CN112036504A - Temperature measurement model training method, device, equipment and storage medium - Google Patents
Temperature measurement model training method, device, equipment and storage medium Download PDFInfo
- Publication number
- CN112036504A CN112036504A CN202010970862.2A CN202010970862A CN112036504A CN 112036504 A CN112036504 A CN 112036504A CN 202010970862 A CN202010970862 A CN 202010970862A CN 112036504 A CN112036504 A CN 112036504A
- Authority
- CN
- China
- Prior art keywords
- model
- temperature measurement
- initial
- training
- parameters
- 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
- 238000009529 body temperature measurement Methods 0.000 title claims abstract description 159
- 238000012549 training Methods 0.000 title claims abstract description 148
- 238000000034 method Methods 0.000 title claims abstract description 65
- 239000013598 vector Substances 0.000 claims abstract description 49
- 230000007613 environmental effect Effects 0.000 claims abstract description 18
- 238000004590 computer program Methods 0.000 claims description 16
- 238000012545 processing Methods 0.000 claims description 16
- 238000000605 extraction Methods 0.000 claims description 9
- 238000004861 thermometry Methods 0.000 claims description 6
- 239000000284 extract Substances 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims description 3
- 238000007781 pre-processing Methods 0.000 claims description 3
- 238000005259 measurement Methods 0.000 abstract description 9
- 238000010276 construction Methods 0.000 abstract 1
- 238000010586 diagram Methods 0.000 description 10
- 238000004364 calculation method Methods 0.000 description 5
- 238000003491 array Methods 0.000 description 2
- 238000004422 calculation algorithm Methods 0.000 description 2
- 238000013136 deep learning model Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
- G06N20/20—Ensemble learning
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Medical Informatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Radiation Pyrometers (AREA)
Abstract
The application relates to the field of model construction, and particularly discloses a training method, a device, equipment and a storage medium for a temperature measurement model, wherein the method comprises the following steps: the method comprises the steps that a plurality of clients obtain model initial parameters from a server side, model training is carried out according to a pre-constructed local training set to obtain a plurality of initial models, and the local training set is a feature vector comprising environmental information; the server receives initial model parameters corresponding to the initial models sent by the clients and processes the initial model parameters to obtain average parameters; the server side sends the average parameters to each client side respectively and judges whether the initial models are converged or not respectively; and if the initial models are all converged, the client adjusts the initial models according to the average parameters to obtain the temperature measurement model. And performing joint training according to the temperature measurement data in multiple scenes to improve the data security and improve the measurement accuracy of the temperature measurement model.
Description
Technical Field
The present application relates to the field of model training, and in particular, to a method, an apparatus, a device, and a storage medium for training a temperature measurement model.
Background
With the development of artificial intelligence, deep learning, machine learning, etc. have been gradually applied to human infrared temperature measurement products. Because the influence factors on the infrared temperature measurement precision of the human body are more, the temperature measurement data which can be collected by the infrared temperature measurement product only depending on a single point has a single scene and few temperature measurement data samples, the temperature measurement precision of the temperature measurement model obtained by training is lower. However, if temperature measurement data under all scenes needs to be collected, not only the workload is large, but also some privacy problems are involved, and the data security is low.
Therefore, how to perform joint training according to temperature measurement data in multiple scenes to improve data security and improve measurement accuracy of a temperature measurement model becomes an urgent problem to be solved.
Disclosure of Invention
The application provides a training method, a training device, equipment and a storage medium of a temperature measurement model, which are used for performing combined training according to temperature measurement data under multiple scenes, so that the data security is improved, and the measurement accuracy of the temperature measurement model is improved.
In a first aspect, the present application provides a method for training a temperature measurement model, the method comprising:
the method comprises the steps that a plurality of clients obtain model initial parameters from a server side, model training is carried out according to a pre-constructed local training set to obtain a plurality of initial models, and the local training set is a feature vector comprising environmental information; the server receives initial model parameters corresponding to the initial models sent by the clients and processes the initial model parameters to obtain average parameters; the server side sends the average parameters to each client side respectively and judges whether the initial models are converged or not respectively; and if the initial models are all converged, the client adjusts the initial models according to the average parameters to obtain the temperature measurement model.
