CN112036293A - Age estimation method, and training method and device of age estimation model - Google Patents
Age estimation method, and training method and device of age estimation model Download PDFInfo
- Publication number
- CN112036293A CN112036293A CN202010882335.6A CN202010882335A CN112036293A CN 112036293 A CN112036293 A CN 112036293A CN 202010882335 A CN202010882335 A CN 202010882335A CN 112036293 A CN112036293 A CN 112036293A
- Authority
- CN
- China
- Prior art keywords
- age
- result
- regression
- network
- classification
- 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
- 238000000034 method Methods 0.000 title claims abstract description 100
- 238000012549 training Methods 0.000 title claims abstract description 68
- 238000012545 processing Methods 0.000 claims abstract description 17
- 238000000605 extraction Methods 0.000 claims description 32
- 230000010354 integration Effects 0.000 abstract description 4
- 230000006870 function Effects 0.000 description 17
- 238000012360 testing method Methods 0.000 description 11
- 230000004913 activation Effects 0.000 description 9
- 238000002372 labelling Methods 0.000 description 8
- 230000008569 process Effects 0.000 description 8
- 238000013136 deep learning model Methods 0.000 description 7
- 238000010586 diagram Methods 0.000 description 7
- 238000010606 normalization Methods 0.000 description 6
- 238000004891 communication Methods 0.000 description 5
- 230000014509 gene expression Effects 0.000 description 4
- 230000036544 posture Effects 0.000 description 4
- 238000013145 classification model Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000012417 linear regression Methods 0.000 description 3
- 206010011469 Crying Diseases 0.000 description 2
- 230000003778 catagen phase Effects 0.000 description 2
- 230000001815 facial effect Effects 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 238000003909 pattern recognition Methods 0.000 description 2
- 238000012706 support-vector machine Methods 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004590 computer program Methods 0.000 description 1
- 238000010924 continuous production Methods 0.000 description 1
- 239000006185 dispersion Substances 0.000 description 1
- 238000000802 evaporation-induced self-assembly Methods 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000011946 reduction process Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000012827 research and development Methods 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/161—Detection; Localisation; Normalisation
- G06V40/166—Detection; Localisation; Normalisation using acquisition arrangements
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- 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
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/168—Feature extraction; Face representation
-
- 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
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/172—Classification, e.g. identification
-
- 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
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/174—Facial expression recognition
-
- 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
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/178—Human faces, e.g. facial parts, sketches or expressions estimating age from face image; using age information for improving recognition
Abstract
The invention provides an age estimation method, an age estimation model training method and an age estimation model training device, wherein the method comprises the following steps: acquiring an image to be processed containing a target object; inputting the image to be processed into an age estimation model which is trained in advance; outputting an age classification result corresponding to the image to be processed through the classification network of the age estimation model, and outputting an age regression result corresponding to the image to be processed through the regression network of the age estimation model; determining an age of the target subject based on the age classification result and the age regression result. When the age of the target object is estimated through the age estimation model, the method can perform classification processing on the image to be processed to obtain an age classification result, perform regression processing on the image to be processed to obtain an age regression result, and determine the age of the target object according to the age classification result and the age regression result, so that the accuracy of age estimation is improved through the classification regression integration mode.
Description
Technical Field
The invention relates to the technical field of image processing, in particular to an age estimation method, an age estimation model training method and an age estimation model training device.
Background
Age is an important human face attribute, and is widely applied to the fields of human-computer interaction, intelligent commerce, safety monitoring, entertainment and the like. In the related art, the age of a person in an image can be estimated through a trained deep learning model; the deep learning model is characterized in that in the training process, an age estimation task is regarded as a classification task or a regression task, a loss value of age estimation is calculated, and then the deep learning model is trained based on the loss value.
Disclosure of Invention
In view of the above, the present invention provides an age estimation method, an age estimation model training method and an age estimation model training device, so as to improve the accuracy of age estimation.
In a first aspect, an embodiment of the present invention provides an age estimation method, where the method includes: acquiring an image to be processed containing a target object; inputting an image to be processed into an age estimation model which is trained in advance; wherein the age estimation model comprises a classification network and a regression network; outputting an age classification result corresponding to the image to be processed through a classification network, and outputting an age regression result corresponding to the image to be processed through a regression network; determining an age of the target subject based on the age classification result and the age regression result.
With reference to the first aspect, an embodiment of the present invention provides a first possible implementation manner of the first aspect, where the age estimation model further includes a feature extraction network; the step of outputting the age classification result corresponding to the image to be processed through the classification network and outputting the age regression result corresponding to the image to be processed through the regression network includes: extracting age characteristics of the image to be processed through a characteristic extraction network; classifying the age characteristics through a classification network to obtain an age classification result; and carrying out regression operation on the age characteristics through a regression network to obtain an age regression result.
With reference to the first aspect, an embodiment of the present invention provides a second possible implementation manner of the first aspect, where the age classification result includes: the probability that the target object belongs to each age value under a plurality of preset age values; the step of determining the age of the target subject based on the age classification result and the age regression result includes: calculating the average value of the age value corresponding to the maximum probability in the age classification result and the age regression result; the age of the target subject is determined based on the average.
In a second aspect, an embodiment of the present invention provides a training method for an age estimation model, where the training method includes: determining a sample image based on a preset training set; the sample image carries an age label, and the age label is used for indicating the age of the target object in the sample image; inputting the sample image into an initial model; the initial model comprises an initial classification network and an initial regression network; outputting a first result of the sample image through an initial classification network, and outputting a second result of the sample image through an initial regression network; updating the weight parameters of the initial model based on the first result, the second result and the age label; and continuing to execute the step of determining the sample image based on the preset training set until the initial model converges to obtain the age estimation model.
