaction; (b) is the network specific details, some of the details of the convolution branches and the transformers can be seen. (c) The network overall flow chart firstly carries out two parallel branch networks through a stem block unit: and (4) carrying out convolution branching and Transformer branching to finally obtain the classification results of the two branches, and summing the prediction results of the two classifiers during reasoning.
The processing of the input vector by the former network is functionally represented as:
wherein FFN refers to a Feed Forward Module (Feed Forward Module) in the network, MHSA refers to an Attention Module (Multi-Head Self attachment Module), Conv refers to a Convolution Module (Convolution Module), and layerorm refers to a normalization Module.
Fig. 2 is a second flowchart of a human recognition method based on gait characteristics according to an embodiment of the present invention, as shown in fig. 2, in an embodiment of the present invention, the moving image sequence in step S101 is generated by step S201 and step S202.
Step S201, acquiring the acquired video data of the target person.
In an embodiment of the invention, the invention can install an image sensor (camera) at places such as bank business halls, network sites and the like which need to carry out gait recognition, and the image sensor (camera) is used for collecting the video data of the target person.
Step S202, aiming at each frame of image in the video data, adopting a second model network to identify the target person and extracting the person image to obtain the motion image sequence of the target person.
In one embodiment of the present invention, the second model network employs a YOLOv5 network.
In one embodiment of the present invention, the present invention applies YOLOv5 target detection algorithm to perform person recognition, tracking and segmentation on each acquired frame image.
In one embodiment of the present invention, the input of the YOLOv5 network is 608 × 608, so the input includes an image pre-processing stage, and an adaptive picture scaling method is applied. The reference network of the YOLOv5 network is usually a network of some excellent classifier, and the module is used to extract some general feature representations. The Neck network of the YOLOv5 network is usually located in the middle of the reference network and the Head network, and the diversity and the robustness of the features can be further improved by using the Neck network. The Head output end of the YOLOv5 network is used for completing the output of the target detection result. For different detection algorithms, the number of branches at the output end is different, and the detection algorithm usually comprises a classification branch and a regression branch.
As shown in fig. 3, in one embodiment of the present invention, the extracting three-dimensional coordinates of each moving joint point of the target person according to the moving image sequence in step S102 specifically includes step S301 and step S302.
In step S301, two-dimensional coordinates of each of the moving joint points of the target person are extracted by openpos 2D for each of the images in the moving image sequence.
Step S302, converting the two-dimensional coordinates into three-dimensional coordinates by openpos 3D.
In an embodiment of the invention, openpos 3D is used for extracting human body 3D information, 2D coordinates of 14 kinematic joint points (as shown in fig. 4) of a segmented client image are firstly acquired through openpos 2D, and then the acquired 2D coordinates are converted into 3D coordinates through openpos 3D to construct a gait 3D coordinate matrix.
In a specific method, use
And 3D coordinates of the i-th moving joint point of the person in the t-th frame are shown. And then splicing all the moving joint points of each frame to be used as column vectors to generate a 3D coordinate matrix. In fig. 4, the node 0 is a head joint, 8 and 11 are hip joints, 9 and 12 are knee joints, and 10 and 13 are ankle joints.
In one embodiment of the present invention, the three-dimensional coordinate matrix includes: three-dimensional coordinates corresponding to the knee joint nodes, three-dimensional coordinates corresponding to the hip joint nodes and three-dimensional coordinates corresponding to the ankle joint nodes; the motion feature matrix specifically includes: knee joint motion characteristics.
In an embodiment of the present invention, the determining the motion characteristic matrix of the target person according to the three-dimensional coordinate matrix in step S103 includes:
and determining the knee joint movement characteristics of the target person according to the three-dimensional coordinates corresponding to the knee joint nodes, the three-dimensional coordinates corresponding to the hip joint nodes and the three-dimensional coordinates corresponding to the ankle joint nodes.
In the present invention, the knee joint movement characteristics
The included angle between the knee joint and the upper and lower nodes, namely the bending angle of the knee, is calculated according to the following formula, wherein k is b Represents the knee joint nodes, nodes 9 and 12;
is a coordinate value of the knee joint node, k i And k j Are adjacent to knee joint nodes, hip joint nodes (8, 11) and ankle joint nodes (10, 13),
is a coordinate value of the hip joint node,
are coordinate values of the ankle joint node.