In a second aspect, the present application further provides a training device for a temperature measurement model, the device comprising:
the initial training module is used for obtaining model initial parameters from the server side by a plurality of client sides and carrying out model training according to a pre-constructed local training set to obtain a plurality of initial models, wherein the local training set is a feature vector comprising environmental information; the parameter processing module is used for receiving initial model parameters corresponding to the initial models sent by the plurality of clients by the server side and processing the plurality of initial model parameters to obtain average parameters; the convergence judging module is used for respectively sending the average parameters to each client by the server and respectively judging whether the plurality of initial models are converged; and the model adjusting module is used for adjusting the initial model according to the average parameter by the client side if the initial model is converged so as to obtain the temperature measurement model.
In a third aspect, the present application further provides a computer device comprising a memory and a processor; the memory is used for storing a computer program; the processor is configured to execute the computer program and to implement the training method of the temperature measurement model as described above when the computer program is executed.
In a fourth aspect, the present application further provides a computer-readable storage medium, in which a computer program is stored, which, when executed by a processor, causes the processor to implement the method for training a temperature measurement model as described above.
The application discloses a training method, a device, equipment and a storage medium of a temperature measurement model, wherein a plurality of client terminals are used for obtaining model initial parameters from a server terminal, and model training is carried out according to a pre-constructed local training set to obtain a plurality of initial models, wherein the local training set is a feature vector comprising environmental information; the server receives initial model parameters corresponding to the initial models sent by the clients and processes the initial model parameters to obtain average parameters; the server side sends the average parameters to each client side respectively and judges whether the initial models are converged or not respectively; and if the initial models are all converged, the client adjusts the initial models according to the average parameters to obtain the temperature measurement model. The training process of the model occurs locally at each client, so that the local training set of each client is not shared with other clients, the data safety and reliability are improved, and meanwhile, the combined training is performed according to the temperature measurement data of a plurality of clients, so that the accuracy of the obtained temperature measurement model on temperature measurement is improved. And the training set used for training the initial model comprises environmental information, so that the prediction precision of the initial model obtained by training is improved.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the following description are some embodiments of the present application, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.
FIG. 1 is a schematic flow chart diagram of a method for training a temperature measurement model according to an embodiment of the present disclosure;
FIG. 2 is a schematic flow chart diagram of another method for training a temperature measurement model provided by an embodiment of the present application;
FIG. 3 is a schematic block diagram of a training apparatus for a temperature measurement model provided by an embodiment of the present application;
FIG. 4 is a schematic block diagram of another temperature measurement model training apparatus provided by an embodiment of the present application;
fig. 5 is a schematic block diagram of a computer device according to an embodiment of the present application.
Detailed Description
The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are some, but not all, embodiments of the present application. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
The flow diagrams depicted in the figures are merely illustrative and do not necessarily include all of the elements and operations/steps, nor do they necessarily have to be performed in the order depicted. For example, some operations/steps may be decomposed, combined or partially combined, so that the actual execution sequence may be changed according to the actual situation.
It is to be understood that the terminology used in the description of the present application herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. As used in the specification of the present application and the appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.
It should also be understood that the term "and/or" as used in this specification and the appended claims refers to and includes any and all possible combinations of one or more of the associated listed items.
The embodiment of the application provides a training method and device of a temperature measurement model, computer equipment and a storage medium. The training method of the temperature measurement model can be used for training the temperature measurement model, so that the measurement accuracy of the temperature measurement model obtained through training is improved, and the convenience and accuracy in temperature measurement are further improved.
Some embodiments of the present application will be described in detail below with reference to the accompanying drawings. The embodiments described below and the features of the embodiments can be combined with each other without conflict.
Referring to fig. 1, fig. 1 is a schematic flow chart of a training method of a temperature measurement model according to an embodiment of the present application. According to the training method of the temperature measurement model, the combined training of the temperature measurement model is carried out on the temperature measurement data of the plurality of clients, so that the measurement accuracy of the temperature measurement model obtained through training is improved.
As shown in fig. 1, the training method of the temperature measurement model specifically includes: step S101 to step S104.
S101, a plurality of clients acquire model initial parameters from a server side, and perform model training according to a pre-constructed local training set to obtain a plurality of initial models.
The client side is a local temperature measurement client side, the plurality of local temperature measurement client sides respectively obtain model initialization parameters from the server side, model training is carried out according to a pre-constructed local training set, and a plurality of initial models are obtained, wherein the local training set can be a feature vector comprising environment information. In the specific implementation process, for example, a Gradient boosting regression (e.g., a ensemble learning algorithm) may be used for model training.
