CN116486341A - Training and identifying method and device for human body behavior identification model based on RFID - Google Patents
Training and identifying method and device for human body behavior identification model based on RFID Download PDFInfo
- Publication number
- CN116486341A CN116486341A CN202310473044.5A CN202310473044A CN116486341A CN 116486341 A CN116486341 A CN 116486341A CN 202310473044 A CN202310473044 A CN 202310473044A CN 116486341 A CN116486341 A CN 116486341A
- Authority
- CN
- China
- Prior art keywords
- domain
- target
- phase value
- behavior recognition
- matrix
- 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
- 238000012549 training Methods 0.000 title claims abstract description 83
- 238000000034 method Methods 0.000 title claims abstract description 56
- 239000011159 matrix material Substances 0.000 claims abstract description 91
- 239000013598 vector Substances 0.000 claims abstract description 40
- 238000007781 pre-processing Methods 0.000 claims abstract description 22
- 238000004364 calculation method Methods 0.000 claims abstract description 18
- 238000001228 spectrum Methods 0.000 claims abstract description 12
- 230000006399 behavior Effects 0.000 claims description 126
- 238000000605 extraction Methods 0.000 claims description 19
- 238000004590 computer program Methods 0.000 claims description 14
- 238000012545 processing Methods 0.000 claims description 9
- 238000001914 filtration Methods 0.000 claims description 5
- 238000010606 normalization Methods 0.000 claims description 5
- 238000012952 Resampling Methods 0.000 claims description 4
- 238000009499 grossing Methods 0.000 claims description 4
- 238000005457 optimization Methods 0.000 claims description 4
- 239000010410 layer Substances 0.000 description 14
- 238000010586 diagram Methods 0.000 description 12
- 230000009471 action Effects 0.000 description 7
- 230000006870 function Effects 0.000 description 7
- 238000004891 communication Methods 0.000 description 6
- 230000008569 process Effects 0.000 description 6
- 230000003287 optical effect Effects 0.000 description 4
- 230000006978 adaptation Effects 0.000 description 2
- 238000004422 calculation algorithm Methods 0.000 description 2
- 239000002355 dual-layer Substances 0.000 description 2
- 239000004973 liquid crystal related substance Substances 0.000 description 2
- 230000000644 propagated effect Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 239000004065 semiconductor Substances 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000002238 attenuated effect Effects 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 210000003127 knee Anatomy 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 230000005012 migration Effects 0.000 description 1
- 238000013508 migration 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
- 239000013307 optical fiber Substances 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000002441 reversible effect Effects 0.000 description 1
- 230000003595 spectral effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06K—GRAPHICAL DATA READING; PRESENTATION OF DATA; RECORD CARRIERS; HANDLING RECORD CARRIERS
- G06K17/00—Methods or arrangements for effecting co-operative working between equipments covered by two or more of main groups G06K1/00 - G06K15/00, e.g. automatic card files incorporating conveying and reading operations
- G06K17/0022—Methods or arrangements for effecting co-operative working between equipments covered by two or more of main groups G06K1/00 - G06K15/00, e.g. automatic card files incorporating conveying and reading operations arrangements or provisious for transferring data to distant stations, e.g. from a sensing device
- G06K17/0025—Methods or arrangements for effecting co-operative working between equipments covered by two or more of main groups G06K1/00 - G06K15/00, e.g. automatic card files incorporating conveying and reading operations arrangements or provisious for transferring data to distant stations, e.g. from a sensing device the arrangement consisting of a wireless interrogation device in combination with a device for optically marking the record carrier
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
- G06V10/774—Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting
-
- 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
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/103—Static body considered as a whole, e.g. static pedestrian or occupant recognition
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D30/00—Reducing energy consumption in communication networks
- Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Human Computer Interaction (AREA)
- Computing Systems (AREA)
- Databases & Information Systems (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Computer Networks & Wireless Communication (AREA)
- General Engineering & Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The embodiment of the application provides a training and identifying method and device for a human behavior identification model based on RFID. The training method comprises the following steps: preprocessing a signal receiving intensity indication sequence and a phase value sequence in a target domain training data set respectively to obtain a corresponding signal receiving intensity indication matrix and a phase value matrix, and inputting the corresponding signal receiving intensity indication matrix and the phase value matrix into a human behavior recognition model which is pre-trained by a source domain training data set to output a corresponding weighted target feature vector; inputting the weighted target feature vector to a domain discriminator to output a corresponding domain discriminating label; and performing batch spectrum punishment calculation, and optimizing the human body behavior recognition model according to the calculation result and the domain discrimination label to obtain a target human body behavior recognition model. According to the technical scheme, the influence of the characteristics which cannot be transferred and the characteristics in the specific field can be eliminated, and the human behavior recognition accuracy is improved.
Description
Technical Field
The application relates to the technical field of computers, in particular to a training and identifying method and device of a human behavior identification model based on RFID.
Background
The human behavior recognition application is wide, and is a hotspot problem of research in the field of artificial intelligence, in particular to a basic technology for various applications such as intelligent monitoring, man-machine interaction robots and the like. In the current technical scheme, the depth network has rich characteristic representation capability, human behavior recognition is usually carried out by adopting the depth network, but most of human behavior recognition researches based on RFID are concentrated on recognition in a specific field, and when the problem of cross-field is solved, the depth network is matched with global characteristics to carry out field adaptation, and consideration of non-transferable characteristics is lacking, so that the recognition accuracy is reduced.