In one embodiment of the present invention, the three-dimensional coordinate matrix includes: three-dimensional coordinates corresponding to the knee joint nodes and three-dimensional coordinates corresponding to the ankle joint nodes; the motion characteristic matrix specifically includes: a corner feature.
In an embodiment of the present invention, the determining the motion characteristic matrix of the target person according to the three-dimensional coordinate matrix in step S103 includes:
and determining the foot angle characteristics of the target character according to the three-dimensional coordinates corresponding to the knee joint nodes and the three-dimensional coordinates corresponding to the ankle joint nodes.
In the present invention, the corner features
Is the included angle between the ankle joint node and the body gravity line normal, and parameters in a calculation formula are as follows,
is a normal vector of the ground surface,
is a coordinate value of the knee joint node,
is the coordinate value of the ankle joint node:
in one embodiment of the present invention, the three-dimensional coordinate matrix comprises: three-dimensional coordinates corresponding to the knee joint nodes and three-dimensional coordinates corresponding to the hip joint nodes; the motion characteristic matrix specifically includes: a center of gravity deviation feature.
In an embodiment of the present invention, the determining the motion characteristic matrix of the target person according to the three-dimensional coordinate matrix in step S103 specifically includes:
and determining the gravity center deviation characteristic of the target person according to the three-dimensional coordinates corresponding to the knee joint nodes and the three-dimensional coordinates corresponding to the hip joint nodes.
In the present invention, the center of gravity deviation feature
Is the body weight line N G The included angle formed between the lower limb joint and the lower limb joint is calculated according to the following formula:
in one embodiment of the present invention, the three-dimensional coordinate matrix includes: three-dimensional coordinates corresponding to the ankle joint nodes and three-dimensional coordinates corresponding to the head joint nodes; the motion characteristic matrix specifically includes: body vertical height and step size characteristics.
In an embodiment of the present invention, the determining the motion characteristic matrix of the target person according to the three-dimensional coordinate matrix in step S103 specifically includes:
and determining the body vertical height and the step length characteristic of the target person according to the three-dimensional coordinates corresponding to the ankle joint nodes and the three-dimensional coordinates corresponding to the head joint nodes.
In the present invention, the body vertical height and step length features
Is the ratio of the height to the step length during walking, and the calculation formula is as follows, wherein x 10 Is the x-axis coordinate, x, of the ankle joint 10 13 Is the x-axis coordinate, y, of the ankle joint 13 10 Is the y-axis coordinate, y, of the ankle joint 10 13 Is the y-axis coordinate, z, of the ankle joint 13 0 As z-axis coordinates of head joint node 0:
in one embodiment of the present invention, the three-dimensional coordinate matrix comprises: three-dimensional coordinates corresponding to the knee joint nodes and three-dimensional coordinates corresponding to the head joint nodes; the motion feature matrix specifically includes: human structure motion characteristics.
In an embodiment of the present invention, the determining the motion characteristic matrix of the target person according to the three-dimensional coordinate matrix in step S103 includes:
and determining the human body structure motion characteristics of the target person according to the three-dimensional coordinates corresponding to the knee joint nodes and the three-dimensional coordinates corresponding to the head joint nodes.
In the present invention, the human body structure movement characteristics
Is the ratio of the height to the knee joint during walking, and the calculation formula is as follows, wherein x 0 Is the x-axis coordinate of the head joint node 0, y 0 Is the y-axis coordinate of head joint node 0, z 0 Is the z-axis coordinate, x, of the head joint node 0 kb Is the x-axis coordinate, y, of the knee joint nodes 9 and 12 kb For the y-axis coordinates of knee joints 9 and 12:
it can be seen from the above embodiments that the present invention has the following advantages:
(1) the invention has long recognition distance to the client, and the recognition distance can reach 20 meters furthest. The gait recognition can recognize the identity of the client at a long distance, and compared with biological recognition modes such as face recognition, fingerprint recognition and the like, the recognition distance is improved qualitatively.