With reference to the second aspect, an embodiment of the present invention provides a first possible implementation manner of the second aspect, where the initial model further includes an initial feature extraction network; the step of outputting a first result of the sample image through an initial classification network and a second result of the sample image through an initial regression network includes: extracting age characteristics of the sample image through an initial characteristic extraction network; classifying the age characteristics through an initial classification network to obtain a first result; and carrying out regression operation on the age characteristics through the initial regression network to obtain a second result.
With reference to the second aspect, an embodiment of the present invention provides a second possible implementation manner of the second aspect, where the step of updating the weight parameter of the initial model based on the first result, the second result, and the age tag includes: determining a first loss value based on the first result and the age label; determining a second loss value based on the second result and the age label; updating the weight parameters of the initial model based on the first loss value and the second loss value.
With reference to the second aspect, the present invention provides a third possible implementation manner of the second aspect, where the first result includes: the probability that the target object in the sample image belongs to each age value under a plurality of preset age values; the first loss value is determined by the following equation:
the second loss value is determined by the following equation:
wherein L isclassificationRepresenting a first loss value; l isregressionRepresenting a second loss value;
representing a first result; a represents an age label;
representing a probability that the age value of the age label corresponds in the first result;
representing a second result; log represents base 2 logarithm operation.
In a third aspect, an embodiment of the present invention provides an age estimation apparatus, including: the image acquisition module is used for acquiring an image to be processed containing a target object; the image input module is used for inputting the image to be processed into the pre-trained age estimation model; wherein the age estimation model comprises a classification network and a regression network; the result output module is used for outputting an age classification result corresponding to the image to be processed through the classification network and outputting an age regression result corresponding to the image to be processed through the regression network; and the age determining module is used for determining the age of the target object based on the age classification result and the age regression result.
In a fourth aspect, an embodiment of the present invention provides a training apparatus for an age estimation model, where the training apparatus includes: the sample determining module is used for determining a sample image based on a preset training set; the sample image carries an age label, and the age label is used for indicating the age of the target object in the sample image; the sample input module is used for inputting a sample image into the initial model; the initial model comprises an initial classification network and an initial regression network; the sample processing module is used for outputting a first result of the sample image through the initial classification network and outputting a second result of the sample image through the initial regression network; the parameter adjusting module is used for updating the weight parameters of the initial model based on the first result, the second result and the age label; and continuing to execute the step of determining the sample image based on the preset training set until the initial model converges to obtain the age estimation model.
In a fifth aspect, an embodiment of the present invention provides an electronic device, which includes a processor and a memory, where the memory stores machine executable instructions capable of being executed by the processor, and the processor executes the machine executable instructions to implement the age estimation method according to any one of the foregoing embodiments or the training method of the age estimation model according to any one of the foregoing embodiments.
In a sixth aspect, embodiments of the present invention provide a machine-readable storage medium storing machine-executable instructions that, when invoked and executed by a processor, cause the processor to implement the age estimation method of any one of the preceding embodiments or the training method of the age estimation model of any one of the preceding embodiments.
The embodiment of the invention has the following beneficial effects:
the invention provides an age estimation method, an age estimation model training method and an age estimation model training device, which are characterized in that firstly, an image to be processed containing a target object is obtained; inputting the image to be processed into an age estimation model trained in advance; outputting an age classification result corresponding to the image to be processed through a classification network of the age estimation model, and outputting an age regression result corresponding to the image to be processed through a regression network of the age estimation model; the age of the target subject is then determined based on the age classification results and the age regression results. When the age of the target object is estimated through the age estimation model, the method can perform classification processing on the image to be processed to obtain an age classification result, perform regression processing on the image to be processed to obtain an age regression result, and determine the age of the target object according to the age classification result and the age regression result, so that the accuracy of age estimation is improved through the classification regression integration mode.
Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by the practice of the invention as set forth above.
In order to make the aforementioned and other objects, features and advantages of the present invention comprehensible, preferred embodiments accompanied with figures are described in detail below.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.
Fig. 1 is a flowchart of an age estimation method according to an embodiment of the present invention;
FIG. 2 is a flow chart of another age estimation method according to an embodiment of the present invention;
fig. 3 is a schematic structural diagram of an age estimation model according to an embodiment of the present invention;
fig. 4 is a flowchart of a training method of an age estimation device according to an embodiment of the present invention;
fig. 5 is a schematic structural diagram of an age estimation apparatus according to an embodiment of the present invention;
fig. 6 is a schematic structural diagram of an age estimation model training apparatus according to an embodiment of the present invention;
fig. 7 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
To make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings, and it is apparent that the described embodiments are some, but not all embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Automatic face age estimation, an important biometric identification technology, has been the subject of intense research in the field of pattern recognition and computer vision. The human face age estimation problem generally refers to that the real age of a person in a human face image is automatically estimated according to an input human face image by adopting a computer vision technology and the like.
In practical implementation, age estimation is a very specific pattern recognition problem. Specifically, if different ages are considered as different categories, the age estimation problem can be regarded as a classification task, for example, assuming that the training data set spans from 1 to 80 years of age, the age estimation can be regarded as an 80-category multi-classification problem on the data set; also, the age of a person represents the time that he/she has elapsed from birth to the present, which is a continuous process, and therefore, the age estimation problem can also be formulated as a regression problem.
In the related art, two age estimation methods are usually adopted, the first one is a traditional face age estimation algorithm, and usually, facial features (such as active appearance features, anthropometric features, biological heuristic features, and the like) in a face image need to be manually extracted, and then a classifier or regressor for obtaining the age from the facial features is trained, through which the age of a person in the face image can be estimated, but the method lacks high-level semantic information of the face, so that the accuracy of the age estimation result obtained by the method is low.
The second method is to estimate the age based on a trained deep learning model; in the training process of the deep learning model, the age estimation task is regarded as a classification task or a regression task, the loss value of age estimation is calculated, and then the deep learning model is trained based on the loss value, so that the mapping relation between the input face and the age can be established.