Disclosure of Invention
The embodiment of the application provides a training and identifying method and device for a human body behavior identification model based on RFID, so that the influence of non-transferable characteristics and specific field characteristics can be eliminated at least to a certain extent, and the human body behavior identification precision is improved.
Other features and advantages of the present application will be apparent from the following detailed description, or may be learned in part by the practice of the application.
According to an aspect of the embodiments of the present application, there is provided a training method of a human behavior recognition model based on RFID, including:
Acquiring a target domain training data set, wherein the target domain training data set comprises a signal receiving intensity indication sequence and a phase value sequence of a received tag signal when different types of human behaviors are performed;
preprocessing the signal receiving intensity indication sequence and the phase value sequence respectively to obtain a corresponding signal receiving intensity indication matrix and a phase value matrix;
inputting the signal receiving intensity indication matrix and the phase value matrix into a human body behavior recognition model pre-trained by a source domain training data set so that the human body behavior recognition model outputs a corresponding weighted target feature vector, wherein the human body behavior recognition model comprises a feature extraction module and a self-attention module which are connected;
inputting the weighted target feature vector to a domain discriminator so that the domain discriminator outputs corresponding domain discrimination tags respectively, wherein the domain discrimination tags are used for determining whether the data come from a source domain or a target domain;
and performing batch spectrum punishment calculation according to target feature vectors respectively from a target domain and a source domain, and optimizing the human body behavior recognition model according to a calculation result and the domain discrimination label to obtain a target human body behavior recognition model.
According to an aspect of the embodiment of the application, there is provided an RFID-based human behavior recognition method, which is applied to a terminal device, where the terminal device is connected to a reader, and the reader is configured to receive tag signals of a plurality of target tags, where the plurality of target tags are respectively disposed at predetermined positions of a collection object;
the method comprises the following steps:
a signal reception intensity indication sequence and a phase value sequence when receiving a tag signal transmitted by the reader;
preprocessing is carried out according to the signal receiving intensity indication sequence and the phase value sequence respectively to obtain a corresponding signal receiving intensity indication matrix and a phase value matrix;
inputting the signal receiving intensity indication matrix and the phase value matrix into a target human body behavior recognition model, so that the target human body behavior recognition model outputs a corresponding behavior discrimination result, and the target human body behavior recognition model is trained by the training method described in the previous embodiment.
According to an aspect of the embodiments of the present application, there is provided a training device for a human behavior recognition model based on RFID, including:
the acquisition module is used for acquiring a target domain training data set, wherein the target domain training data set comprises a signal receiving intensity indication sequence and a phase value sequence of a received tag signal when different types of human behaviors are performed;
The first preprocessing module is used for preprocessing the signal receiving intensity indication sequence and the phase value sequence respectively to obtain a corresponding signal receiving intensity indication matrix and a phase value matrix;
the first input module is used for inputting the signal receiving intensity indication matrix and the phase value matrix into a human body behavior recognition model pre-trained by a source domain training data set so that the human body behavior recognition model outputs a corresponding weighted target feature vector, and the human body behavior recognition model comprises a feature extraction module and a self-attention module which are connected;
the second input module is used for inputting the weighted target feature vector to a domain discriminator so as to enable the domain discriminator to output a corresponding domain discrimination tag, wherein the domain discrimination tag is used for determining whether the data come from a source domain or a target domain;
and the processing module is used for carrying out batch spectrum punishment calculation according to target feature vectors respectively from a target domain and a source domain, and carrying out optimization on the human body behavior recognition model according to a calculation result and the domain discrimination label so as to obtain a target human body behavior recognition model.
According to an aspect of the embodiment of the application, there is provided an RFID-based human behavior recognition device, which is applied to a terminal device, where the terminal device is connected to a reader, and the reader is configured to receive tag signals of a plurality of target tags, where the plurality of target tags are respectively disposed at predetermined positions of a collection object;
The device comprises:
a receiving module for receiving a signal reception intensity indication sequence and a phase value sequence transmitted by the reader when receiving a tag signal;
the second preprocessing module is used for preprocessing the signal receiving intensity indication sequence and the phase value sequence respectively to obtain a corresponding signal receiving intensity indication matrix and a phase value matrix;
the recognition module is used for inputting the signal receiving intensity indication matrix and the phase value matrix into a target human body behavior recognition model so that the target human body behavior recognition model outputs a corresponding behavior discrimination result, and the target human body behavior recognition model is trained by the training method described in the previous embodiment.
According to an aspect of the embodiments of the present application, there is provided a computer readable medium having stored thereon a computer program which, when executed by a processor, implements a method as described in the above embodiments.
According to an aspect of an embodiment of the present application, there is provided an electronic device including: one or more processors; and storage means for storing one or more programs which, when executed by the one or more processors, cause the one or more processors to implement the method as described in the above embodiments.
According to an aspect of embodiments of the present application, there is provided a computer program product or computer program comprising computer instructions stored in a computer readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device performs the method provided in the above-described embodiment.