When training, each local temperature measurement client can perform model training according to a training set locally constructed by the local temperature measurement client, that is, training samples of local training sets used by different local temperature measurement clients can be the same or different, but structures and data formats in the local training sets used by different local temperature measurement clients are the same.
Because the initial model training is carried out on each local temperature measurement client, each local temperature measurement client does not need to carry out intercommunication sharing of a training set with other clients when carrying out model training, and therefore, the data safety and reliability are improved. And moreover, the environment information is added into the feature vector, and model training is carried out according to the feature vector comprising the environment information, so that the influence of the environment information on the initial model obtained by training is reduced, and the prediction precision of the initial model obtained by training is improved.
In some embodiments, when each client performs training of the initial model, whether the model is trained completely or not can be determined by calculating a loss function of the initial model.
S102, the server receives initial model parameters corresponding to the initial models sent by the clients, and processes the initial model parameters to obtain average parameters.
Each local temperature measurement client can train an initial model based on a local training set, after the initial model training is completed, each local temperature measurement client sends model parameters of the initial model trained by the local temperature measurement client to the server, and the server processes the initial model parameters after receiving the initial model parameters sent by each client to obtain average parameters. The processing may include various ways, such as averaging parameters, etc.
In some embodiments, said processing a plurality of said initial model parameters comprises: and carrying out weighted average processing on a plurality of initial model parameters.
For a plurality of initial model parameters sent by a plurality of clients, when the weighted average algorithm is adopted to calculate the average parameter, the weights of the plurality of clients can be adjusted, so that the accuracy of the finally obtained average parameter is improved. For example, clients with larger amounts of data in the local training set may be adjusted to a higher weight, while clients with smaller amounts of data in the local training set may be adjusted to a lower weight.
S103, the server side sends the average parameters to each client side respectively, and judges whether the initial models are converged or not respectively.
After the average parameters are obtained through calculation, the server side can send the average parameters obtained through calculation to each client side respectively, and the convergence state of the initial model trained by each local temperature measurement client side is judged.
When determining whether the initial model trained by each local temperature measurement client has converged, the state of the initial model may be determined in various ways, for example, by calculating the loss function of each initial model, or by the number of iterative training of each initial model.
In some embodiments, said determining whether a plurality of said initial models converge comprises: judging whether the iteration times of the initial model reach preset times or not; and if the iteration times of the initial model reach preset times, judging that the initial model is converged.
The server can set a preset number of times, the preset number of times represents the number of iterations required for the initial model, and the initial model is considered to be converged only when the number of iterative training times of the initial model reaches the preset number of times.
For the initial models respectively trained by a plurality of local temperature measurement clients, the same preset times can be adopted for judgment, and when judging whether the initial models are converged, each initial model needs to be judged respectively.
And S104, if the initial models are all converged, the client adjusts the initial models according to the average parameters to obtain a temperature measurement model.
When the initial model trained by each local temperature measurement client is judged to be convergent by the server, the client adjusts the initial model parameters of the initial model according to the received average parameters, and the model after parameter adjustment is used as a temperature measurement model. The client can measure the temperature according to the temperature measurement model.
In the training method for the temperature measurement model provided by the embodiment, the initial parameters of the model are obtained from the server by using the plurality of clients, and the model training is performed according to the pre-constructed local training set to obtain a plurality of initial models; the server receives initial model parameters corresponding to the initial models sent by the clients and processes the initial model parameters to obtain average parameters; the server side sends the average parameters to each client side respectively and judges whether the initial models are converged or not respectively; and if the initial models are all converged, the client adjusts the initial models according to the average parameters to obtain the temperature measurement model. The training process of the model occurs locally at each client, so that the local training set of each client is not shared with other clients, the data safety and reliability are improved, and meanwhile, the combined training is performed according to the temperature measurement data of a plurality of clients, so that the accuracy of the obtained temperature measurement model on temperature measurement is improved.
Referring to fig. 2, fig. 2 is a schematic flow chart of another temperature measurement model training method according to an embodiment of the present application.
As shown in fig. 2, the training method of the temperature measurement model specifically includes: step S201 to step S208.
S201, the client side obtains a local temperature measurement image and temperature information corresponding to the local temperature measurement image.
The local temperature measurement client can acquire a human body infrared temperature measurement data picture by using the infrared camera, take the acquired human body infrared temperature measurement data picture as a local temperature measurement image, and record temperature information obtained by measuring the human body infrared temperature measurement data picture.
