CN113762082B - Unsupervised skeleton action recognition method based on cyclic graph convolution automatic encoder - Google Patents
Unsupervised skeleton action recognition method based on cyclic graph convolution automatic encoder Download PDFInfo
- Publication number
- CN113762082B CN113762082B CN202110908006.9A CN202110908006A CN113762082B CN 113762082 B CN113762082 B CN 113762082B CN 202110908006 A CN202110908006 A CN 202110908006A CN 113762082 B CN113762082 B CN 113762082B
- Authority
- CN
- China
- Prior art keywords
- skeleton
- representing
- graph convolution
- action sequence
- sequence
- 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.)
- Active
Links
- 230000009471 action Effects 0.000 title claims abstract description 85
- 238000000034 method Methods 0.000 title claims abstract description 43
- 125000004122 cyclic group Chemical group 0.000 title claims abstract description 41
- 239000013598 vector Substances 0.000 claims abstract description 34
- 238000012512 characterization method Methods 0.000 claims abstract description 19
- 238000007635 classification algorithm Methods 0.000 claims abstract description 8
- 239000011159 matrix material Substances 0.000 claims description 20
- 238000012549 training Methods 0.000 claims description 18
- 230000004913 activation Effects 0.000 claims description 9
- 238000004364 calculation method Methods 0.000 claims description 8
- 230000008569 process Effects 0.000 claims description 7
- 238000010586 diagram Methods 0.000 claims description 4
- 238000011478 gradient descent method Methods 0.000 claims description 3
- 230000033001 locomotion Effects 0.000 description 27
- 230000006870 function Effects 0.000 description 14
- 238000011160 research Methods 0.000 description 4
- 238000010998 test method Methods 0.000 description 4
- 238000012360 testing method Methods 0.000 description 4
- 238000011161 development Methods 0.000 description 3
- 230000018109 developmental process Effects 0.000 description 3
- 238000005096 rolling process Methods 0.000 description 3
- 241000282412 Homo Species 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 230000004438 eyesight Effects 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 230000007774 longterm Effects 0.000 description 2
- 241001465754 Metazoa Species 0.000 description 1
- 230000003042 antagnostic effect Effects 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000004880 explosion Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 241000894007 species Species 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
- 230000016776 visual perception Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2413—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on distances to training or reference patterns
- G06F18/24147—Distances to closest patterns, e.g. nearest neighbour classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/088—Non-supervised learning, e.g. competitive learning
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Mathematical Physics (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to an unsupervised skeleton action recognition method based on a cyclic graph convolution automatic encoder, which is characterized by comprising the following steps of: inputting the human skeleton action sequence to a cyclic graph convolution encoder; outputting a representation vector of the action sequence by a cyclic graph convolution encoder; calculating a characterization vector of the action sequence through a weighted nearest neighbor classification algorithm to obtain an identification category of the human skeleton action sequence; the cyclic graph convolution encoder includes: the multi-layer space joint attention module is used for combining the human skeleton action sequence and a hidden layer of the cyclic graph convolution encoder, and adaptively measuring the importance of different joints with different actions to obtain a weighted skeleton sequence; and the multi-layer graph convolution gating circulating unit layer is used for integrating the connection relation characteristics of the weighted skeleton sequences to obtain the characterization vector of the action sequence. Compared with the prior art, the method can remarkably improve the recognition precision of the non-supervision action recognition system, and has wide application prospect.
Description
Technical Field
The invention relates to the technical field of computer vision and motion recognition, in particular to an unsupervised skeleton motion recognition method based on a cyclic graph convolution automatic encoder.
Background
Movements of living organisms in nature, such as humans and animals, including whole body movements, and movements of parts of the body, such as the head, limbs, hands, eyes, etc., are commonly referred to as biological movements. These forms of motion are critical to human perception of dynamic environmental changes and to infer intent of others or other species. Identifying and understanding the motion of an observed individual is a fundamental attribute of human visual perception, and the ability to identify motion in different scenarios is also critical. For the above reasons, human motion recognition tasks have attracted the attention of a large number of researchers in the field of computer vision. The motion recognition task is widely applied, for example, in the fields of video monitoring, man-machine interaction, motion analysis and the like, so that the motion recognition task is gradually developed into an important research direction. The research on human body motion recognition can be traced to 1973, johansson finds that human body motion is mainly realized through movement of a plurality of key skeleton points of the human body through experimental observation, and the combination and tracking of 10-12 key nodes can characterize such motions as walking, running, dancing and the like, so that the recognition of human body motion is realized.