(2) The identification process of the invention is not influenced by behaviors such as customer occlusion and the like. Under the condition of wearing the mask, the face recognition cannot identify a person wearing the mask, the hidden health hazard exists when the mask is taken off, and the gait recognition is not influenced by shielding of clothes, hats and the like of clients.
(3) The identification process of the present invention does not require customer cooperation. The customer can finish the identification by walking before the sensor (camera), so that more matching steps can not be added for the customer, and the customer feels convenient and comfortable.
It should be noted that the steps illustrated in the flowcharts of the figures may be performed in a computer system such as a set of computer-executable instructions and that, although a logical order is illustrated in the flowcharts, in some cases, the steps illustrated or described may be performed in an order different than presented herein.
Based on the same inventive concept, the embodiment of the present invention further provides a person identification apparatus based on gait characteristics, which can be used to implement the person identification method based on gait characteristics described in the above embodiment, as described in the following embodiment. Because the principle of solving the problem of the person identification device based on the gait features is similar to that of the person identification method based on the gait features, the embodiment of the person identification device based on the gait features can be referred to as the embodiment of the person identification method based on the gait features, and repeated details are not repeated. As used hereinafter, the term "unit" or "module" may be a combination of software and/or hardware that implements a predetermined function. Although the means described in the embodiments below are preferably implemented in software, an implementation in hardware, or a combination of software and hardware is also possible and contemplated.
Fig. 7 is a block diagram showing a configuration of a person recognition apparatus based on gait characteristics according to an embodiment of the present invention, and as shown in fig. 7, in an embodiment of the present invention, the person recognition apparatus based on gait characteristics of the present invention includes:
an image data acquisition unit 1 for acquiring a moving image sequence of a target person;
a three-dimensional coordinate matrix constructing unit 2, configured to extract three-dimensional coordinates of each moving joint point of the target person according to the moving image sequence, and construct a three-dimensional coordinate matrix of the target person according to the three-dimensional coordinates;
the motion characteristic matrix generating unit 3 is used for determining a motion characteristic matrix of the target person according to the three-dimensional coordinate matrix;
and the identity recognition unit 4 is configured to extract the gait features of the target person from the motion feature matrix through a first model network, and perform identity recognition on the target person according to the gait features.
In one embodiment of the present invention, the human recognition apparatus based on gait characteristics of the present invention further includes:
the video data acquisition unit is used for acquiring the acquired video data of the target person;
and the moving image sequence extraction unit is used for identifying the target person by adopting a second model network aiming at each frame of image in the video data and extracting the person image to obtain the moving image sequence of the target person.
In an embodiment of the present invention, the three-dimensional coordinate matrix building unit 2 includes:
a two-dimensional coordinate extraction module for extracting two-dimensional coordinates of each of the moving joint points of the target person by openpos 2D for each of the images in the moving image sequence, respectively;
and the coordinate conversion module is used for converting the two-dimensional coordinates into three-dimensional coordinates by using Openpos 3D.
In one embodiment of the present invention, the three-dimensional coordinate matrix comprises: three-dimensional coordinates corresponding to the knee joint nodes, three-dimensional coordinates corresponding to the hip joint nodes and three-dimensional coordinates corresponding to the ankle joint nodes; the motion feature matrix specifically includes: knee joint motion characteristics; the motion characteristic matrix generation unit 3 is specifically configured to determine the knee joint motion characteristic of the target person according to the three-dimensional coordinate corresponding to the knee joint node, the three-dimensional coordinate corresponding to the hip joint node, and the three-dimensional coordinate corresponding to the ankle joint node.
In one embodiment of the present invention, the three-dimensional coordinate matrix comprises: three-dimensional coordinates corresponding to the knee joint nodes and three-dimensional coordinates corresponding to the ankle joint nodes; the motion feature matrix specifically includes: a foot corner feature; the motion characteristic matrix generating unit 3 is specifically configured to determine the toe characteristic of the target person according to the three-dimensional coordinates corresponding to the knee joint nodes and the three-dimensional coordinates corresponding to the ankle joint nodes.