In some embodiments of the present application, a target domain training data set is obtained, where the target domain training data set includes a signal receiving intensity indication sequence and a phase value sequence of a tag signal received during different types of human behaviors, the signal receiving intensity indication sequence and the phase value sequence are preprocessed respectively to obtain a corresponding signal receiving intensity indication matrix and a corresponding phase value matrix, the signal receiving intensity indication matrix and the phase value matrix are input to a human behavior recognition model pre-trained by a source domain training data set, so that the human behavior recognition model outputs a corresponding weighted target feature vector, the human behavior recognition model includes a feature extraction module and a self-attention module that are connected, the weighted target feature vector is input to a domain discriminator, so that the domain discriminator outputs a corresponding domain discrimination tag, the domain discrimination tag is used to determine that data is from the source domain or the target domain, and performs batch spectrum calculation according to the target feature vector from the target domain and the source domain, and then performs optimization on the human behavior recognition model according to a calculation result and the domain discrimination tag, so as to obtain a target human behavior recognition model. Therefore, the target human behavior recognition model can adapt to transferable attention and resistance learning, so that the influence of non-transferable characteristics and specific field characteristics can be eliminated, and the recognition accuracy of human behavior recognition is improved.
It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the application.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the application and together with the description, serve to explain the principles of the application. It is apparent that the drawings in the following description are only some embodiments of the present application, and that other drawings may be obtained from these drawings without inventive effort for a person of ordinary skill in the art. In the drawings:
FIG. 1 illustrates a flow diagram of a training method for RFID-based human behavior recognition models, according to one embodiment of the present application;
FIG. 2 illustrates a block diagram of a training device for RFID-based human behavior recognition models, according to one embodiment of the present application;
fig. 3 shows a schematic diagram of a computer system suitable for use in implementing the electronic device of the embodiments of the present application.
Detailed Description
Example embodiments will now be described more fully with reference to the accompanying drawings. However, the exemplary embodiments may be embodied in many forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the concept of the example embodiments to those skilled in the art.
Furthermore, the described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments. In the following description, numerous specific details are provided to give a thorough understanding of embodiments of the present application. One skilled in the relevant art will recognize, however, that the aspects of the application can be practiced without one or more of the specific details, or with other methods, components, devices, steps, etc. In other instances, well-known methods, devices, implementations, or operations are not shown or described in detail to avoid obscuring aspects of the application.
The block diagrams depicted in the figures are merely functional entities and do not necessarily correspond to physically separate entities. That is, the functional entities may be implemented in software, or in one or more hardware modules or integrated circuits, or in different networks and/or processor devices and/or microcontroller devices.
The flow diagrams depicted in the figures are exemplary only, and do not necessarily include all of the elements and operations/steps, nor must they be performed in the order described. For example, some operations/steps may be decomposed, and some operations/steps may be combined or partially combined, so that the order of actual execution may be changed according to actual situations.
FIG. 1 shows a flow diagram of a training method for RFID-based human behavior recognition models, according to one embodiment of the present application. The method may be applied to a terminal device or server, which may include, but is not limited to, one or more of a smart phone, a tablet computer, a laptop computer, and a desktop computer. The following description will take application of the training method to a terminal device as an example.
As shown in fig. 1, the training method at least includes steps S110 to S150, which are described in detail as follows:
in step S110, a target domain training data set is acquired, the target domain training data set including a signal reception intensity indication sequence and a phase value sequence of a received tag signal when different kinds of human behaviors are performed.
The target domain training data set may be a data set to be migrated of the depth model, and the target domain training data set and the source domain training data set may have a certain difference, where the difference may be caused by different data attributes, different data distribution or different acquisition modes, and other reasons. Unlike the source domain training dataset, the target domain training dataset does not have a corresponding behavior class label, i.e., it does not need to be manually labeled, thereby reducing the data acquisition cost.
The application also provides a data acquisition system, which comprises a terminal device, an RFID reader and a plurality of RFID tags, wherein the terminal device is taken as a desktop computer for example, the desktop computer is in communication connection with the RFID reader, the RFID reader can receive tag signals of the RFID tags, and the RFID tags can be respectively arranged at each key point of an acquisition object, such as a head, a shoulder, a hand, a knee, a foot and the like.
It should be noted that, the signal receiving strength indication (Received Signal Strength Indicator, RSSI) of the tag signal received by the RFID reader may be attenuated as the distance from the RFID tag increases, and the phase value may also be changed as the propagation path changes, so when the human body performs a specific action, both the signal receiving strength indication and the phase value received by the RFID reader may show a specific rule, so the signal receiving strength indication and the phase value may provide useful position and action information, and have lower cost and complexity.
In particular, the data acquisition system employs a single RFID reader in order to prevent collision problems that may be caused by signals of the RFID readers interfering with each other, thereby resulting in reduced communication efficiency. The RFID reader can record the signal receiving intensity indication and the phase value in each action process and transmit the signal receiving intensity indication and the phase value to a desktop computer for subsequent processing. The desktop computer can record real-time signal receiving intensity indication and phase value of each tag, and when acquiring the source domain training data set, the acquisition personnel can manually intercept the signal receiving intensity indication sequence and the phase value sequence when the action occurs and mark the signal receiving intensity indication sequence and the phase value sequence so as to determine the behavior type tag corresponding to the signal receiving intensity indication sequence and the phase value sequence, for example, the behavior type tag can comprise but is not limited to standing, walking, bowing, swinging, waving and the like. When the target domain training data set is acquired, the target domain training data set does not need to be marked.
Therefore, according to the data acquisition system, an acquisition person can acquire a source domain training data set and a target domain training data set respectively, and the source domain training data set and the target domain training data set are stored in a local storage space for later use.