In recent years, with the successive advent and rapid development of depth sensors such as Kinect and RealSense, humans can more conveniently obtain RGB information, depth information, and skeleton information of an image. This has also led to tremendous developments in the field of motion recognition. The early action recognition method is mostly based on video sequences, but has the defects of high computational complexity, easiness in being influenced by other factors and the like, but has the advantages that skeleton information is very robust to factors such as human appearance, environment interaction, visual angle change and the like, and meanwhile, the computational complexity is low, and data is easy to store. Action recognition based on skeleton data is a rapidly developing research direction, and effective action recognition can be performed by using the change information of the key points.
At present, research in the field of motion recognition based on skeleton data also changes rapidly along with the development of deep learning technology, and the recognition is performed from the earliest characteristic extracted manually to the current cyclic neural network and convolutional neural network. However, these methods cannot utilize the topological features of the skeleton data itself, and thus the recognition accuracy has yet to be improved.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provide an unsupervised skeleton action recognition method based on a cyclic graph convolution automatic encoder.
The aim of the invention can be achieved by the following technical scheme:
an unsupervised skeleton action recognition method based on a cyclic graph convolution automatic encoder comprises the following steps:
s1, inputting a human skeleton action sequence to a cyclic graph convolution encoder;
s2, outputting and obtaining a characterization vector of the action sequence by a cyclic graph convolution encoder;
s3, calculating a characterization vector of the action sequence through a weighted nearest neighbor classification algorithm to obtain an identification category of the human skeleton action sequence;
the cyclic graph convolution encoder includes: the multi-layer space joint attention module is used for combining the human skeleton action sequence and a hidden layer of the cyclic graph convolution encoder, and adaptively measuring the importance of different joints with different actions to obtain a weighted skeleton sequence; and the multi-layer graph convolution gating circulating unit layer is used for integrating the connection relation characteristics of the weighted skeleton sequences to obtain the characterization vector of the action sequence.
Further, in the spatial joint attention module, the weighted skeleton sequence calculation expression is:
x′ t ＝(α t +1)·x t
s t ＝U s φ(W x x t +W h h t-1 +b s )+b u
wherein x 'is' t Representing a weighted skeleton sequence, alpha t Representing the importance of each joint s t Representing the importance score of each joint,representing the sequence coordinates of N joints at time t, h t-1 Representing hidden layer information, W x And W is h Represents a matrix of learnable parameters, phi represents an activation function, b s And b u Representing the bias.
Further, in the graph rolling gate control circulation unit layer, the expression of the connection relation characteristic of the integrated weighted skeleton sequence is as follows:
wherein H is (l+1) Indicating that the output of layer l +1 is convolved,representing a symmetrical adjacency matrix with spin, A representing the adjacency matrix, I representing the identity matrix,/->For the degree matrix, τ represents the activation function, H (l) Representing the output of layer I of the graph convolution, Θ (l) Representing a matrix of learnable parameters of the first layer.
Further, the representation of the graph roll-up gating cycle cell layer is:
wherein z is t Representing an update gate, r t Representing a reset gate and,activation vector representing candidate->Representative graph convolution sum H (l +1) Correspondingly, W xz 、W hz 、W xr 、W hr 、W xh And W is hh Representing the parameter matrix in different gating, +..
Further, the training step of the cyclic graph convolutional encoder comprises the following steps of
A1, inputting a training action sequence set to a cyclic graph convolution encoder so as to obtain a characterization vector of an action sequence;
a2, inputting the characterization vector and the hidden layer vector of the action sequence to a decoder, and performing sequence restoration to obtain a reconstructed action sequence set;
a3, comparing the reconstructed action sequence set with the training action sequence set, and calculating a loss function value through a reconstruction loss function;
repeating the steps A1 to A3 until the loss function value reaches the preset cut-off condition.
Further, the hidden layer vector is a vector with the same length as the human skeleton action sequence and the value of zero.
Further, the expression of the reconstruction loss function is:
in the method, in the process of the invention,representing training action sequence set,/->Representing a set of reconstructed action sequences, I.I F Representing the Frobenius norm and L representing the loss function value.
Further, the cyclic graph convolutional encoder is trained by adopting a gradient descent method.