Based on the above description, the embodiment of the invention provides an age estimation method, an age estimation model training method and an age estimation model training device. The technology can be applied to the age estimation scene in the fields of human-computer interaction, intelligent commerce, safety monitoring, entertainment and the like. To facilitate understanding of the present embodiment, an age estimation method disclosed in the present embodiment will be described in detail first, and as shown in fig. 1, the method includes the following steps:
step S102, acquiring an image to be processed containing a target object.
The target object generally refers to a human face of a person in an image, and the human face may have various postures and expressions, but one human face in one image has only one posture or expression, and for example, the human face may be a front face, a side face, a smiling face, a crying face and the like. The image to be processed may be a picture or a photograph taken by a video camera or a still camera, or may be a certain video frame obtained from a specific video file. In a specific implementation, the image to be processed may be acquired by: the images are taken by a camera, a camera head and the like connected through communication and then transmitted into the storage device, or are acquired from the storage device storing the images which are already taken.
Step S104, inputting the image to be processed into an age estimation model which is trained in advance; wherein the age estimation model comprises a classification network and a regression network.
The age estimation model may adopt a deep learning model or a neural network model, and includes an independent classification network and a regression network in order to fully consider the complementarity of classification and regression of the age estimation problem. The age estimation model is usually obtained through machine learning training according to a preset training set, in the training process, the age estimation task is regarded as a classification task and a regression task, then the age estimation model is trained according to a loss value corresponding to a classification network and a loss value corresponding to a regression network until the model converges, and the trained age estimation model is obtained.
And S106, outputting an age classification result corresponding to the image to be processed through the classification network, and outputting an age regression result corresponding to the image to be processed through the regression network.
The classification network in the age estimation model can classify the characteristics of the image to be processed to obtain an age classification result; the regression network in the age estimation model can perform regression operation on the features of the image to be processed to obtain an age regression result. In a specific implementation, the age classification result may be a probability value that the age of the target object is equal to each age value under a plurality of preset age values, for example, the plurality of preset age values are integers between 0 and 100, that is, represent 0 to 100 years old, and then the age classification result is a probability value that the target object is equal to 0 to 100 years old, respectively. The age regression result is usually a real number representing the age prediction value of the regression network output.
Step S108, determining the age of the target object based on the age classification result and the age regression result.
The age classification result and the age result are fused to obtain the age of the target object, and in concrete implementation, the age value corresponding to the maximum value in the age classification result and the age regression result can be averaged to determine the average value as the age of the target object; or obtaining an estimated age value from an age classification result according to a preset rule (the preset rule can be set according to research and development requirements), and then taking the average value of the estimated age value and an age regression result as the age of the target object; the larger or smaller of the age value corresponding to the maximum value in the age classification results and the age value corresponding to the age regression result may also be taken as the age of the target object.
The invention provides an age estimation method, firstly, acquiring an image to be processed containing a target object; inputting the image to be processed into an age estimation model trained in advance; outputting an age classification result corresponding to the image to be processed through a classification network of the age estimation model, and outputting an age regression result corresponding to the image to be processed through a regression network of the age estimation model; the age of the target subject is then determined based on the age classification results and the age regression results. When the age of the target object is estimated through the age estimation model, the method can perform classification processing on the image to be processed to obtain an age classification result, perform regression processing on the image to be processed to obtain an age regression result, and determine the age of the target object according to the age classification result and the age regression result, so that the accuracy of age estimation is improved through the classification regression integration mode.
The embodiment of the invention also provides another age estimation method, which is realized on the basis of the method of the embodiment; in the method, a specific process (realized by steps S206 to S210 described below) of outputting an age classification result corresponding to an image to be processed through the classification network, outputting an age regression result corresponding to the image to be processed through the regression network, and a specific process (realized by steps S212 to S214 described below) of determining the age of a target object based on the age classification result and the age regression result is described in detail, in a case where the age estimation model includes a feature extraction network, a classification network, and a regression network; as shown in fig. 2, the method comprises the steps of:
step S202, acquiring an image to be processed containing a target object.
Step S204, inputting the image to be processed into an age estimation model which is trained in advance; the age estimation model includes a feature extraction network, a classification network, and a regression network.
And step S206, extracting the age characteristics of the image to be processed through the characteristic extraction network.
In the age estimation model, a feature extraction network is connected with a classification network and a regression network, respectively. The feature extraction network is used for receiving an input image to be processed, extracting age features in the image to be processed, and inputting the age features into the classification network and the regression network respectively.
The feature extraction network can extract the image features of the image to be processed to obtain the age features, so that high-level semantic information of the image features can be obtained. The feature extraction network may include a convolutional layer and an activation function layer connected in sequence, where the activation function layer may perform function transformation on a feature map output by the convolutional layer, the transformation process may break a linear combination of inputs of the convolutional layer, and the activation function layer may specifically be a Sigmoid function, a tanh function, a Relu function, or the like. In order to improve the performance of the feature extraction network, the feature extraction network may generally include multiple groups of sequentially connected convolution layers and activation function layers, specifically including how many groups of sequentially connected convolution layers and activation function layers, and how many sequentially connected convolution layers and activation function layers each group includes, which may be determined by the speed and precision requirements of the specific application.
In some embodiments, the convolutional layer and the activation function in the feature extraction network can be connected through a normalization layer, the normalization layer can perform normalization processing on the feature map output by the convolutional layer, the convergence speed of the feature extraction network and the model can be increased, and the problem of gradient dispersion in multilayer convolution can be solved, so that the feature extraction network is more stable.
Step S208, classifying the age characteristics through a classification network to obtain an age classification result; the age classification result includes: the probability that the target object belongs to each age value under a plurality of preset age values.