In this step S110, the terminal device may acquire the target domain training data set from the local storage space, or may acquire the target domain training data set in real time, which is not limited in particular.
In step S120, the signal reception intensity indication sequence and the phase value sequence are preprocessed respectively, so as to obtain a corresponding signal reception intensity indication matrix and a corresponding phase value matrix.
In one embodiment of the present application, preprocessing may include, but is not limited to, at least one of resampling, phase unwrapping, filtering and smoothing, and normalization.
Specifically, for resampling, the length of the data sequence corresponding to each action may be preset, for example, the length of the data sequence is 5s, and linear interpolation is performed on each data sequence to achieve that the data sequence is 5Hz. It should be noted that the above numbers are merely exemplary distances, and those skilled in the art may determine the corresponding data sequence length and frequency according to actual implementation needs, which are not limited in particular.
For phase unwrapping, it is only for the sequence of phase values, so that the periodicity of the phase values can be removed, so that the phase values become continuous and interpretable.
For filtering and smoothing, the existing filter can be used for noise reduction and filtering of the data sequence processed by the steps, for example, a Savitzky-Golay filter can be used.
For the signal normalization process, it may normalize the data sequence of each RFID tag.
Thus, through the above preprocessing step, an n×m signal reception intensity indication matrix M can be obtained rssi And a matrix of phase values M phase Where n is the number of tags, m is the number of signal received strength indicators or phase values for a predetermined length of time after interpolation, e.g. the predetermined length of time is 5s, the frequency is 5Hz, then m is 25, etc.
In step S130, the signal receiving intensity indication matrix and the phase value matrix are input to a human body behavior recognition model pre-trained by a source domain training data set, so that the human body behavior recognition model outputs a corresponding weighted target feature vector, and the human body behavior recognition model includes a connected feature extraction module and a self-attention module.
The human behavior recognition model may be a depth model for human behavior recognition, and the human behavior recognition model may include a connected feature extraction module and a self-attention module.
In this embodiment, the terminal device may input the signal receiving intensity indication matrix and the phase value matrix into the training data set of the source domain and the trained human behavior recognition model, where the feature extraction module may convert the signal receiving intensity indication matrix and the phase value matrix into a space-time feature stream (i.e. a third feature vector described later) with higher hierarchy and richer information, and then the self-attention module weights the features of the space-time feature stream, so that the influence of the non-transferable features may be effectively reduced, and the self-attention mechanism retains the original input tensor, so that the information loss may be reduced, and the robustness of the model may be improved.
In one embodiment of the present application, the feature extraction module performs feature extraction on the signal reception intensity indication matrix and the phase value matrix, including:
respectively inputting the signal receiving intensity indication matrix and the phase value matrix into a plurality of convolution layers to obtain corresponding first feature vectors;
And splicing the first eigenvector corresponding to the signal receiving intensity indication matrix and the first eigenvector corresponding to the phase value matrix to obtain a second eigenvector, and inputting the second eigenvector to a double-layer GRU so that the double-layer GRU outputs a corresponding third eigenvector as the input of the self-attention module.
In this embodiment, the feature extraction module may pass the signal received strength indication matrix and the phase value matrix through respective multi-layer convolution layers, preferably 3 layers, each employing a convolution kernel of 1*3. In addition, in the convolution process, batch normalization processing can be adopted, so that the condition of overfitting in the training process is prevented, and it is understood that the multi-layer convolution layer can effectively extract the space information in the signal.
Then, the first feature vectors output by the two multi-layer convolution layers may be spliced to obtain a spliced second feature vector, and the second feature vector is input into a dual-layer GRU (gate-controlled loop unit, gated recurrent unit) to extract timing information through the dual-layer GRU, and output a third feature vector as an input of the self-attention module.
In an embodiment, the third feature vector is input to the self-attention module, and the self-attention module may pass the third feature vector through a linear layer to obtain Q, K, V terms, and obtain an attention map, where the formula is as follows:
wherein d k Is the dimension of K.
And multiplying V by the attention force and adding the original input x (i.e., the third feature vector) to obtain the final Output (i.e., the target feature vector) of the self-attention module, where the formula is as follows:
Output(x，Q，K，V)＝x+AttentionMap(Q，K)。
thus, by weighting the features extracted by the feature extraction module using the self-attention module, the impact of the non-transferable features can be effectively mitigated.
Referring to fig. 1, in step S140, the weighted target feature vectors are input to a domain arbiter, so that the domain arbiter outputs corresponding domain discrimination tags respectively, where the domain discrimination tags are used to determine whether the data is from a source domain or a target domain.
In this embodiment, the terminal device may input the weighted target feature vector to the domain arbiter so that the domain arbiter may output a domain discrimination tag for determining whether the data is from the source domain or the target domain.
It can be understood that the source domain training data set is a training data set with a large number of behavior class labels, the human body behavior recognition model can train out a parameter which performs well in the training data set through the source domain training data set, but the human body behavior recognition model trained through the source domain training data set is directly used for predicting the target domain training data set, so that a good result cannot be obtained, because the human body behavior recognition model may excessively fit the data of the source domain, the characteristics of the data of the target domain cannot be extracted well, but because the behavior class labels of the target domain are not available, the human body behavior recognition model cannot be trained directly in the target domain.