Further, the space joint attention module and the graph rolling gate control circulation unit layer are three layers.
Further, in the weighted nearest neighbor classification algorithm, after k nearest samples are obtained, k is a set value, the number of votes of each category is calculated, and a recognition result is obtained through weighted voting, wherein the calculation expression of the weight is as follows:
wherein w is i And d i The voting weights and cosine distances of the samples i are shown, respectively.
Compared with the prior art, the invention has the following beneficial effects:
the cyclic graph convolution encoder is applied to skeleton action recognition, and the cyclic graph convolution encoder is provided with the multi-layer spatial joint attention module, so that the spatial topological relation of skeleton sequence data is considered in the recognition process, and the recognition precision is improved by utilizing the space-time dependency relation of an action sequence; meanwhile, the weighted nearest neighbor classification algorithm is adopted as the classifier for final recognition, and the idea of exponential explosion is utilized to ensure that samples beneficial to the result have larger voting weight, so that the recognition accuracy is further improved.
Drawings
FIG. 1 is a schematic overall flow chart of the present invention.
Fig. 2 is a schematic diagram of a spatial joint attention module.
FIG. 3 is a schematic diagram of a layer of a graph roll-up gating cycle cell.
Detailed Description
The invention will now be described in detail with reference to the drawings and specific examples. The present embodiment is implemented on the premise of the technical scheme of the present invention, and a detailed implementation manner and a specific operation process are given, but the protection scope of the present invention is not limited to the following examples.
The embodiment provides an unsupervised skeleton action recognition method based on a cyclic graph convolution automatic encoder, which is used for solving the problem that the existing unsupervised action recognition method ignores the spatial dependency relationship of an action sequence and improving the accuracy of action recognition.
As shown in the straight line flow in fig. 1, the specific steps of this embodiment are as follows:
step S1, preprocessing a human skeleton action sequence and inputting the human skeleton action sequence into a cyclic graph convolution encoder.
And S2, outputting and obtaining a characterization vector of the action sequence by the cyclic graph convolution encoder. Wherein the cyclic graph convolution encoder comprises: the multi-layer space joint attention module is used for combining the human skeleton action sequence and a hidden layer of the cyclic graph convolution encoder, and adaptively measuring the importance of different joints with different actions to obtain a weighted skeleton sequence; and the multi-layer graph convolution gating circulating unit layer (graph convolution GRU layer) is used for integrating the connection relation characteristics of the weighted skeleton sequences to obtain the characterization vector of the action sequence. The space joint attention module and the picture convolution gating circulation unit layer are preferably three layers.
And step S3, calculating the characterization vector of the action sequence through a weighted nearest neighbor classification algorithm to obtain the identification category of the human skeleton action sequence, and completing the action identification flow.
The training steps of the cyclic graph rolling encoder are as described in the dashed flow in fig. 1, and the training is performed by adopting a gradient descent method as a whole, specifically as follows:
and A1, inputting a training action sequence set into a cyclic graph convolution encoder so as to obtain a characterization vector of the action sequence.
A2, inputting the characterization vector and the hidden layer vector of the action sequence to a decoder for sequence restoration to obtain a reconstructed action sequence set;
and A3, comparing the reconstructed action sequence set with the training action sequence set, and calculating a loss function value through a reconstruction loss function.
And repeating the steps A1 to A3 until the loss function value reaches a preset cut-off condition.
In the training process, the expression of the reconstruction loss function of the training is as follows:
in the method, in the process of the invention,representing training action sequence set,/->Representing a set of reconstructed action sequences, I.I F Representing the Frobenius norm and L representing the loss function value.
Next, the present embodiment will be described in detail in sections.
1. The spatial joint attention module is shown in fig. 2. The space joint attention module is used for combining the human skeleton action sequence x t And hidden layer h of a cyclic graph convolution encoder t-1 The importance of different joints with different actions is adaptively measured to obtain a weighted skeleton sequence x' t . Computing a weighted skeleton sequence x' t The method comprises the following steps:
first, an importance score s for each joint is calculated t The calculation expression is as follows:
s t ＝U s φ(W x x t +W h h t-1 +b s )+b u
in the method, in the process of the invention,representing the sequence coordinates of N joints at time t, h t-1 Representing hidden layer information, W x And W is h Represents a matrix of learnable parameters, phi represents an activation function, b s And b u Representing the bias.