The preset age values may be an age range set by the user according to actual needs, and may be, for example, an integer between 0 and 80, representing 81 categories of 0 to 80 years old; may be an integer between 0 and 100 and represents the 101 categories of 0 to 100 years old. When the classification network receives the age characteristics input by the characteristic extraction network, the classification network classifies the age characteristics according to a plurality of preset age values to obtain a probability value of classifying the age characteristics into each age value in the plurality of preset age values, wherein the probability value is the probability of the target object belonging to each age value.
In a specific implementation, the classification network may be an existing classification model, such as a Support Vector Machine (SVM), a Linear Regression (LR) or the like; or may be a network composed of one or more Fully connected layers (FC), where the Fully connected layers may output a vector corresponding to an age classification result of a specified dimension (the dimension is usually the same as the number of preset multiple age values), and a numerical value of each dimension in the vector corresponds to a probability of one age value.
The probability in the age classification result may be normalized probability or non-normalized probability, and if the non-normalized probability needs to be converted into the normalized probability, a normalization layer may be added at the end of the classification network, where the normalization layer includes a softmax function, that is, the probability is normalized by the softmax function to obtain the probability in the range of 0 to 1. For example, assume that the non-normalized age classification results in
Will be provided with
Each element in the table is sent into a softmax function, and an age classification result after probability normalization can be obtained
Wherein the content of the first and second substances,
representing the probability corresponding to the jth age value in the normalized age classification result;
representing the probability corresponding to the jth age value in the non-normalized age classification result;
representing the mth age in the non-normalized age classification to the corresponding probability.
Step S210, performing regression operation on the age characteristics through a regression network to obtain an age regression result.
In a specific implementation, the regression network may be an existing regression model, such as a neural network regression model, an autoregressive model, a linear regression model, etc.; or the real number is a predicted value of the age of the target object in the image to be processed by the regression network.
In order to facilitate understanding of the age estimation model in the embodiment of the present invention, a schematic structural diagram of an age estimation model is shown in fig. 3. Block1, Block2 and FC1 in fig. 3 form a feature extraction network, wherein Block1 is composed of a set of convolutional layers and activation function layers, Block2 is also composed of a set of convolutional layers and activation function layers, and FC1 represents a fully connected layer; FC2 in fig. 3 represents a classification network and FC3 represents a regression network, where FC2 and FC3 are both fully connected layers.
In specific implementation, an image to be processed is input into Block1, and after processing of Block2 and FC1, an age vector with dimension c (the value of c is set according to task requirements, and the larger the value of c is, the better the effect is) is obtained, and the age vector is also the extracted age feature of the image to be processed; the dimension of the data output by Block2 is usually larger, and FC1 can be understood as a dimension reduction process, that is, the feature vector output by Block2 is reduced to c dimension to obtain an age vector. And then the age characteristics are respectively sent to a classification network and a regression network, so that an age classification result which is output by the classification network and has the same dimension with the number of the preset multiple age values and a one-dimensional age regression result output by the regression network can be obtained.
In a specific implementation, the weight parameters of each network in the age estimation model are determined according to the loss values in the process of machine learning; the loss value is determined according to an age classification result output by the classification network, an age regression result output by the regression network and an age label corresponding to the sample image; the age label is used for indicating the age of a target object contained in the sample image; the specific training process of the age estimation model will be described in detail in the following embodiments of the training method of the age estimation model, and will not be described herein again.
In step S212, an average value of the age regression result and the age value corresponding to the age classification result with the highest probability is calculated.
In step S214, the age of the target object is determined based on the average value.
In a particular implementation, if the average is an integer, the average may be determined as the age of the target subject; if the age average is decimal, the target age can be obtained by rounding up, rounding down, or rounding down.
Firstly, acquiring an image to be processed containing a target object, and then inputting the image to be processed into an age estimation model which is trained in advance; extracting the age characteristics of the image to be processed through a characteristic extraction network; classifying the age characteristics through a classification network to obtain an age classification result; performing regression operation on the age characteristics through a regression network to obtain an age regression result; then calculating the average value of the age value corresponding to the maximum probability in the age classification result and the age regression result; finally, the age of the target subject is determined based on the average. When the age estimation model in the method estimates the age, the multi-level semantic features related to the age in the image to be processed can be automatically learned, so that the accuracy of age estimation can be improved; meanwhile, the method fuses the age classification result output by the classification network and the age regression result output by the regression network to obtain the age of the target object, and the fusion of the results can be regarded as a type of integrated learning, so that the method can further improve the accuracy of age estimation compared with a method of estimating the age by using classification or regression alone.
For the embodiment of the age estimation method, an embodiment of the present invention further provides a training method of an age estimation model, where the age estimation model is the model for estimating age adopted in the embodiment, and as shown in fig. 4, the training method includes the following steps:
step S402, determining a sample image based on a preset training set; the sample image carries an age label indicating the age of the target object in the sample image.
The training set includes a plurality of samples, each sample includes a sample image and an age tag corresponding to the sample image, and each sample image includes a target object, which is the face of a person in the sample image, and the face may have a plurality of postures and expressions, but the face in the sample image has only one posture or expression, for example, the face may be a front face, a side face, a smiling face, a crying face, or the like. The sample image may be a picture or a photograph taken by a video camera or a still camera in advance, or may be a certain video frame obtained from a video file stored in advance, and specifically, the sample image may be identified by X, X e RH×W×3Where H represents height, W represents width, 3 represents RGB (R represents Red, G represents Green, B represents Blue, Blue) three channels, and R represents the training set.
The age label identifies the age of the person in the sample image, which can be determined by the following steps 10-11:
step 10, obtaining a plurality of labeling results corresponding to the sample image; the labeling result is used for identifying the age value of the target object in the sample image; the labeled age value in the labeling result is one of a plurality of preset age values.
The plurality of preset age values are age values within an age range set by a user, and for example, the age values may be set to be integers between 0 and 100, which represent 0 to 100 years, respectively. In specific implementation, preset n persons perform age annotation on a target object in a sample image to obtain n annotation results, where the n annotation results are a plurality of annotation results corresponding to the sample image.
And 11, calculating the average value of the age values corresponding to the plurality of labeling results, and determining the average value as the age label of the sample image.
For example, assuming that a plurality of preset age values are integers between 0 and 100, n persons perform age labeling on a target object in a sample image to obtain n labeling results
Wherein k has a value ranging from 1 to n,
and the age label which represents the labeling result of the kth person on the sample image and can obtain the sample image according to the n labeling results is as follows:
wherein a represents an age label of the sample image;
represents rounding down.
Step S404, inputting the sample image into an initial model; the initial model includes an initial classification network and an initial regression network.
Step S406, outputting a first result of the sample image through the initial classification network, and outputting a second result of the sample image through the initial regression network.
In specific implementation, the initial model comprises an initial feature extraction network, an initial classification model and an initial regression model; the above step S406 can be realized by the following steps 20 to 22:
and 20, extracting the age characteristics of the sample image through the initial characteristic extraction network.
In the initial model, an initial feature extraction network is respectively connected with an initial classification network and an initial regression network, and the initial feature extraction network is used for receiving an input sample image, extracting age features in the sample image and respectively inputting the extracted age features to the classification network and the regression network.
And step 21, classifying the age characteristics through an initial classification network to obtain a first result.
And step 22, performing regression operation on the age characteristics through the initial regression network to obtain a second result.
The first result includes: the probability that the target object in the sample image belongs to each age value under a plurality of preset age values; the second result is an age value of the target object predicted by the regression network. The structures of the initial feature extraction network, the initial classification model and the initial regression model may refer to the structures in the schematic diagram shown in fig. 3, but the structure in the schematic diagram in fig. 3 is only an example, and may also be implemented by other structures.
Step S408, updating the weight parameters of the initial model based on the first result, the second result and the age label; and continuing to execute the step of determining the sample image based on the preset training set until the initial model converges to obtain the age estimation model.
In particular implementations, a first loss value may be determined based on the first result and the age tag; determining a second loss value based on the second result and the age label; the weight parameters of the initial model are then updated based on the first penalty value and the second penalty value. Specifically, the first loss value may be determined by the following equation:
the second loss value may be determined by the following equation:
wherein L isclassificationRepresenting a first loss value; l isregressionRepresenting a second loss value;
representing a first result; a represents an age label;
representing a probability that the age value of the age label corresponds in the first result;
representing a second result; log represents base 2 logarithm operation.
The first loss value may also be referred to as a classification loss value, and generally the greater the probability of the age value corresponding to the age tag in the first result, the smaller the first loss value, and therefore, as the training is performed, the age value corresponding to the age tag in the first result is larger, that is, the classification result is closer to the true value (that is, the age value corresponding to the age tag); the second loss value may also be referred to as a regression loss value, and represents a distance between the age regression result and the age label, and the closer the age regression result is to the age label, the smaller the second loss value is, and thus the age regression result gets closer to the true value as the training progresses.
In particular implementations, the model loss value of the initial model may be summed with the first loss value and the second loss value:
L＝Lclassification+Lregression；
wherein, L represents a model loss value, and then the weight parameter of the initial model may be updated according to the model loss value, and the specific implementation manner may be the following steps 30 to 33:
step 30, calculating the derivative of the model loss value to the weight parameter to be updated in the initial model
W represents a weight parameter to be updated; the weight parameters to be updated can be all parameters in the initial model, and can also be partial parameters randomly determined from the initial model; the updated weight parameter is also the weight of each layer of network in the initial model. The derivative of the weight parameter to be updated can be solved according to a back propagation algorithm in general; if the model loss value is larger, the difference between the estimation result of the current initial model and the expected result is more, the derivative of the model loss value to the weight parameter to be updated in the initial model is solved, and the derivative can be used as the basis for updating the weight parameter to be updated.
Step 31, updating the weight parameter to be updated to obtain the updated weight parameter to be updated
Where α is a predetermined coefficient, which is a manually predetermined hyper-parameter, for example, a value of 0.01, 0.001, etc.
Step 32, judging whether the weight parameters of the updated initial model are all converged, and if the weight parameters are all converged, executing step 402; otherwise, step 33 is executed.
And step 33, determining the initial model after the parameters are updated as the trained age estimation model.
In a specific implementation, sample images in a preset sample data set may be divided into a training set for training a model and a test set for verifying the model according to a preset ratio (e.g., 10:1 or 20: 1). The identification precision of the trained age estimation model can be determined through the test set; generally, a test sample can be determined from a test set, the test sample comprises a sample image and an age label corresponding to the sample image, the test sample is input into an age estimation model trained by a training set, and an age classification result and an age regression result corresponding to the sample image can be output; determining the age of the target object based on the age classification result and the age regression result, comparing the age with the age label, judging whether the age is correct, and continuously determining the test samples from the test set until all the samples in the test set are selected; and counting the correctness corresponding to the test result corresponding to each test sample to obtain the prediction precision of the trained age estimation model.
The training method of the age estimation model comprises the steps of firstly determining a sample image based on a preset training set; inputting the sample image into an initial model; outputting a first result of the sample image through an initial classification network of the initial model, and outputting a second result of the sample image through an initial regression network of the initial model; and then updating the weight parameters of the initial model based on the first result, the second result and the age label until the initial model converges to obtain an age estimation model. The age estimation model in the method can automatically learn the multi-level semantic features related to the age, so that the accuracy of age estimation can be improved; meanwhile, the method takes the age estimation task as both the classification task and the regression task, and trains the age estimation model through the classification loss and the regression loss at the same time, so that the method integrates the classification regression phase and can further improve the accuracy of the age estimation compared with a method of carrying out the age estimation by using the classification or the regression alone.
Corresponding to the embodiment of the age estimation method, an embodiment of the present invention further provides an age estimation apparatus, as shown in fig. 5, the apparatus including:
an image obtaining module 50, configured to obtain an image to be processed including a target object.
An image input module 51, configured to input an image to be processed into an age estimation model trained in advance; wherein the age estimation model comprises a classification network and a regression network.
And the result output module 52 is configured to output an age classification result corresponding to the image to be processed through the classification network, and output an age regression result corresponding to the image to be processed through the regression network.
And an age determining module 53 for determining the age of the target object based on the age classification result and the age regression result.
The age estimation device firstly acquires an image to be processed containing a target object; inputting the image to be processed into an age estimation model trained in advance; outputting an age classification result corresponding to the image to be processed through a classification network of the age estimation model, and outputting an age regression result corresponding to the image to be processed through a regression network of the age estimation model; the age of the target subject is then determined based on the age classification results and the age regression results. When the age of the target object is estimated through the age estimation model, the method can perform classification processing on the image to be processed to obtain an age classification result, perform regression processing on the image to be processed to obtain an age regression result, and determine the age of the target object according to the age classification result and the age regression result, so that the accuracy of age estimation is improved through the classification regression integration mode.
Specifically, the age estimation model further includes a feature extraction network; the result output module 52 is further configured to: extracting age characteristics of the image to be processed through a characteristic extraction network; classifying the age characteristics through a classification network to obtain an age classification result; and carrying out regression operation on the age characteristics through a regression network to obtain an age regression result.
Further, the age classification result includes: the probability that the target object belongs to each age value under a plurality of preset age values; the age determining module 53 is configured to: calculating the average value of the age value corresponding to the maximum probability in the age classification result and the age regression result; the age of the target subject is determined based on the average.
The age estimation apparatus provided in the embodiment of the present invention has the same implementation principle and technical effect as those of the age estimation method embodiment, and for the sake of brief description, reference may be made to the corresponding contents in the method embodiment for the part not mentioned in the apparatus embodiment.
Corresponding to the above embodiment of the training method of the age estimation model, an embodiment of the present invention further provides a training apparatus of an age estimation model, as shown in fig. 6, the training apparatus includes:
a sample determination module 60, configured to determine a sample image based on a preset training set; the sample image carries an age label indicating the age of the target object in the sample image.
A sample input module 61, configured to input a sample image into the initial model; the initial model includes an initial classification network and an initial regression network.
And a sample processing module 62, configured to output a first result of the sample image through the initial classification network and output a second result of the sample image through the initial regression network.
A parameter adjusting module 63, configured to update the weight parameter of the initial model based on the first result, the second result and the age label; and continuing to execute the step of determining the sample image based on the preset training set until the initial model converges to obtain the age estimation model.
The training device of the age estimation model firstly determines a sample image based on a preset training set; inputting the sample image into an initial model; outputting a first result of the sample image through an initial classification network of the initial model, and outputting a second result of the sample image through an initial regression network of the initial model; and then updating the weight parameters of the initial model based on the first result, the second result and the age label until the initial model converges to obtain an age estimation model. The age estimation model in the method can automatically learn the multi-level semantic features related to the age, so that the accuracy of age estimation can be improved; meanwhile, the method takes the age estimation task as both the classification task and the regression task, and trains the age estimation model through the classification loss and the regression loss at the same time, so that the method integrates the classification regression phase and can further improve the accuracy of the age estimation compared with a method of carrying out the age estimation by using the classification or the regression alone.
Specifically, the initial model further includes an initial feature extraction network; the sample processing module 62 is configured to: extracting age characteristics of the sample image through an initial characteristic extraction network; classifying the age characteristics through an initial classification network to obtain a first result; and carrying out regression operation on the age characteristics through an initial regression network to obtain a second result.
Further, the parameter adjusting module 63 is configured to: determining a first loss value based on the first result and the age label; determining a second loss value based on the second result and the age label; the weight parameters of the initial model are updated based on the first loss value and the second loss value.
In a specific implementation, the first result includes: the probability that the target object in the sample image belongs to each age value under a plurality of preset age values; the first loss value is determined by the following equation:
the second loss value is determined by the following equation:
wherein L isclassificationRepresenting a first loss value; l isregressionRepresenting a second loss value;
representing a first result; a represents an age label;
representing a probability that the age value of the age label corresponds in the first result;
representing a second result; log represents base 2 logarithm operation.
The implementation principle and the generated technical effect of the training device of the age estimation model provided by the embodiment of the invention are the same as those of the embodiment of the training method of the age estimation model, and for the sake of brief description, corresponding contents in the embodiment of the method can be referred to where the embodiment of the device is not mentioned.
An embodiment of the present invention further provides an electronic device, which is shown in fig. 7 and includes a processor 101 and a memory 100, where the memory 100 stores machine executable instructions that can be executed by the processor 101, and the processor executes the machine executable instructions to implement the age estimation method or the training method of the age estimation model.
Further, the electronic device shown in fig. 7 further includes a bus 102 and a communication interface 103, and the processor 101, the communication interface 103, and the memory 100 are connected through the bus 102.
The Memory 100 may include a high-speed Random Access Memory (RAM) and may further include a non-volatile Memory (non-volatile Memory), such as at least one disk Memory. The communication connection between the network element of the system and at least one other network element is realized through at least one communication interface 103 (which may be wired or wireless), and the internet, a wide area network, a local network, a metropolitan area network, and the like can be used. The bus 102 may be an ISA bus, PCI bus, EISA bus, or the like. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of illustration, only one double-headed arrow is shown in FIG. 7, but this does not indicate only one bus or one type of bus.