To this end, a domain arbiter is introduced, which in one example comprises a user domain arbiter for determining whether the currently input target feature vector is the same acquisition object as the source domain training dataset, and an environment domain arbiter for determining whether the currently input target feature vector is the same acquisition environment as the source domain training dataset. Therefore, under the condition that the target domain does not have the corresponding behavior category label, the user domain label and the environment domain label can be marked for the data of the source domain and the target domain simultaneously through the user domain label and the environment domain label, and at the moment, each data comprising the source domain and the target domain has two additional labels which are different. Thus, under such training, the human behavior recognition model can gradually grasp the characteristics of the data from the source domain and the target domain at the same time to help achieve better results in the classification task.
In step S150, batch spectrum penalty calculation is performed according to the target feature vectors from the target domain and the source domain, and the human behavior recognition model is optimized according to the calculation result and the domain discrimination label, so as to obtain the target human behavior recognition model.
In this embodiment, a batch spectrum penalty calculation may be performed according to the target feature vectors from the target domain and the source domain, where the batch spectrum penalty is composed of singular value decomposition, so that the maximum singular value can be obtained from the source feature matrix and the target feature matrix to determine a loss of the corresponding batch spectrum penalty, and the amplitude of the feature matrix is adjusted to slow down the overfitting of the human behavior recognition model, so as to improve the generalization capability of the human behavior recognition model.
And, based on the domain discrimination tag, cross entropy loss between the domain discrimination tag and the real tag can be determined, and based on the loss, gradient of each parameter in the model can be calculated by using a back propagation algorithm for updating.
Therefore, according to the calculation result of the batch spectrum punishment and the domain discrimination label, the human body behavior recognition model can be optimized, namely, the accuracy of the recognition result of the human body behavior recognition model can be improved by adjusting the parameters of the human body behavior recognition model, so that the target human body behavior recognition model can be obtained.
In another embodiment of the present application, a training method of a human behavior recognition model is provided, the method including the steps of:
s1: the data collected by one environment and volunteers is set as source domain data, and the data collected by other environments or other volunteers is set as target domain data.
S2: the source domain data or the target domain data is processed by a feature extractor and a self-attention module to obtain f' source Or f' target 。
S3: will f' source Obtaining predicted behavior labels through a behavior predictor, and calculating losses L with the actual behavior labels y The loss uses cross entropy loss.
S4: a pair of f' source And f' target The user domain label and the environment domain label are obtained through the user domain discriminator and the environment domain discriminator respectively, and are compared with the real user domain label or the environment domain label to obtain lossAnd->Then summing to obtain L d ，
S5: a pair of f' source And f' target For calculating a batch spectral penalty,
specifically, singular value matrices Σ are calculated for both respectively source Sum sigma target Then each take out their ith big value beta s，i And beta t，i， Summing the values of k large before them to obtain BSP loss
S6: the total loss of the model is
L＝L y -αL d +βL bsp
Alpha and beta are hyper-parameters that are used to balance losses in the domain discriminators and the attention migration.
Based on this total loss, the gradients of the various parameters in the model can be calculated using a back propagation algorithm to update the gradients to complete the training.
Therefore, the human behavior recognition model obtained through training by the training method can eliminate the influence of the characteristics which cannot be transferred and the characteristics in the specific field, and improve the human behavior recognition precision.
In one embodiment of the application, there is further provided an RFID-based human behavior recognition method, which may be applied to a terminal device, where the terminal device is connected to a reader, and the reader is configured to receive tag signals of a plurality of target tags, where the plurality of target tags are respectively disposed at predetermined positions of a collection object;
the method comprises the following steps:
a signal reception intensity indication sequence and a phase value sequence when receiving a tag signal transmitted by the reader;
preprocessing is carried out according to the signal receiving intensity indication sequence and the phase value sequence respectively to obtain a corresponding signal receiving intensity indication matrix and a phase value matrix;
inputting the signal receiving intensity indication matrix and the phase value matrix into a target human body behavior recognition model, so that the target human body behavior recognition model outputs a corresponding behavior discrimination result, and the target human body behavior recognition model is trained by the training method described in the previous embodiment.
Therefore, the human body behavior recognition model obtained by the training method in the embodiment is adopted to carry out the human body behavior recognition model, so that the influence of the characteristics which cannot be transferred and the characteristics in the specific field can be eliminated, and the human body behavior recognition precision can be improved. For details not disclosed in the embodiment of the RFID-based human behavior recognition method of the present application, please refer to the embodiment of the training method of the RFID-based human behavior recognition model described in the present application.
The following describes an embodiment of the apparatus of the present application, which may be used to perform the training method of the RFID-based human behavior recognition model or the RFID-based human behavior recognition method in the above-described embodiments of the present application. For details not disclosed in the embodiments of the apparatus of the present application, please refer to the embodiments of the training method of the RFID-based human behavior recognition model or the RFID-based human behavior recognition method described in the present application.
FIG. 2 illustrates a block diagram of a training device for RFID-based human behavior recognition models, according to one embodiment of the present application.