S202, the client extracts the characteristics of the local temperature measurement image to obtain a characteristic vector corresponding to the local temperature measurement image.
When the local temperature measurement image is subjected to feature extraction, a deep learning model can be adopted for feature extraction, high-dimensional features of the image are extracted from the local temperature measurement image, and an abstract structure of an object to be detected, namely a human body, in the image is displayed by a group of high-dimensional arrays. For example, the deep learning model may be resnet 50.
In some embodiments, prior to said feature extracting the local thermometry image, the method further comprises: and the client side preprocesses the local temperature measurement image.
Wherein the pre-processing may include at least one of image enhancement and image normalization. The local temperature measurement image is preprocessed to eliminate irrelevant information in the image, filter interference and noise, enhance the detectability of available information and simplify data, thereby improving the reliability of feature extraction.
S203, adding the environmental information corresponding to the local temperature measurement image into the characteristic vector to obtain a target characteristic vector, and storing the target characteristic vector and the temperature information corresponding to the local temperature measurement image to obtain a local training set.
Acquiring environment information when a local temperature measurement image is acquired, adding the environment information into the characteristic vector to obtain a target characteristic vector, taking the temperature information corresponding to the local temperature measurement image as a label value, respectively storing the label value and the target characteristic vector of the same local temperature measurement image in a local database in a one-to-one correspondence manner, and constructing a local training set.
For each local thermometric client, a local training set can be constructed based on the above manner. The data of the training set is stored locally, and the mutual transmission of sample data among a plurality of local temperature measurement clients is not needed, so that the data security is improved.
In some embodiments, before the storing the feature vector and the temperature information corresponding to the local thermometry image, the method further includes: the client performs dimensionality reduction operation on the feature vector to obtain a dimensionality reduced feature vector; and the client acquires environment information when a local temperature measurement image is acquired, and adds the environment information into the reduced feature vector to obtain a target feature vector.
After the local temperature measurement image is subjected to feature extraction to obtain high-dimensional features of the image, in order to reduce the number of predicted variables and ensure the independence between the variables, the feature vector can be subjected to dimension reduction operation. In the dimension reduction, the PCA model may be used for the dimension reduction operation, but other dimension reduction methods may also be used, such as Lasso, wavelet analysis, LDA, and so on.
The environmental information includes air humidity, room temperature, measurement angle, and measurement distance. Because the precision of human body infrared temperature measurement is influenced by factors such as air humidity, room temperature, measurement angle and measurement distance, the variables can influence the temperature measured by the infrared camera. And the acquired local temperature measurement images may have different environmental information due to different positions and times of the images acquired by the local clients, so that the environmental information is added into the reduced-dimension characteristic vectors, and when the clients train the initial model by using the target characteristic vectors, the model trained by the clients has better robustness.
S204, the plurality of clients acquire model initial parameters from the server side, and perform model training according to a pre-constructed local training set to obtain a plurality of initial models.
The client side is a local temperature measurement client side, the plurality of local temperature measurement client sides respectively obtain model initialization parameters from the server side, model training is carried out according to a local training set which is constructed in advance, and a plurality of initial models are obtained.
S205, the server receives the initial model parameters corresponding to the initial models sent by the plurality of clients, and processes the plurality of initial model parameters to obtain average parameters.
Each local temperature measurement client can train an initial model based on a local training set, after the initial model training is completed, each local temperature measurement client sends model parameters of the initial model trained by the local temperature measurement client to the server, and the server processes the initial model parameters after receiving the initial model parameters sent by each client to obtain average parameters.
S206, the server side sends the average parameters to each client side respectively, and judges whether the plurality of initial models are converged or not respectively.
After the average parameters are obtained through calculation, the server side can send the average parameters obtained through calculation to each client side respectively, and the convergence state of the initial model trained by each local temperature measurement client side is judged.
And S207, if the initial models are all converged, the client adjusts the initial models according to the average parameters to obtain the temperature measurement model.
When the initial model trained by each local temperature measurement client is judged to be convergent by the server, the client adjusts the initial model parameters of the initial model according to the received average parameters, and the model after parameter adjustment is used as a temperature measurement model.
And S208, if at least one initial model is not converged, the client side continues to train the model according to the average parameters.
And if the server side judges that at least one initial model is not converged, each client side continues to train the model according to the average parameters sent by the server side.