Then, the importance alpha of each joint is calculated t The calculation expression is as follows:
finally, a weighted skeleton sequence x 'is calculated' t The calculation expression is as follows:
x′ t ＝(α t +1)·x t
wherein, representative point multiplication.
2. The multi-layer graph convolution gating loop cell layer is shown in fig. 3. The multi-layer graph convolution gating circulating unit layer is used for integrating the connection relation characteristics of the weighted skeleton sequences, fully utilizing the space dependency relation between joints of each frame, and simultaneously retaining the characteristics of time dimension to obtain the characterization vector of the action sequence.
In the graph convolution gating circulating unit layer, the expression of the connection relation characteristic of the integrated weighted skeleton sequence is as follows:
wherein H is (l+1) Indicating that the output of layer l +1 is convolved,representing a symmetrical adjacency matrix with spin, A represents the adjacency matrix of the diagram, I represents the identity matrix,/>For the degree matrix, τ represents the activation function, H (l) Representing the output of layer I of the graph convolution, Θ (l) Representing a matrix of learnable parameters of the first layer.
The graph convolution gating and circulating unit layer combines graph convolution and gating and circulating units, and the expression is as follows:
wherein z is t Representing an update gate, r t Representing a reset gate and,activation vector representing candidate->Representative graph convolution sum H (l +1) Correspondingly, W xz 、W hz 、W xr 、W hr 、W xh And W is hh Representing the parameter matrix in different gating, +..
3. In the training step of the cyclic graph convolutional encoder, the decoder input is a representation vector of the motion sequence and an implicit layer vector, wherein in order to make the decoder depend completely on the state delivered by the encoder, the encoder is forced to learn better feature representation, the implicit layer vector is adopted as all 0 and x t Vectors of the same size.
4. In the embodiment, a weighted nearest neighbor classification algorithm is adopted as a classifier, and the identification category of the human skeleton action sequence is obtained from the characterization vector of the action sequence. Specifically, after the first k nearest samples are obtained, the number of votes of each category is calculated, and a recognition result is obtained through weighted voting. In this example k=9. The calculation expression of the weight is as follows:
wherein w is i And d i The voting weights and cosine distances of the samples i are shown, respectively.
In order to support and verify the performance of the motion recognition method provided by the invention, the embodiment adopts recognition accuracy as an evaluation index on three public standard data sets, and compares the invention with other conventional unsupervised skeleton motion recognition methods, including LongT GAN (Long-Term Dynamics GAN, long-term dynamic generation of an antagonistic network), P & C (predictive & Cluster), MS2L (Multi-Task Self-Supervised Learning, multi-Task Self-supervision learning).
Table 1 shows the comparison of recognition accuracy of the invention with other non-supervision motion recognition methods based on the skeleton on the NTU-RGB+D60 data set. Wherein CS (Cross-Subject) and CV (Cross-View) represent two different test methods for the dataset, wherein CS refers to dividing the training set and the test set according to different volunteers collecting data, and CV refers to dividing the training set and the test set according to the results of different View cameras used for collecting data.
Table 1 comparison of recognition accuracy (%) on NTU-RGB+D60 dataset
As can be seen from table 1, the unsupervised human motion recognition method based on the cyclic graph convolution automatic encoder provided by the invention is superior to the existing method in two test methods CS and CV of NTU-rgb+d60 data sets, and is respectively higher than the existing methods by 1.8 and 2.9 percentage points.
Table 2 shows comparison of recognition accuracy of the invention with other non-supervision motion recognition methods based on the skeleton on the NW-UCLA data set.
Table 2 NW-comparison of recognition accuracy (%) on UCLA dataset
As can be seen from table 2, although the existing method has achieved an accuracy of 80% or more on the NW-UCLA dataset, the method proposed by the present invention can still further improve the recognition accuracy.
Table 3 shows comparison of recognition accuracy of the invention with other non-supervision motion recognition methods based on the skeleton on the UWA3D data set. Wherein V3 and V4 represent two test methods on the UWA3D dataset.
Table 3 comparison of recognition accuracy (%) on UWA3D dataset
As can be seen from table 3, compared with other non-supervision action recognition methods based on the skeleton, the method has more excellent recognition accuracy in all the test methods on the UWA3D data set, and improves the recognition accuracy by 2.4% under the V4 test condition. The embodiments on the three data sets together illustrate that the unsupervised human skeleton action recognition method based on the cyclic graph convolution automatic encoder provided by the invention can stably obtain excellent recognition accuracy under different test conditions in different data sets.