The processor 101 may be an integrated circuit chip having signal processing capabilities. In implementation, the steps of the above method may be performed by integrated logic circuits of hardware or instructions in the form of software in the processor 101. The Processor 101 may be a general-purpose Processor, and includes a Central Processing Unit (CPU), a Network Processor (NP), and the like; the device can also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other Programmable logic device, a discrete Gate or transistor logic device, or a discrete hardware component. The various methods, steps and logic blocks disclosed in the embodiments of the present invention may be implemented or performed. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like. The steps of the method disclosed in connection with the embodiments of the present invention may be directly implemented by a hardware decoding processor, or implemented by a combination of hardware and software modules in the decoding processor. The software module may be located in ram, flash memory, rom, prom, or eprom, registers, etc. storage media as is well known in the art. The storage medium is located in the memory 100, and the processor 101 reads the information in the memory 100, and completes the steps of the method of the foregoing embodiment in combination with the hardware thereof.
An embodiment of the present invention further provides a machine-readable storage medium, where the machine-readable storage medium stores machine-executable instructions, and when the machine-executable instructions are called and executed by a processor, the machine-executable instructions cause the processor to implement the age estimation method or the training method for the age estimation model, and specific implementation may refer to method embodiments and will not be described herein again.
The age estimation method, the age estimation model training method, and the computer program product of the apparatus provided in the embodiments of the present invention include a computer-readable storage medium storing a program code, where instructions included in the program code may be used to execute the method described in the foregoing method embodiments, and specific implementations may refer to the method embodiments and are not described herein again.
The functions, if implemented in the form of software functional units and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, an electronic device, or a network device) to perform all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
Finally, it should be noted that: the above embodiments are only used to illustrate the technical solution of the present invention, and not to limit the same; while the invention has been described in detail and with reference to the foregoing embodiments, it will be understood by those skilled in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some or all of the technical features may be equivalently replaced; and the modifications or the substitutions do not make the essence of the corresponding technical solutions depart from the scope of the technical solutions of the embodiments of the present invention.
Claims (11)
1. A method of age estimation, the method comprising:
acquiring an image to be processed containing a target object;
inputting the image to be processed into an age estimation model which is trained in advance; wherein the age estimation model comprises a classification network and a regression network;
outputting an age classification result corresponding to the image to be processed through the classification network, and outputting an age regression result corresponding to the image to be processed through the regression network;
determining an age of the target subject based on the age classification result and the age regression result.
2. The method of claim 1, wherein the age estimation model further comprises a feature extraction network;
the step of outputting the age classification result corresponding to the image to be processed through the classification network and outputting the age regression result corresponding to the image to be processed through the regression network comprises the following steps:
extracting age characteristics of the image to be processed through the characteristic extraction network;
classifying the age characteristics through the classification network to obtain an age classification result;
and carrying out regression operation on the age characteristics through the regression network to obtain an age regression result.
3. The method of claim 1, wherein the age classification result comprises: a probability that the target object belongs to each of a plurality of preset age values;
the step of determining the age of the target subject based on the age classification result and the age regression result comprises:
calculating the average value of the age value corresponding to the maximum probability in the age classification result and the age regression result;
determining an age of the target subject based on the average.
4. A method for training an age estimation model, the method comprising:
determining a sample image based on a preset training set; the sample image carries an age label, and the age label is used for indicating the age of the target object in the sample image;
inputting the sample image into an initial model; the initial model comprises an initial classification network and an initial regression network;
outputting a first result of the sample image through the initial classification network and outputting a second result of the sample image through the initial regression network;
updating a weight parameter of the initial model based on the first result, the second result, and the age tag; and continuing to execute the step of determining the sample image based on the preset training set until the initial model converges to obtain the age estimation model.
5. The training method of claim 4, wherein the initial model further comprises an initial feature extraction network; the step of outputting a first result of the sample image through the initial classification network and a second result of the sample image through the initial regression network comprises:
extracting age characteristics of the sample image through the initial characteristic extraction network;
classifying the age characteristics through the initial classification network to obtain a first result;
and carrying out regression operation on the age characteristics through the initial regression network to obtain a second result.
6. The training method of claim 4, wherein the step of updating the weight parameters of the initial model based on the first result, the second result and the age label comprises:
determining a first loss value based on the first result and the age tag; determining a second loss value based on the second result and the age label;
updating a weight parameter of the initial model based on the first penalty value and the second penalty value.
7. Training method according to claim 6, characterized in that said first result comprises: a probability that a target object in the sample image belongs to each of the age values at a preset plurality of age values; the first loss value is determined by the following equation:
the second loss value is determined by the following equation:
wherein L isclassificationRepresenting the first loss value; l isregressionRepresenting the second loss value;
representing the first result; a represents the age label;
representing a probability that the age value of the age tag corresponds in the first result;
representing the second result; log represents base 2 logarithm operation.
8. An age estimation apparatus, characterized in that the apparatus comprises:
the image acquisition module is used for acquiring an image to be processed containing a target object;
the image input module is used for inputting the image to be processed into an age estimation model which is trained in advance; wherein the age estimation model comprises a classification network and a regression network;
the result output module is used for outputting the age classification result corresponding to the image to be processed through the classification network and outputting the age regression result corresponding to the image to be processed through the regression network;
an age determination module to determine an age of the target subject based on the age classification result and the age regression result.
9. An apparatus for training an age estimation model, the apparatus comprising:
the sample determining module is used for determining a sample image based on a preset training set; the sample image carries an age label, and the age label is used for indicating the age of the target object in the sample image;
a sample input module for inputting the sample image into an initial model; the initial model comprises an initial classification network and an initial regression network;
the sample processing module is used for outputting a first result of the sample image through the initial classification network and outputting a second result of the sample image through the initial regression network;
a parameter adjustment module for updating a weight parameter of the initial model based on the first result, the second result and the age label; and continuing to execute the step of determining the sample image based on the preset training set until the initial model converges to obtain the age estimation model.