That is, the local temperature measurement clients use the average parameters as model parameters, iterative training of the models is performed by using the local training set, after model training of each local temperature measurement client is completed, the parameters obtained by training are sent to the server again, and the server determines whether all the retrained models are converged again until the initial models are converged.
In the training method for the temperature measurement model provided by the embodiment, the client establishes the local training set including the environmental information during the acquisition of the local temperature measurement image, and performs model training to obtain a plurality of initial models; the server receives initial model parameters corresponding to the initial models sent by the clients and processes the initial model parameters to obtain average parameters; the server side sends the average parameters to each client side respectively and judges whether the initial models are converged or not respectively; and if the initial models are all converged, the client adjusts the initial models according to the average parameters to obtain the temperature measurement model. On the basis of improving the data safety and reliability, model training is carried out by utilizing the environmental information, so that the initial model obtained by training has better robustness. And the average parameter is calculated by the server side, the average parameter is used as the model parameter under the condition that all initial models are converged, the sample data size of the models in the training process is increased, combined training is carried out according to the temperature measurement data of a plurality of client sides, and the accuracy of the obtained temperature measurement model on temperature measurement is further improved.
Referring to fig. 3, fig. 3 is a schematic block diagram of a training apparatus for a temperature measurement model according to an embodiment of the present application, the training apparatus for a temperature measurement model being used for executing the aforementioned training method for a temperature measurement model. Wherein, the training device of the temperature measurement model can be configured in a server or a terminal.
The server may be an independent server or a server cluster. The terminal can be an electronic device such as a mobile phone, a tablet computer, a notebook computer, a desktop computer, a personal digital assistant and a wearable device.
As shown in fig. 3, the temperature measurement model device 300 includes: an initial training module 301, a parameter processing module 302, a convergence determination module 303, and a model adjustment module 304.
The initial training module 301 is configured to obtain model initial parameters from a server by multiple clients, and perform model training according to a pre-constructed local training set to obtain multiple initial models, where the local training set is a feature vector including environment information.
A parameter processing module 302, configured to receive, by the server, initial model parameters corresponding to the initial models sent by the multiple clients, and process the multiple initial model parameters to obtain average parameters.
And a convergence judging module 303, configured to send the average parameter to each client by the server, and respectively judge whether the plurality of initial models converge.
A model adjusting module 304, configured to, if the initial models are all converged, adjust the initial models by the client according to the average parameters to obtain a temperature measurement model.
Referring to fig. 4, fig. 4 is a schematic block diagram of a training device for providing another temperature measurement model according to an embodiment of the present application.
As shown in fig. 4, the temperature measurement model device 400 includes: an image acquisition module 401, a feature extraction module 402, an information storage module 403, an initial training module 404, a parameter processing module 405, a convergence judgment module 406, a model adjustment module 407, and a continuous training module 408.
The image obtaining module 401 is configured to obtain, by a client, a local temperature measurement image and temperature information corresponding to the local temperature measurement image.
And a feature extraction module 402, configured to perform feature extraction on the local temperature measurement image by the client to obtain a feature vector corresponding to the local temperature measurement image.
And an information storage module 403, where the client stores the feature vector and the temperature information corresponding to the local temperature measurement image to obtain a local training set.
The initial training module 404 is configured to obtain model initial parameters from a server by a plurality of clients, and perform model training according to a pre-constructed local training set to obtain a plurality of initial models, where the local training set is a feature vector including environment information.
A parameter processing module 405, configured to receive, by the server, initial model parameters corresponding to the initial models sent by the multiple clients, and process the multiple initial model parameters to obtain average parameters.
And a convergence judging module 406, configured to send the average parameters to each client by the server, and respectively judge whether the multiple initial models converge.
And a model adjusting module 407, configured to, if the initial models are all converged, adjust, by the client, the initial model according to the average parameter to obtain a temperature measurement model.
A continuing training module 408, configured to, if at least one of the initial models does not converge, continue model training by the client according to the average parameter.
It should be noted that, as will be clearly understood by those skilled in the art, for convenience and brevity of description, the specific working processes of the training apparatus for a temperature measurement model and each module described above may refer to the corresponding processes in the foregoing embodiments of the training method for a temperature measurement model, and are not described herein again.
The training means of the temperature measurement model described above may be implemented in the form of a computer program which can be run on a computer device as shown in fig. 5.