Referring to fig. 2, a training apparatus for an RFID-based human behavior recognition model according to an embodiment of the present application includes:
The acquisition module is used for acquiring a target domain training data set, wherein the target domain training data set comprises a signal receiving intensity indication sequence and a phase value sequence of a received tag signal when different types of human behaviors are performed;
the first preprocessing module is used for preprocessing the signal receiving intensity indication sequence and the phase value sequence respectively to obtain a corresponding signal receiving intensity indication matrix and a phase value matrix;
the first input module is used for inputting the signal receiving intensity indication matrix and the phase value matrix into a human body behavior recognition model pre-trained by a source domain training data set so that the human body behavior recognition model outputs a corresponding weighted target feature vector, and the human body behavior recognition model comprises a feature extraction module and a self-attention module which are connected;
the second input module is used for inputting the weighted target feature vector to a domain discriminator so as to enable the domain discriminator to output a corresponding domain discrimination tag, wherein the domain discrimination tag is used for determining whether the data come from a source domain or a target domain;
and the processing module is used for carrying out batch spectrum punishment calculation according to target feature vectors respectively from a target domain and a source domain, and carrying out optimization on the human body behavior recognition model according to a calculation result and the domain discrimination label so as to obtain a target human body behavior recognition model.
In one embodiment of the present application, the feature extraction module performs feature extraction on the signal reception intensity indication matrix and the phase value matrix, including: respectively inputting the signal receiving intensity indication matrix and the phase value matrix into a plurality of convolution layers to obtain corresponding first feature vectors; and splicing the first eigenvector corresponding to the signal receiving intensity indication matrix and the first eigenvector corresponding to the phase value matrix to obtain a second eigenvector, and inputting the second eigenvector to a double-layer GRU so that the double-layer GRU outputs a corresponding third eigenvector as the input of the self-attention module.
In one embodiment of the present application, the preprocessing includes at least one of resampling, phase unwrapping, filtering and smoothing, and normalization.
In one embodiment of the present application, the domain arbiter includes a user domain arbiter for determining whether the source domain training data set is the same acquisition object and an environment domain arbiter for determining whether the source domain training data set is the same acquisition environment.
In one embodiment of the present application, there is further provided an RFID-based human behavior recognition device, which is applied to a terminal device, where the terminal device is connected to a reader, and the reader is configured to receive tag signals of a plurality of target tags, where the plurality of target tags are respectively disposed at predetermined positions of an acquisition object;
the device comprises:
a receiving module for receiving a signal reception intensity indication sequence and a phase value sequence transmitted by the reader when receiving a tag signal;
the second preprocessing module is used for preprocessing the signal receiving intensity indication sequence and the phase value sequence respectively to obtain a corresponding signal receiving intensity indication matrix and a phase value matrix;
the recognition module is used for inputting the signal receiving intensity indication matrix and the phase value matrix into a target human body behavior recognition model so that the target human body behavior recognition model outputs a corresponding behavior discrimination result, and the target human body behavior recognition model is trained by the training method according to any one of claims 1-4.
Fig. 3 shows a schematic diagram of a computer system suitable for use in implementing the electronic device of the embodiments of the present application.
It should be noted that, the computer system of the electronic device shown in fig. 3 is only an example, and should not impose any limitation on the functions and the application scope of the embodiments of the present application.
As shown in fig. 3, the computer system includes a central processing unit (Central Processing Unit, CPU) 301 that can perform various appropriate actions and processes, such as performing the methods described in the above embodiments, according to a program stored in a Read-Only Memory (ROM) 302 or a program loaded from a storage section 308 into a random access Memory (Random Access Memory, RAM) 303. In the RAM 303, various programs and data required for the system operation are also stored. The CPU 301, ROM 302, and RAM 303 are connected to each other through a bus 304. An Input/Output (I/O) interface 305 is also connected to bus 304.
The following components are connected to the I/O interface 305: an input section 306 including a keyboard, a mouse, and the like; an output portion 307 including a Cathode Ray Tube (CRT), a liquid crystal display (Liquid Crystal Display, LCD), and the like, a speaker, and the like; a storage section 308 including a hard disk or the like; and a communication section 309 including a network interface card such as a LAN (Local Area Network ) card, a modem, or the like. The communication section 309 performs communication processing via a network such as the internet. The drive 310 is also connected to the I/O interface 305 as needed. A removable medium 311 such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, or the like is installed on the drive 310 as needed, so that a computer program read therefrom is installed into the storage section 308 as needed.
In particular, according to embodiments of the present application, the processes described above with reference to flowcharts may be implemented as computer software programs. For example, embodiments of the present application include a computer program product comprising a computer program embodied on a computer readable medium, the computer program comprising a computer program for performing the method shown in the flowchart. In such an embodiment, the computer program may be downloaded and installed from a network via the communication portion 309, and/or installed from the removable medium 311. When executed by a Central Processing Unit (CPU) 301, performs the various functions defined in the system of the present application.
It should be noted that, the computer readable medium shown in the embodiments of the present application may be a computer readable signal medium or a computer readable storage medium, or any combination of the two. The computer readable storage medium can be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or a combination of any of the foregoing. More specific examples of the computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-Only Memory (ROM), an erasable programmable read-Only Memory (Erasable Programmable Read Only Memory, EPROM), flash Memory, an optical fiber, a portable compact disc read-Only Memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the context of this document, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. In the present application, however, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, with a computer-readable computer program embodied therein. Such a propagated data signal may take any of a variety of forms, including, but not limited to, electro-magnetic, optical, or any suitable combination of the foregoing. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. A computer program embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wired, etc., or any suitable combination of the foregoing.
The flowcharts and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present application. Where each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams or flowchart illustration, and combinations of blocks in the block diagrams or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.
The units involved in the embodiments of the present application may be implemented by means of software, or may be implemented by means of hardware, and the described units may also be provided in a processor. Wherein the names of the units do not constitute a limitation of the units themselves in some cases.