10. An electronic device comprising a processor and a memory, the memory storing machine executable instructions executable by the processor, the processor executing the machine executable instructions to implement the age estimation method of any one of claims 1 to 3 or the training method of the age estimation model of any one of claims 4 to 7.
11. A machine-readable storage medium having stored thereon machine-executable instructions which, when invoked and executed by a processor, cause the processor to implement the age estimation method of any of claims 1 to 3 or the training method of the age estimation model of any of claims 4 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010882335.6A CN112036293A (en) | 2020-08-27 | 2020-08-27 | Age estimation method, and training method and device of age estimation model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010882335.6A CN112036293A (en) | 2020-08-27 | 2020-08-27 | Age estimation method, and training method and device of age estimation model |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112036293A true CN112036293A (en) | 2020-12-04 |
Family
ID=73586737
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010882335.6A Pending CN112036293A (en) | 2020-08-27 | 2020-08-27 | Age estimation method, and training method and device of age estimation model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112036293A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113298004A (en) * | 2021-06-03 | 2021-08-24 | 南京佑驾科技有限公司 | Lightweight multi-head age estimation method based on face feature learning |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2010125916A1 (en) * | 2009-04-28 | 2010-11-04 | Ｎｅｃソフト株式会社 | Age estimation device, age estimation method, and program |
| CN103984948A (en) * | 2014-06-03 | 2014-08-13 | 南京大学 | Soft double-layer age estimation method based on facial image fusion features |
| WO2019091402A1 (en) * | 2017-11-09 | 2019-05-16 | 北京达佳互联信息技术有限公司 | Method and device for age estimation |
| CN110287942A (en) * | 2019-07-03 | 2019-09-27 | 成都旷视金智科技有限公司 | Training method, age estimation method and the corresponding device of age estimation model |
| CN111126207A (en) * | 2019-12-12 | 2020-05-08 | 北京金山云网络技术有限公司 | Training method and device of age estimation model and electronic equipment |
-
2020
- 2020-08-27 CN CN202010882335.6A patent/CN112036293A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2010125916A1 (en) * | 2009-04-28 | 2010-11-04 | Ｎｅｃソフト株式会社 | Age estimation device, age estimation method, and program |
| CN103984948A (en) * | 2014-06-03 | 2014-08-13 | 南京大学 | Soft double-layer age estimation method based on facial image fusion features |
| WO2019091402A1 (en) * | 2017-11-09 | 2019-05-16 | 北京达佳互联信息技术有限公司 | Method and device for age estimation |
| CN110287942A (en) * | 2019-07-03 | 2019-09-27 | 成都旷视金智科技有限公司 | Training method, age estimation method and the corresponding device of age estimation model |
| CN111126207A (en) * | 2019-12-12 | 2020-05-08 | 北京金山云网络技术有限公司 | Training method and device of age estimation model and electronic equipment |
Non-Patent Citations (3)
| Title |
|---|
| 赵一丁;田森平;: "基于分类与回归混合模型的人脸年龄估计方法", 计算机应用, no. 07 * |
| 赵一丁等: "基于分类与回归混合模型的人脸年龄估计方法", 《计算机应用》, vol. 37, no. 7, pages 1 - 2 * |
| 陈敬;李寿山;王晶晶;周国栋;: "混合分类/回归模型的用户年龄识别方法", 中国科学:信息科学, no. 08 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113298004A (en) * | 2021-06-03 | 2021-08-24 | 南京佑驾科技有限公司 | Lightweight multi-head age estimation method based on face feature learning |
| CN113298004B (en) * | 2021-06-03 | 2022-04-29 | 南京佑驾科技有限公司 | Lightweight multi-head age estimation method based on face feature learning |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110070067B (en) | Video classification method, training method and device of video classification method model and electronic equipment | |
| CN111898547B (en) | Training method, device, equipment and storage medium of face recognition model | |
| CN109815826B (en) | Method and device for generating face attribute model | |
| CN109902202B (en) | Video classification method and device | |
| US20210182535A1 (en) | Method for selecting images in video of faces in the wild | |
| CN111401339B (en) | Method and device for identifying age of person in face image and electronic equipment | |
| CN108197592B (en) | Information acquisition method and device | |
| CN111582150A (en) | Method and device for evaluating face quality and computer storage medium | |
| CN111967382A (en) | Age estimation method, and training method and device of age estimation model | |
| CN112085000A (en) | Age identification method, and training method and device of age identification model | |
| CN112070044A (en) | Video object classification method and device | |
| CN111401343A (en) | Method for identifying attributes of people in image and training method and device for identification model | |
| CN109325435B (en) | Video action recognition and positioning method based on cascade neural network | |
| CN112950579A (en) | Image quality evaluation method and device and electronic equipment | |
| CN110414541B (en) | Method, apparatus, and computer-readable storage medium for identifying an object | |
| CN111046747B (en) | Crowd counting model training method, crowd counting method, device and server | |
| CN111915595A (en) | Image quality evaluation method, and training method and device of image quality evaluation model | |
| CN111814846A (en) | Training method and recognition method of attribute recognition model and related equipment | |
| CN111967383A (en) | Age estimation method, and training method and device of age estimation model | |
| CN115205615A (en) | Vehicle cleanliness detection method and device and computer readable storage medium | |
| CN112036293A (en) | Age estimation method, and training method and device of age estimation model | |
| CN113222043A (en) | Image classification method, device, equipment and storage medium | |
| CN111914772A (en) | Method for identifying age, and training method and device of age identification model | |
| CN112380919A (en) | Vehicle category statistical method | |
| CN116959099A (en) | Abnormal behavior identification method based on space-time diagram convolutional neural network |
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 |