Referring to fig. 5, fig. 5 is a schematic block diagram of a computer device according to an embodiment of the present disclosure. The computer device may be a server or a terminal.
Referring to fig. 5, the computer device includes a processor, a memory, and a network interface connected through a system bus, wherein the memory may include a nonvolatile storage medium and an internal memory.
The non-volatile storage medium may store an operating system and a computer program. The computer program includes program instructions that, when executed, cause a processor to perform any one of the methods of training a temperature measurement model.
The processor is used for providing calculation and control capability and supporting the operation of the whole computer equipment.
The internal memory provides an environment for the execution of a computer program on a non-volatile storage medium, which computer program, when executed by the processor, causes the processor to perform any one of the methods for training a temperature measurement model.
The network interface is used for network communication, such as sending assigned tasks and the like. Those skilled in the art will appreciate that the architecture shown in fig. 5 is merely a block diagram of some of the structures associated with the disclosed aspects and is not intended to limit the computing devices to which the disclosed aspects apply, as particular computing devices may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.
It should be understood that the Processor may be a Central Processing Unit (CPU), and the Processor may be other general purpose processors, Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field Programmable Gate Arrays (FPGAs) or other Programmable logic devices, discrete Gate or transistor logic devices, discrete hardware components, etc. Wherein a general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
Wherein, in one embodiment, the processor is configured to execute a computer program stored in the memory to implement the steps of:
the method comprises the steps that a plurality of clients obtain model initial parameters from a server side, model training is carried out according to a pre-constructed local training set to obtain a plurality of initial models, and the local training set is a feature vector comprising environmental information; the server receives initial model parameters corresponding to the initial models sent by the clients and processes the initial model parameters to obtain average parameters; the server side sends the average parameters to each client side respectively and judges whether the initial models are converged or not respectively; and if the initial models are all converged, the client adjusts the initial models according to the average parameters to obtain the temperature measurement model.
In one embodiment, the processor is configured to implement: and if at least one initial model is not converged, the client side continues to train the model according to the average parameters.
In one embodiment, the processor is configured to implement: the client acquires a local temperature measurement image and temperature information corresponding to the local temperature measurement image; the client extracts the characteristics of the local temperature measurement image to obtain a characteristic vector corresponding to the local temperature measurement image; and adding the environmental information corresponding to the local temperature measurement image into the characteristic vector to obtain a target characteristic vector, and storing the target characteristic vector and the temperature information corresponding to the local temperature measurement image to obtain a local training set.
In one embodiment, before performing the feature extraction on the local thermometry image, the processor is configured to perform: the client preprocesses the local thermometry image, wherein the preprocessing may include at least one of image enhancement and image normalization.
In one embodiment, when the processor adds the environmental information corresponding to the local temperature measurement image to the feature vector to obtain a target feature vector, the processor is configured to: the client performs dimensionality reduction operation on the feature vector to obtain a dimensionality reduced feature vector; and the client acquires environment information when a local temperature measurement image is acquired, and adds the environment information into the reduced feature vector to obtain a target feature vector.
In one embodiment, the processor, when performing the processing of the plurality of initial model parameters, is configured to perform: and carrying out weighted average processing on a plurality of initial model parameters.
In one embodiment, the processor, in performing the determining whether the plurality of initial models converge, is configured to perform: judging whether the iteration times of the initial model reach preset times or not; and if the iteration times of the initial model reach preset times, judging that the initial model is converged.
The embodiment of the application also provides a computer-readable storage medium, wherein a computer program is stored in the computer-readable storage medium, the computer program comprises program instructions, and the processor executes the program instructions to implement any one of the training methods for the temperature measurement model provided by the embodiment of the application.
The computer-readable storage medium may be an internal storage unit of the computer device described in the foregoing embodiment, for example, a hard disk or a memory of the computer device. The computer readable storage medium may also be an external storage device of the computer device, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), and the like provided on the computer device.
While the invention has been described with reference to specific embodiments, the scope of the invention is not limited thereto, and those skilled in the art can easily conceive various equivalent modifications or substitutions within the technical scope of the invention. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.
Claims (10)
1. A method for training a temperature measurement model, comprising:
the method comprises the steps that a plurality of clients obtain model initial parameters from a server side, model training is carried out according to a pre-constructed local training set to obtain a plurality of initial models, and the local training set is a feature vector comprising environmental information;
the server receives initial model parameters corresponding to the initial models sent by the clients and processes the initial model parameters to obtain average parameters;
the server side sends the average parameters to each client side respectively and judges whether the initial models are converged or not respectively;
and if the initial models are all converged, the client adjusts the initial models according to the average parameters to obtain the temperature measurement model.