As another aspect, the present application also provides a computer-readable medium that may be contained in the electronic device described in the above embodiment; or may exist alone without being incorporated into the electronic device. The computer-readable medium carries one or more programs which, when executed by the electronic device, cause the electronic device to implement the methods described in the above embodiments.
It should be noted that although in the above detailed description several modules or units of a device for action execution are mentioned, such a division is not mandatory. Indeed, the features and functions of two or more modules or units described above may be embodied in one module or unit, in accordance with embodiments of the present application. Conversely, the features and functions of one module or unit described above may be further divided into a plurality of modules or units to be embodied.
From the above description of embodiments, those skilled in the art will readily appreciate that the example embodiments described herein may be implemented in software, or may be implemented in software in combination with the necessary hardware. Thus, the technical solution according to the embodiments of the present application may be embodied in the form of a software product, which may be stored in a non-volatile storage medium (may be a CD-ROM, a usb disk, a mobile hard disk, etc.) or on a network, and includes several instructions to cause a computing device (may be a personal computer, a server, a touch terminal, or a network device, etc.) to perform the method according to the embodiments of the present application.
Other embodiments of the present application will be apparent to those skilled in the art from consideration of the specification and practice of the embodiments disclosed herein. This application is intended to cover any variations, uses, or adaptations of the application following, in general, the principles of the application and including such departures from the present disclosure as come within known or customary practice within the art to which the application pertains.
It is to be understood that the present application is not limited to the precise arrangements and instrumentalities shown in the drawings, which have been described above, and that various modifications and changes may be effected without departing from the scope thereof. The scope of the application is limited only by the appended claims.
Claims (10)
1. The training method of the human behavior recognition model based on the RFID is characterized by comprising the following steps of:
acquiring a target domain training data set, wherein the target domain training data set comprises a signal receiving intensity indication sequence and a phase value sequence of a received tag signal when different types of human behaviors are performed;
preprocessing the signal receiving intensity indication sequence and the phase value sequence respectively to obtain a corresponding signal receiving intensity indication matrix and a phase value matrix;
Inputting the signal receiving intensity indication matrix and the phase value matrix into a human body behavior recognition model pre-trained by a source domain training data set so that the human body behavior recognition model outputs a corresponding weighted target feature vector, wherein the human body behavior recognition model comprises a feature extraction module and a self-attention module which are connected;
inputting the weighted target feature vector to a domain discriminator so that the domain discriminator outputs a corresponding domain discriminating label, wherein the domain discriminating label is used for determining whether the data come from a source domain or a target domain;
and performing batch spectrum punishment calculation according to target feature vectors respectively from a target domain and a source domain, and optimizing the human body behavior recognition model according to a calculation result and the domain discrimination label to obtain a target human body behavior recognition model.
2. The method of claim 1, wherein the feature extraction module performs feature extraction on the signal reception intensity indication matrix and the phase value matrix, comprising:
respectively inputting the signal receiving intensity indication matrix and the phase value matrix into a plurality of convolution layers to obtain corresponding first feature vectors;
And splicing the first eigenvector corresponding to the signal receiving intensity indication matrix and the first eigenvector corresponding to the phase value matrix to obtain a second eigenvector, and inputting the second eigenvector to a double-layer GRU so that the double-layer GRU outputs a corresponding third eigenvector as the input of the self-attention module.
3. The method of claim 1, wherein the preprocessing comprises at least one of resampling, phase unwrapping, filtering and smoothing, and normalization.
4. A method according to any of claims 1-3, wherein the domain arbiter comprises a user domain arbiter for determining whether the source domain training data set is the same acquisition object and an environment domain arbiter for determining whether the source domain training data set is the same acquisition environment.
5. The human behavior recognition method based on RFID is characterized by being applied to terminal equipment, wherein the terminal equipment is connected with a reader, the reader is used for receiving tag signals of a plurality of target tags, and the target tags are respectively arranged at preset positions of an acquisition object;
The method comprises the following steps:
a signal reception intensity indication sequence and a phase value sequence when receiving a tag signal transmitted by the reader;
preprocessing is carried out according to the signal receiving intensity indication sequence and the phase value sequence respectively to obtain a corresponding signal receiving intensity indication matrix and a phase value matrix;
inputting the signal receiving intensity indication matrix and the phase value matrix into a target human body behavior recognition model, so that the target human body behavior recognition model outputs a corresponding behavior discrimination result, wherein the target human body behavior recognition model is trained by the training method according to any one of claims 1-4.
6. A training device for a human behavior recognition model based on RFID, comprising:
the acquisition module is used for acquiring a target domain training data set, wherein the target domain training data set comprises a signal receiving intensity indication sequence and a phase value sequence of a received tag signal when different types of human behaviors are performed;
the first preprocessing module is used for preprocessing the signal receiving intensity indication sequence and the phase value sequence respectively to obtain a corresponding signal receiving intensity indication matrix and a phase value matrix;
The first input module is used for inputting the signal receiving intensity indication matrix and the phase value matrix into a human body behavior recognition model pre-trained by a source domain training data set so that the human body behavior recognition model outputs a corresponding weighted target feature vector, and the human body behavior recognition model comprises a feature extraction module and a self-attention module which are connected;
the second input module is used for inputting the weighted target feature vector to a domain discriminator so as to enable the domain discriminator to output a corresponding domain discrimination tag, wherein the domain discrimination tag is used for determining whether the data come from a source domain or a target domain;
and the processing module is used for carrying out batch spectrum punishment calculation according to target feature vectors respectively from a target domain and a source domain, and carrying out optimization on the human body behavior recognition model according to a calculation result and the domain discrimination label so as to obtain a target human body behavior recognition model.
7. The apparatus of claim 6, wherein the feature extraction module performs feature extraction on the signal reception intensity indication matrix and the phase value matrix, comprising:
respectively inputting the signal receiving intensity indication matrix and the phase value matrix into a plurality of convolution layers to obtain corresponding first feature vectors;
And splicing the first eigenvector corresponding to the signal receiving intensity indication matrix and the first eigenvector corresponding to the phase value matrix to obtain a second eigenvector, and inputting the second eigenvector to a double-layer GRU so that the double-layer GRU outputs a corresponding third eigenvector as the input of the self-attention module.
8. The human behavior recognition device based on RFID is characterized by being applied to terminal equipment, wherein the terminal equipment is connected with a reader, the reader is used for receiving tag signals of a plurality of target tags, and the target tags are respectively arranged at preset positions of an acquisition object;
the device comprises:
a receiving module for receiving a signal reception intensity indication sequence and a phase value sequence transmitted by the reader when receiving a tag signal;
the second preprocessing module is used for preprocessing the signal receiving intensity indication sequence and the phase value sequence respectively to obtain a corresponding signal receiving intensity indication matrix and a phase value matrix;
the recognition module is used for inputting the signal receiving intensity indication matrix and the phase value matrix into a target human body behavior recognition model so that the target human body behavior recognition model outputs a corresponding behavior discrimination result, and the target human body behavior recognition model is trained by the training method according to any one of claims 1-4.
9. A computer readable medium, on which a computer program is stored, which computer program, when being executed by a processor, implements the method according to any one of claims 1 to 5.
10. An electronic device, comprising:
one or more processors;
storage means for storing one or more programs which when executed by the one or more processors cause the one or more processors to implement the method of any of claims 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310473044.5A CN116486341A (en) | 2023-04-27 | 2023-04-27 | Training and identifying method and device for human body behavior identification model based on RFID |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310473044.5A CN116486341A (en) | 2023-04-27 | 2023-04-27 | Training and identifying method and device for human body behavior identification model based on RFID |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116486341A true CN116486341A (en) | 2023-07-25 |
Family
ID=87222935
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310473044.5A Pending CN116486341A (en) | 2023-04-27 | 2023-04-27 | Training and identifying method and device for human body behavior identification model based on RFID |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116486341A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117435902A (en) * | 2023-12-20 | 2024-01-23 | 武汉华威科智能技术有限公司 | Method and device for determining RFID tag movement behavior based on machine learning |
-
2023
- 2023-04-27 CN CN202310473044.5A patent/CN116486341A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117435902A (en) * | 2023-12-20 | 2024-01-23 | 武汉华威科智能技术有限公司 | Method and device for determining RFID tag movement behavior based on machine learning |
| CN117435902B (en) * | 2023-12-20 | 2024-04-02 | 武汉华威科智能技术有限公司 | Method and device for determining RFID tag movement behavior based on machine learning |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111860573B (en) | Model training method, image category detection method and device and electronic equipment | |
| CN111160462B (en) | Unsupervised human activity recognition method based on multi-sensor data alignment | |
| CN111950596A (en) | Training method for neural network and related equipment | |
| CN108197592B (en) | Information acquisition method and device | |
| CN111401433A (en) | User information acquisition method and device, electronic equipment and storage medium | |
| CN109636047A (en) | User activity prediction model training method, system, equipment and storage medium | |
| CN113656558B (en) | Method and device for evaluating association rule based on machine learning | |
| CN108875836B (en) | Simple-complex activity collaborative recognition method based on deep multitask learning | |
| CN113569554B (en) | Entity pair matching method and device in database, electronic equipment and storage medium | |
| CN116486341A (en) | Training and identifying method and device for human body behavior identification model based on RFID | |
| CN113705534A (en) | Behavior prediction method, behavior prediction device, behavior prediction equipment and storage medium based on deep vision | |
| CN113724010A (en) | Customer loss prediction method and device | |
| CN110363121A (en) | Fingerprint image processing method and processing device, storage medium and electronic equipment | |
| CN115034315A (en) | Business processing method and device based on artificial intelligence, computer equipment and medium | |
| CN116385850A (en) | Multi-target detection method, device, electronic equipment and storage medium | |
| CN111950647A (en) | Classification model training method and device | |
| CN113065634B (en) | Image processing method, neural network training method and related equipment | |
| CN116578925B (en) | Behavior prediction method, device and storage medium based on feature images | |
| CN116739787B (en) | Transaction recommendation method and system based on artificial intelligence | |
| CN111325578B (en) | Sample determination method and device of prediction model, medium and equipment | |
| CN116705196A (en) | Drug target interaction prediction method and device based on symbolic graph neural network | |
| CN116738144A (en) | Human behavior detection method, device, medium and equipment based on RFID | |
| CN116152551A (en) | Classification model training method, classification method, device, equipment and medium | |
| CN115223157A (en) | Power grid equipment nameplate optical character recognition method based on recurrent neural network | |
| CN114886383A (en) | Electroencephalogram signal emotional feature classification method based on transfer 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 |