The foregoing describes in detail preferred embodiments of the present invention. It should be understood that numerous modifications and variations can be made in accordance with the concepts of the invention by one of ordinary skill in the art without undue burden. Therefore, all technical solutions which can be obtained by logic analysis, reasoning or limited experiments based on the prior art by the person skilled in the art according to the inventive concept shall be within the scope of protection defined by the claims.
Claims (7)
1. An unsupervised skeleton action recognition method based on a cyclic graph convolution automatic encoder is characterized by comprising the following steps:
s1, inputting a human skeleton action sequence to a cyclic graph convolution encoder;
s2, outputting and obtaining a characterization vector of the action sequence by a cyclic graph convolution encoder;
s3, calculating a characterization vector of the action sequence through a weighted nearest neighbor classification algorithm to obtain an identification category of the human skeleton action sequence;
the cyclic graph convolution encoder includes: the multi-layer space joint attention module is used for combining the human skeleton action sequence and a hidden layer of the cyclic graph convolution encoder, and adaptively measuring the importance of different joints with different actions to obtain a weighted skeleton sequence; the multi-layer diagram convolution gating circulating unit layer is used for integrating the connection relation characteristics of the weighted skeleton sequences to obtain the characterization vector of the action sequence;
in the spatial joint attention module, the weighted skeleton sequence calculation expression is:
x′ t ＝(α t +1)·x t
s t ＝U s φ(W x x t +W h h t-1 +b t )+b u
wherein x 'is' t Representing a weighted skeleton sequence, alpha t Representing the importance of each joint s t Representing the importance score of each joint,representing the sequence coordinates of N joints at time t, h t-1 Representing hidden layer information, W x And W is h Represents a matrix of learnable parameters, phi represents an activation function, b s And b u Representing the bias;
in the graph convolution gating circulating unit layer, the expression of the connection relation characteristic of the integrated weighted skeleton sequence is as follows:
wherein H is (l+1) Indicating that the output of layer l +1 is convolved,representing a symmetric adjacency matrix with spin, A represents the adjacency matrixI represents an identity matrix,>for the degree matrix, τ represents the activation function, H (l) Representing the output of layer I of the graph convolution, Θ (l) A matrix of learnable parameters representing the first layer;
the representation of the graph convolution gating loop unit layer is as follows:
wherein z is t Representing an update gate, r t Representing a reset gate and,activation vector representing candidate->Representative graph convolution sum H (l+1) Correspondingly, W xz 、W hz 、W xr 、W hr 、W xh And W is hh Representing the parameter matrix in different gating, +..
2. The method for recognizing actions of an unsupervised skeleton based on a cyclic graph convolution automatic encoder according to claim 1, wherein the training step of the cyclic graph convolution encoder comprises
A1, inputting a training action sequence set to a cyclic graph convolution encoder so as to obtain a characterization vector of an action sequence;
a2, inputting the characterization vector and the hidden layer vector of the action sequence to a decoder, and performing sequence restoration to obtain a reconstructed action sequence set;
a3, comparing the reconstructed action sequence set with the training action sequence set, and calculating a loss function value through a reconstruction loss function;
repeating the steps A1 to A3 until the loss function value reaches the preset cut-off condition.
3. The method for recognizing the actions of the unsupervised skeleton based on the automatic loop-chart convolution encoder according to claim 2, wherein the hidden layer vector is a vector with a value of zero and the same length as the action sequence of the human skeleton.
4. The method for recognizing the actions of the unsupervised skeleton based on the automatic loop-chart convolution encoder according to claim 2, wherein the expression of the reconstruction loss function is:
in the method, in the process of the invention,representing training action sequence set,/->Representing a set of reconstructed action sequences, I.I F Representing the Frobenius norm and L representing the loss function value.
5. The method for recognizing actions of an unsupervised skeleton based on a cyclic graph convolution automatic encoder according to claim 2, wherein the cyclic graph convolution encoder is trained by a gradient descent method.
6. The method for recognizing actions of an unsupervised skeleton based on a loop-based convolution automatic encoder according to claim 1, wherein the spatial joint attention module and the loop-based convolution gating loop unit layer are three layers.
7. The method for recognizing the actions of the unsupervised skeleton based on the automatic loop-chart convolution encoder according to claim 1, wherein in the weighted nearest neighbor classification algorithm, k is a set value after k nearest samples are obtained, the number of votes in each category is calculated, and recognition results are obtained through weighted voting, wherein the calculation expression of the weight is:
wherein w is i And d i The voting weights and cosine distances of the samples i are shown, respectively.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110908006.9A CN113762082B (en) | 2021-08-09 | 2021-08-09 | Unsupervised skeleton action recognition method based on cyclic graph convolution automatic encoder |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110908006.9A CN113762082B (en) | 2021-08-09 | 2021-08-09 | Unsupervised skeleton action recognition method based on cyclic graph convolution automatic encoder |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113762082A CN113762082A (en) | 2021-12-07 |
| CN113762082B true CN113762082B (en) | 2024-02-27 |
Family
ID=78788720
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110908006.9A Active CN113762082B (en) | 2021-08-09 | 2021-08-09 | Unsupervised skeleton action recognition method based on cyclic graph convolution automatic encoder |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113762082B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112597883A (en) * | 2020-12-22 | 2021-04-02 | 武汉大学 | Human skeleton action recognition method based on generalized graph convolution and reinforcement learning |
| CN112733656A (en) * | 2020-12-30 | 2021-04-30 | 杭州电子科技大学 | Skeleton action identification method based on multi-stream space attention diagram convolution SRU network |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11449537B2 (en) * | 2018-12-18 | 2022-09-20 | Adobe Inc. | Detecting affective characteristics of text with gated convolutional encoder-decoder framework |
-
2021
- 2021-08-09 CN CN202110908006.9A patent/CN113762082B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112597883A (en) * | 2020-12-22 | 2021-04-02 | 武汉大学 | Human skeleton action recognition method based on generalized graph convolution and reinforcement learning |
| CN112733656A (en) * | 2020-12-30 | 2021-04-30 | 杭州电子科技大学 | Skeleton action identification method based on multi-stream space attention diagram convolution SRU network |
Non-Patent Citations (1)
| Title |
|---|
| 多模型融合动作识别研究;田曼;张艺;;电子测量技术(第20期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113762082A (en) | 2021-12-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Xia et al. | Multi-stage feature constraints learning for age estimation | |
| CN110728209B (en) | Gesture recognition method and device, electronic equipment and storage medium | |
| CN112800903B (en) | Dynamic expression recognition method and system based on space-time diagram convolutional neural network | |
| CN109101865A (en) | A kind of recognition methods again of the pedestrian based on deep learning | |
| CN105069413A (en) | Human body gesture identification method based on depth convolution neural network | |
| CN106909938B (en) | Visual angle independence behavior identification method based on deep learning network | |
| CN109508686B (en) | Human behavior recognition method based on hierarchical feature subspace learning | |
| CN106548194B (en) | The construction method and localization method of two dimensional image human joint points location model | |
| CN113313123B (en) | Glance path prediction method based on semantic inference | |
| CN111259735B (en) | Single-person attitude estimation method based on multi-stage prediction feature enhanced convolutional neural network | |
| CN116524593A (en) | Dynamic gesture recognition method, system, equipment and medium | |
| CN111723667A (en) | Human body joint point coordinate-based intelligent lamp pole crowd behavior identification method and device | |
| CN116091551B (en) | Target retrieval tracking method and system based on multi-mode fusion | |
| CN105930793B (en) | A kind of human body detecting method based on the study of SAE feature visualization | |
| CN112906520A (en) | Gesture coding-based action recognition method and device | |
| CN117671787A (en) | Rehabilitation action evaluation method based on transducer | |
| CN117854155B (en) | Human skeleton action recognition method and system | |
| CN116246338A (en) | Behavior recognition method based on graph convolution and transducer composite neural network | |
| CN110135435B (en) | Saliency detection method and device based on breadth learning system | |
| CN113780129B (en) | Action recognition method based on unsupervised graph sequence predictive coding and storage medium | |
| Niu et al. | An indoor pool drowning risk detection method based on improved YOLOv4 | |
| Mohana et al. | Emotion recognition from facial expression using hybrid CNN–LSTM network | |
| Waheed et al. | An automated human action recognition and classification framework using deep learning | |
| Abedi et al. | Modification of deep learning technique for face expressions and body postures recognitions | |
| CN113158756A (en) | Posture and behavior analysis module and method based on HRNet deep learning |
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 |