2. A method for training a temperature measurement model according to claim 1, the method comprising:
and if at least one initial model is not converged, the client side continues to train the model according to the average parameters.
3. A method for training a temperature measurement model according to claim 1, the method comprising:
the client acquires a local temperature measurement image and temperature information corresponding to the local temperature measurement image;
the client extracts the characteristics of the local temperature measurement image to obtain a characteristic vector corresponding to the local temperature measurement image;
and adding the environmental information corresponding to the local temperature measurement image into the characteristic vector to obtain a target characteristic vector, and storing the target characteristic vector and the temperature information corresponding to the local temperature measurement image to obtain a local training set.
4. The method for training the temperature measurement model according to claim 3, wherein before the feature extraction of the local thermometry image, the method comprises:
the client preprocesses the local thermometry image, wherein the preprocessing may include at least one of image enhancement and image normalization.
5. The method for training the temperature measurement model according to claim 3, wherein the adding environmental information corresponding to the local temperature measurement image into the feature vector to obtain a target feature vector comprises:
the client performs dimensionality reduction operation on the feature vector to obtain a dimensionality reduced feature vector;
and the client acquires environment information when a local temperature measurement image is acquired, and adds the environment information into the reduced feature vector to obtain a target feature vector.
6. The method of claim 1, wherein the processing the plurality of initial model parameters comprises:
and carrying out weighted average processing on a plurality of initial model parameters.
7. The method for training a temperature measurement model according to claim 1, wherein the determining whether the plurality of initial models converge comprises:
judging whether the iteration times of the initial model reach preset times or not;
and if the iteration times of the initial model reach preset times, judging that the initial model is converged.
8. A training device for a temperature measurement model, comprising:
the initial training module is used for obtaining model initial parameters from the server side by a plurality of client sides and carrying out model training according to a pre-constructed local training set to obtain a plurality of initial models, wherein the local training set is a feature vector comprising environmental information;
the parameter processing module is used for receiving initial model parameters corresponding to the initial models sent by the plurality of clients by the server side and processing the plurality of initial model parameters to obtain average parameters;
the convergence judging module is used for respectively sending the average parameters to each client by the server and respectively judging whether the plurality of initial models are converged;
and the model adjusting module is used for adjusting the initial model according to the average parameter by the client side if the initial model is converged so as to obtain the temperature measurement model.
9. A computer device, wherein the computer device comprises a memory and a processor;
the memory is used for storing a computer program;
the processor for executing the computer program and for implementing the method of training a temperature measurement model according to any of claims 1 to 7 when executing the computer program.
10. A computer-readable storage medium, characterized in that the computer-readable storage medium stores a computer program which, when executed by a processor, causes the processor to carry out a method of training a temperature measurement model according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010970862.2A CN112036504A (en) | 2020-09-15 | 2020-09-15 | Temperature measurement model training method, device, equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010970862.2A CN112036504A (en) | 2020-09-15 | 2020-09-15 | Temperature measurement model training method, device, equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112036504A true CN112036504A (en) | 2020-12-04 |
Family
ID=73589484
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010970862.2A Pending CN112036504A (en) | 2020-09-15 | 2020-09-15 | Temperature measurement model training method, device, equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112036504A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115230067A (en) * | 2021-09-16 | 2022-10-25 | 健大电业制品（昆山）有限公司 | Intelligent determination method and system for injection molding process parameters |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110070579A (en) * | 2019-03-16 | 2019-07-30 | 平安科技（深圳）有限公司 | Localization method, device, equipment and storage medium based on image detection |