In one embodiment of the present invention, the three-dimensional coordinate matrix includes: three-dimensional coordinates corresponding to the knee joint nodes and three-dimensional coordinates corresponding to the hip joint nodes; the motion feature matrix specifically includes: a center of gravity deviation feature; the motion characteristic matrix generation unit 3 is specifically configured to determine the gravity center deviation characteristic of the target person according to the three-dimensional coordinates corresponding to the knee joint nodes and the three-dimensional coordinates corresponding to the hip joint nodes.
In one embodiment of the present invention, the three-dimensional coordinate matrix includes: three-dimensional coordinates corresponding to the ankle joint nodes and three-dimensional coordinates corresponding to the head joint nodes; the motion characteristic matrix specifically includes: body vertical height and step length features; the motion characteristic matrix generation unit 3 is specifically configured to determine the body vertical height and the step length characteristic of the target person according to the three-dimensional coordinates corresponding to the ankle joint nodes and the three-dimensional coordinates corresponding to the head joint nodes.
In one embodiment of the present invention, the three-dimensional coordinate matrix comprises: three-dimensional coordinates corresponding to the knee joint nodes and three-dimensional coordinates corresponding to the head joint nodes; the motion characteristic matrix specifically includes: human body structure motion characteristics; the motion characteristic matrix generation unit 3 is specifically configured to determine the human body structure motion characteristic of the target person according to the three-dimensional coordinates corresponding to the knee joint nodes and the three-dimensional coordinates corresponding to the head joint nodes.
To achieve the above object, according to another aspect of the present application, there is also provided a computer apparatus. As shown in fig. 8, the computer device comprises a memory, a processor, a communication interface and a communication bus, wherein a computer program that can be run on the processor is stored in the memory, and the steps of the method of the above embodiment are realized when the processor executes the computer program.
The processor may be a Central Processing Unit (CPU). The Processor may also be other general purpose Processor, Digital Signal Processor (DSP), Application Specific Integrated Circuit (ASIC), Field Programmable Gate Array (FPGA) or other Programmable logic device, discrete Gate or transistor logic device, discrete hardware component, or a combination thereof.
The memory, which is a non-transitory computer readable storage medium, may be used to store non-transitory software programs, non-transitory computer executable programs, and units, such as the corresponding program units in the above-described method embodiments of the present invention. The processor executes the non-transitory software programs, instructions and modules stored in the memory so as to execute various functional applications of the processor and processing of the work data, i.e., to implement the methods in the above method embodiments.
The memory may include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required for at least one function; the storage data area may store data created by the processor, and the like. Further, the memory may include high speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid state storage device. In some embodiments, the memory optionally includes memory located remotely from the processor, and such remote memory may be coupled to the processor via a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.
The one or more units are stored in the memory and, when executed by the processor, perform the method of the above embodiment.
The specific details of the computer device may be understood by referring to the corresponding related description and effects in the foregoing embodiments, which are not described herein again.
In order to achieve the above object, according to another aspect of the present application, there is also provided a computer-readable storage medium storing a computer program which, when executed in a computer processor, implements the steps in the above-described human figure recognition method based on gait characteristics. Those skilled in the art will appreciate that all or part of the processes of the methods of the embodiments described above can be implemented by a computer program, which can be stored in a computer-readable storage medium and can include the processes of the embodiments of the methods described above when executed. The storage medium may be a magnetic Disk, an optical Disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a Flash Memory (Flash Memory), a Hard Disk Drive (Hard Disk Drive, abbreviated as HDD), or a Solid State Drive (SSD); the storage medium may also comprise a combination of memories of the kind described above.
To achieve the above object, according to another aspect of the present application, there is also provided a computer program product comprising a computer program/instructions which, when executed by a processor, implement the steps of the above-mentioned person identification method based on gait characteristics.
It will be apparent to those skilled in the art that the modules or steps of the present invention described above may be implemented by a general purpose computing device, they may be centralized on a single computing device or distributed across a network of multiple computing devices, and they may alternatively be implemented by program code executable by a computing device, such that they may be stored in a storage device and executed by a computing device, or fabricated separately as individual integrated circuit modules, or fabricated as a single integrated circuit module from multiple modules or steps. Thus, the present invention is not limited to any specific combination of hardware and software.
The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (13)
1. A human recognition method based on gait characteristics is characterized by comprising the following steps:
acquiring a motion image sequence of a target person;
extracting three-dimensional coordinates of each moving joint point of the target person according to the moving image sequence, and constructing a three-dimensional coordinate matrix of the target person according to the three-dimensional coordinates;
determining a motion characteristic matrix of the target person according to the three-dimensional coordinate matrix;
and extracting the gait characteristics of the target person from the motion characteristic matrix through a first model network, and identifying the identity of the target person according to the gait characteristics.
2. The human recognition method based on gait characteristics according to claim 1, characterized by further comprising:
acquiring the collected video data of the target person;
and identifying the target person by adopting a second model network aiming at each frame of image in the video data and extracting the person image to obtain the motion image sequence of the target person.
3. The human recognition method based on gait features according to claim 1, wherein the extracting three-dimensional coordinates of each of the moving joint points of the target human from the moving image sequence includes:
extracting two-dimensional coordinates of each moving joint point of the target person through openposition 2D for each image in the moving image sequence respectively;
the two-dimensional coordinates are converted to three-dimensional coordinates with openpos 3D.
4. The method of claim 1, wherein the three-dimensional coordinate matrix comprises: three-dimensional coordinates corresponding to the knee joint nodes, three-dimensional coordinates corresponding to the hip joint nodes and three-dimensional coordinates corresponding to the ankle joint nodes; the motion feature matrix specifically includes: knee joint motion characteristics;
the determining the motion characteristic matrix of the target person according to the three-dimensional coordinate matrix specifically includes:
and determining the knee joint movement characteristics of the target person according to the three-dimensional coordinates corresponding to the knee joint nodes, the three-dimensional coordinates corresponding to the hip joint nodes and the three-dimensional coordinates corresponding to the ankle joint nodes.
5. The method of claim 1, wherein the three-dimensional coordinate matrix comprises: three-dimensional coordinates corresponding to the knee joint nodes and three-dimensional coordinates corresponding to the ankle joint nodes; the motion feature matrix specifically includes: a foot corner feature;
the determining the motion characteristic matrix of the target person according to the three-dimensional coordinate matrix specifically includes:
and determining the foot angle characteristics of the target character according to the three-dimensional coordinates corresponding to the knee joint nodes and the three-dimensional coordinates corresponding to the ankle joint nodes.
6. The human recognition method based on gait features according to claim 1, wherein the three-dimensional coordinate matrix includes: three-dimensional coordinates corresponding to the knee joint nodes and three-dimensional coordinates corresponding to the hip joint nodes; the motion feature matrix specifically includes: a center of gravity deviation feature;
the determining the motion characteristic matrix of the target person according to the three-dimensional coordinate matrix specifically includes:
and determining the gravity center deviation characteristic of the target person according to the three-dimensional coordinates corresponding to the knee joint nodes and the three-dimensional coordinates corresponding to the hip joint nodes.
7. The method of claim 1, wherein the three-dimensional coordinate matrix comprises: three-dimensional coordinates corresponding to the ankle joint nodes and three-dimensional coordinates corresponding to the head joint nodes; the motion feature matrix specifically includes: body vertical height and step length features;
the determining the motion characteristic matrix of the target person according to the three-dimensional coordinate matrix specifically includes:
and determining the body vertical height and the step length characteristic of the target person according to the three-dimensional coordinates corresponding to the ankle joint nodes and the three-dimensional coordinates corresponding to the head joint nodes.
8. The human recognition method based on gait features according to claim 1, wherein the three-dimensional coordinate matrix includes: the three-dimensional coordinates corresponding to the knee joint nodes and the three-dimensional coordinates corresponding to the head joint nodes; the motion characteristic matrix specifically includes: human body structure motion characteristics;
the determining the motion characteristic matrix of the target person according to the three-dimensional coordinate matrix specifically includes:
and determining the human body structure motion characteristics of the target person according to the three-dimensional coordinates corresponding to the knee joint nodes and the three-dimensional coordinates corresponding to the head joint nodes.
9. The gait feature-based person identification method according to claim 1, wherein the first model network includes: a former network; the second model network comprises: YOLOv5 network.
10. A person recognition device based on gait characteristics, comprising:
an image data acquisition unit for acquiring a moving image sequence of a target person;
a three-dimensional coordinate matrix construction unit, configured to extract three-dimensional coordinates of each moving joint point of the target person according to the moving image sequence, and construct a three-dimensional coordinate matrix of the target person according to the three-dimensional coordinates;
the motion characteristic matrix generating unit is used for determining a motion characteristic matrix of the target person according to the three-dimensional coordinate matrix;
and the identity recognition unit is used for extracting the gait features of the target person from the motion feature matrix through a first model network and carrying out identity recognition on the target person according to the gait features.
11. A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the steps of the method of any of claims 1 to 9 are implemented when the computer program is executed by the processor.
12. A computer-readable storage medium on which a computer program/instructions are stored, characterized in that the computer program/instructions, when executed by a processor, implement the steps of the method of any one of claims 1 to 9.
13. A computer program product comprising computer program/instructions, characterized in that the computer program/instructions, when executed by a processor, implement the steps of the method of any one of claims 1 to 9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210466060.7A CN114792444A (en) | 2022-04-29 | 2022-04-29 | Human recognition method and device based on gait characteristics |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210466060.7A CN114792444A (en) | 2022-04-29 | 2022-04-29 | Human recognition method and device based on gait characteristics |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114792444A true CN114792444A (en) | 2022-07-26 |
Family
ID=82460972
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210466060.7A Pending CN114792444A (en) | 2022-04-29 | 2022-04-29 | Human recognition method and device based on gait characteristics |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114792444A (en) |
-
2022
- 2022-04-29 CN CN202210466060.7A patent/CN114792444A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Wechsler | Reliable Face Recognition Methods: System Design, Impementation and Evaluation | |
| Goffredo et al. | Self-calibrating view-invariant gait biometrics | |
| DE69910757T2 (en) | WAVELET-BASED FACIAL MOTION DETECTION FOR AVATAR ANIMATION | |
| CN109299659A (en) | A kind of human posture recognition method and system based on RGB camera and deep learning | |
| WO2019120108A1 (en) | Image coding method, action recognition method, and computer device | |
| CN113706699A (en) | Data processing method and device, electronic equipment and computer readable storage medium | |
| Shinzaki et al. | Gait-based person identification method using shadow biometrics for robustness to changes in the walking direction | |
| Parashar et al. | Intra-class variations with deep learning-based gait analysis: A comprehensive survey of covariates and methods | |
| Lei et al. | Gart: Gaussian articulated template models | |
| CN109919137B (en) | Pedestrian structural feature expression method | |
| CN106815855A (en) | Based on the human body motion tracking method that production and discriminate combine | |
| CN111414840A (en) | Gait recognition method, device, equipment and computer readable storage medium | |
| Yan et al. | Human-object interaction recognition using multitask neural network | |
| JP2019003565A (en) | Image processing apparatus, image processing method and image processing program | |
| CN111507239A (en) | Local feature face recognition method based on image pyramid | |
| Krzeszowski et al. | Gait recognition based on marker-less 3D motion capture | |
| CN111652018A (en) | Face registration method and authentication method | |
| Bakchy et al. | Human gait analysis using gait energy image | |
| Alashkar et al. | A 3D dynamic database for unconstrained face recognition | |
| Lobachev et al. | Machine learning models and methods for human gait recognition | |
| CN114792444A (en) | Human recognition method and device based on gait characteristics | |
| Bakchy et al. | Human identification on the basis of gait analysis using Kohonen self-organizing mapping technique | |
| CN111444374B (en) | Human body retrieval system and method | |
| CN115880766A (en) | Method and device for training posture migration and posture migration models and storage medium | |
| CN113610969A (en) | Three-dimensional human body model generation method and device, electronic equipment and storage medium |
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 |