| KR20190096311A (en) * | 2019-06-04 | 2019-08-19 | 엘지전자 주식회사 | A device for generating a temperature prediction model and a method for providing a simulation environment |
| CN110795477A (en) * | 2019-09-20 | 2020-02-14 | 平安科技（深圳）有限公司 | Data training method, device and system |
| CN111340242A (en) * | 2020-05-15 | 2020-06-26 | 支付宝(杭州)信息技术有限公司 | Model joint training method and device for protecting privacy |
| CN111458030A (en) * | 2020-03-11 | 2020-07-28 | 华瑞新智科技（北京）有限公司 | Infrared human body temperature measurement calibration method and device |
| CN111582240A (en) * | 2020-05-29 | 2020-08-25 | 上海依图网络科技有限公司 | Object quantity identification method, device, equipment and medium |
| CN111625321A (en) * | 2020-07-30 | 2020-09-04 | 上海有孚智数云创数字科技有限公司 | Virtual machine migration planning and scheduling method based on temperature prediction, system and medium thereof |
| CN111652863A (en) * | 2020-05-27 | 2020-09-11 | 刘君茹 | Medical image detection method, device, equipment and storage medium |
-
2020
- 2020-09-15 CN CN202010970862.2A patent/CN112036504A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110070579A (en) * | 2019-03-16 | 2019-07-30 | 平安科技（深圳）有限公司 | Localization method, device, equipment and storage medium based on image detection |
| KR20190096311A (en) * | 2019-06-04 | 2019-08-19 | 엘지전자 주식회사 | A device for generating a temperature prediction model and a method for providing a simulation environment |
| CN110795477A (en) * | 2019-09-20 | 2020-02-14 | 平安科技（深圳）有限公司 | Data training method, device and system |
| CN111458030A (en) * | 2020-03-11 | 2020-07-28 | 华瑞新智科技（北京）有限公司 | Infrared human body temperature measurement calibration method and device |
| CN111340242A (en) * | 2020-05-15 | 2020-06-26 | 支付宝(杭州)信息技术有限公司 | Model joint training method and device for protecting privacy |
| CN111652863A (en) * | 2020-05-27 | 2020-09-11 | 刘君茹 | Medical image detection method, device, equipment and storage medium |
| CN111582240A (en) * | 2020-05-29 | 2020-08-25 | 上海依图网络科技有限公司 | Object quantity identification method, device, equipment and medium |
| CN111625321A (en) * | 2020-07-30 | 2020-09-04 | 上海有孚智数云创数字科技有限公司 | Virtual machine migration planning and scheduling method based on temperature prediction, system and medium thereof |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115230067A (en) * | 2021-09-16 | 2022-10-25 | 健大电业制品（昆山）有限公司 | Intelligent determination method and system for injection molding process parameters |
| CN115230067B (en) * | 2021-09-16 | 2023-11-14 | 健大电业制品（昆山）有限公司 | Intelligent determining method and system for injection molding process parameters |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108537152B (en) | Method and apparatus for detecting living body | |
| US10726244B2 (en) | Method and apparatus detecting a target | |
| CN109214343B (en) | Method and device for generating face key point detection model | |
| EP3637317B1 (en) | Method and apparatus for generating vehicle damage information | |
| US20230081645A1 (en) | Detecting forged facial images using frequency domain information and local correlation | |
| CN112884092B (en) | AI model generation method, electronic device, and storage medium | |
| US9443325B2 (en) | Image processing apparatus, image processing method, and computer program | |
| EP2879080B1 (en) | Image processing device and method, and computer readable medium | |
| CN108229673B (en) | Convolutional neural network processing method and device and electronic equipment | |
| EP3836029A1 (en) | Apparatus and method with neural network implementation of domain adaptation | |
| CN110648289B (en) | Image noise adding processing method and device | |
| WO2022121130A1 (en) | Power target detection method and apparatus, computer device, and storage medium | |
| JP6071002B2 (en) | Reliability acquisition device, reliability acquisition method, and reliability acquisition program | |
| CN112381782A (en) | Human face image quality evaluation method and device, computer equipment and storage medium | |
| CN107918688B (en) | Scene model dynamic estimation method, data analysis method and device and electronic equipment | |
| CN108847941B (en) | Identity authentication method, device, terminal and storage medium | |
| EP3674974A1 (en) | Apparatus and method with user verification | |
| CN111062324A (en) | Face detection method and device, computer equipment and storage medium | |
| CN111382791B (en) | Deep learning task processing method, image recognition task processing method and device | |
| CN110163151B (en) | Training method and device of face model, computer equipment and storage medium | |
| CN113627361B (en) | Training method and device for face recognition model and computer program product | |
| WO2015176502A1 (en) | Image feature estimation method and device | |
| CN114120454A (en) | Training method and device of living body detection model, electronic equipment and storage medium | |
| CN112036504A (en) | Temperature measurement model training method, device, equipment and storage medium | |
| CN115731341A (en) | Three-dimensional human head reconstruction method, device, equipment and medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |