CN112307900A - Method and device for evaluating facial image quality and electronic equipment - Google Patents
Method and device for evaluating facial image quality and electronic equipment Download PDFInfo
- Publication number
- CN112307900A CN112307900A CN202011037197.8A CN202011037197A CN112307900A CN 112307900 A CN112307900 A CN 112307900A CN 202011037197 A CN202011037197 A CN 202011037197A CN 112307900 A CN112307900 A CN 112307900A
- Authority
- CN
- China
- Prior art keywords
- face orientation
- face
- evaluated
- image
- training
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
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
-
- 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
-
- 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
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- Software Systems (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Mathematical Physics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computing Systems (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Human Computer Interaction (AREA)
- Multimedia (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Medical Informatics (AREA)
- Image Analysis (AREA)
Abstract
The invention provides a method and a device for evaluating the quality of a face image and electronic equipment, wherein the method comprises the following steps: acquiring a facial image to be evaluated of a target object; adopting a face orientation detection model to perform face orientation detection on a face image to be evaluated to obtain the confidence coefficient of face orientation angle information of a target object; and taking the confidence as the image quality of the facial image to be evaluated. In the invention, the training of the face orientation detection model is simple, only the face orientation angle of the training object in the training sample image needs to be labeled, the labeling cost is low, in addition, the face orientation angle parameter is more objective and global, when the face orientation angle information output by the model is more accurate (namely the confidence coefficient is higher), the more effective information the model acquires from the face image to be evaluated is, and the quality of the face image to be evaluated is better.
Description
Technical Field
The present invention relates to the field of image processing technologies, and in particular, to a method and an apparatus for evaluating facial image quality, and an electronic device.
Background
There are two methods for evaluating the quality of facial images. One is to give labels to all cases such as occlusion, large angle, blur, non-target object face, etc., and to learn these data and labels directly from the model. And the other method is to label the occlusion degree, the occlusion area and the face direction to the face image sample respectively, learn all the information by the model, and finally manually give the weight corresponding to the attributes to calculate the final face image quality.
Both of the above methods define image quality manually, and the quality of these facial images is not good for the "human eye", but for the model, some human eyes consider that low quality images may belong to recognizable images for the model, and some human eyes consider that recognizable images may belong to low quality images for the model. The results obtained by the above method therefore do not fit well to the characteristics of the model itself. In addition, the two methods need to label various types of image quality (including the faces of the shielding, blurring, large-angle and non-target objects), the labeling types are wide, and the labeling cost is high; in addition, in the two methods, the annotation data often only covers image quality attributes of some aspects, and it is difficult to traverse all the image quality attributes (for example, whether the image quality is good or bad includes whether noise exists, whether coding and decoding errors exist, and the like), so that the image quality information learned by the model is incomplete, and the accuracy of the finally determined face image quality is poor.
In conclusion, in the existing facial image quality evaluation method, the labeling cost is high during model training, the labeled samples are not comprehensive and objective enough, and the accuracy of the finally determined facial image quality is poor.
Disclosure of Invention
In view of the above, the present invention provides a method, an apparatus, and an electronic device for evaluating facial image quality, so as to alleviate the technical problems of high labeling cost, insufficient comprehensive and objective labeling samples, and poor accuracy of the finally determined facial image quality in the existing method for evaluating facial image quality.
In a first aspect, an embodiment of the present invention provides a method for evaluating facial image quality, including: acquiring a facial image to be evaluated of a target object; performing face orientation detection on the face image to be evaluated by adopting a face orientation detection model to obtain a confidence coefficient of the face orientation angle information of the target object, wherein the confidence coefficient is used for representing the accuracy degree of the face orientation angle information of the target object; and taking the confidence as the image quality of the facial image to be evaluated.
Further, after obtaining the image quality of the facial image to be evaluated, the method further includes: acquiring a preset confidence level threshold; and screening the facial image to be evaluated based on the preset confidence level threshold value and the image quality of the facial image to be evaluated to obtain a high-quality facial image to be evaluated, wherein the image quality of the high-quality facial image to be evaluated is greater than the preset confidence level threshold value.
Further, the face orientation detection model includes: a face orientation branch and a confidence branch; the face orientation branch is used for detecting face orientation angle information of a target object in the face image to be evaluated according to the face image to be evaluated; the confidence branch is used for detecting the confidence of the face orientation angle information of the target object in the facial image to be evaluated according to the facial image to be evaluated.
Further, the method further comprises: acquiring a plurality of training sample images containing faces of training objects and face orientation angle information of the training objects in each training sample image; and training an original model of the face orientation detection model according to the plurality of training sample images and the face orientation angle information of the training object in each training sample image to obtain the face orientation detection model.
Further, the original model of the face orientation detection model includes: the face to be trained is branched towards and the confidence to be trained is branched; training an original model of the face orientation detection model by the plurality of training sample images and the face orientation angle information of the training subject in each training sample image comprises: training the face orientation branch to be trained through the plurality of training sample images and the face orientation angle information of the training object in each training sample image to obtain the face orientation branch of the face orientation detection model; training the confidence branch to be trained based on the plurality of training sample images, the face orientation angle information of the training object in each training sample image and the face orientation branch of the face orientation detection model to obtain the confidence branch of the face orientation detection model, and obtaining the face orientation detection model according to the face orientation branch of the face orientation detection model and the confidence branch of the face orientation detection model.
Further, the face-directing branch includes: the convolution layer structure comprises convolution layers with a first preset number and full connection layers with a second preset number, wherein the full connection layers are connected with the convolution layers with the first preset number; the confidence branch includes: the convolution layers with the first preset number and the full connection layers with the third preset number are connected with the convolution layers with the first preset number.
Further, the face orientation angle information includes: a face pitch angle of the target object, a face yaw angle of the target object, and a face roll angle of the target object.
Further, the target object includes any one of: human, animal.
In a second aspect, an embodiment of the present invention further provides an apparatus for evaluating facial image quality, including: an acquisition unit configured to acquire a face image to be evaluated of a target object; the detection unit is used for detecting the face orientation of the face image to be evaluated by adopting a face orientation detection model to obtain the confidence coefficient of the face orientation angle information of the target object, wherein the confidence coefficient is used for representing the accuracy degree of the face orientation angle information of the target object; and the setting unit is used for taking the confidence coefficient as the image quality of the facial image to be evaluated.
In a third aspect, an embodiment of the present invention further provides an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, where the processor implements the steps of the method according to any one of the above first aspects when executing the computer program.
In a fourth aspect, the present invention further provides a computer-readable medium having non-volatile program code executable by a processor, where the program code causes the processor to execute the steps of the method according to any one of the above first aspects.
In the embodiment of the invention, firstly, a facial image to be evaluated of a target object is obtained; performing face orientation detection on a face image to be evaluated by adopting a face orientation detection model to obtain a confidence coefficient of face orientation angle information of a target object, wherein the confidence coefficient is used for representing the accuracy degree of the face orientation angle information of the target object; and taking the confidence as the image quality of the facial image to be evaluated. It can be known from the above description that, in the embodiment of the present invention, the facial orientation detection model is used to evaluate the image quality of the facial image to be evaluated, the training of the facial orientation detection model is simple, only the facial orientation angle of the training object in the training sample image needs to be labeled, the labeling cost is low, in addition, the parameter of the facial orientation angle is more objective and global, when the facial orientation angle information output by the model is more accurate (i.e., the confidence is higher), it indicates that the more effective information the model acquires from the facial image to be evaluated is, i.e., the better the quality of the facial image to be evaluated is, i.e., the image quality of the facial image to be evaluated is represented more objectively by the confidence of the facial orientation angle information, the accuracy is good, the reference value is high, and the problems that the labeling cost is high, and the labeling cost, The labeled sample is not comprehensive and objective enough, and the finally determined face image quality is poor in accuracy.
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 schematic diagram of an electronic device according to an embodiment of the present invention;
fig. 2 is a flowchart of a method for evaluating the quality of a face image according to an embodiment of the present invention;
FIG. 3 is a flowchart of a method for training a face orientation detection model according to an embodiment of the present invention;
FIG. 4 is a schematic diagram of a method for evaluating the quality of a facial image according to an embodiment of the present invention;
fig. 5 is a schematic diagram of an apparatus for evaluating facial image quality according to an embodiment of the present invention.
Detailed Description
The technical solutions of the present invention will be described clearly and completely with reference to the following embodiments, and it should be understood 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.
Example 1:
first, an electronic apparatus 100 for implementing an embodiment of the present invention, which can be used to execute the evaluation method of the face image quality of the embodiments of the present invention, is described with reference to fig. 1.
As shown in FIG. 1, electronic device 100 includes one or more processors 102, one or more memories 104, an input device 106, an output device 108, and a camera 110, which are interconnected via a bus system 112 and/or other form of connection mechanism (not shown). It should be noted that the components and structure of the electronic device 100 shown in fig. 1 are exemplary only, and not limiting, and the electronic device may have other components and structures as desired.
The processor 102 may be implemented in at least one hardware form of a Digital Signal Processor (DSP), a Field-Programmable Gate Array (FPGA), a Programmable Logic Array (PLA), and an asic (application Specific Integrated circuit), and the processor 102 may be a Central Processing Unit (CPU) or other form of Processing Unit having data Processing capability and/or instruction execution capability, and may control other components in the electronic device 100 to perform desired functions.
The memory 104 may include one or more computer program products that may include various forms of computer-readable storage media, such as volatile memory and/or non-volatile memory. The volatile memory may include, for example, Random Access Memory (RAM), cache memory (cache), and/or the like. The non-volatile memory may include, for example, Read Only Memory (ROM), hard disk, flash memory, etc. On which one or more computer program instructions may be stored that may be executed by processor 102 to implement client-side functionality (implemented by the processor) and/or other desired functionality in embodiments of the invention described below. Various applications and various data, such as various data used and/or generated by the applications, may also be stored in the computer-readable storage medium.
The input device 106 may be a device used by a user to input instructions and may include one or more of a keyboard, a mouse, a microphone, a touch screen, and the like.
The output device 108 may output various information (e.g., images or sounds) to the outside (e.g., a user), and may include one or more of a display, a speaker, and the like.
The camera 110 is configured to capture a facial image to be evaluated, where the facial image to be evaluated captured by the camera is processed by the facial image quality evaluation method to obtain the image quality of the facial image to be evaluated, for example, the camera may capture a facial image (e.g., a photo, a video, etc.) to be evaluated desired by a user, and then process the facial image to be evaluated by the facial image quality evaluation method to obtain the image quality of the facial image to be evaluated, and the camera may further store the captured facial image to be evaluated in the memory 104 for use by other components.
Exemplarily, an electronic device for implementing the evaluation method of facial image quality according to an embodiment of the present invention may be implemented as an intelligent mobile terminal such as a smartphone, a tablet computer, or the like.
Example 2:
in accordance with an embodiment of the present invention, there is provided an embodiment of a method for evaluating facial image quality, it being noted that the steps illustrated in the flowchart of the drawings may be performed in a computer system such as a set of computer-executable instructions and that, although a logical order is illustrated in the flowchart, in some cases the steps illustrated or described may be performed in an order different than here.
Fig. 2 is a flowchart of an evaluation method of facial image quality according to an embodiment of the present invention, as shown in fig. 2, the method including the steps of:
step S202, acquiring a facial image to be evaluated of a target object;
in the embodiment of the present invention, the target object may be a human being, an animal, or another real object having face orientation information. When the target object is a person, the face image to be evaluated is the face image to be evaluated; and when the target object is an animal, the facial image to be evaluated is the facial image of the animal to be evaluated.
When the target object is a person, the facial image to be evaluated of the target object may be a plurality of facial images to be evaluated of the same person, may also be a plurality of facial images to be evaluated of different persons (each facial image to be evaluated includes a face of one person), may also be one facial image to be evaluated including a plurality of different faces, and may also be one facial image to be evaluated including only one face.
Step S204, a face orientation detection model is adopted to detect the face orientation of the face image to be evaluated, so that the confidence coefficient of the face orientation angle information of the target object is obtained, wherein the confidence coefficient is used for representing the accuracy degree of the face orientation angle information of the target object;
the face orientation detection model is obtained by training an original model of the face orientation detection model in advance by using a training sample image labeled with training target face orientation angle information. In the training sample image, the face orientation angle information of the artificially labeled training object is more objective and global and more adaptive to the characteristics of the model compared with the information such as the shielding degree and the shielding area of the artificial label. The inventor considers that: when the face orientation angle information output by the face orientation detection model is more accurate (i.e. the confidence coefficient is higher), it indicates that the model acquires more effective information from the face image to be evaluated, that is, the quality of the face image to be evaluated is better, so that the image quality of the face image to be evaluated can be objectively and accurately represented by the confidence coefficient of the face orientation angle information.
In an embodiment of the present invention, the face orientation angle information includes: a face pitch angle of the target object (corresponding to an angle at which the face of the target object is noded up and down), a face yaw angle of the target object (corresponding to an angle at which the face of the target object is shaken left and right), and a face roll angle of the target object (corresponding to an angle at which the face of the target object is turned left and right).
In step S206, the confidence is used as the image quality of the face image to be evaluated.
In the embodiment of the invention, firstly, a facial image to be evaluated of a target object is obtained; performing face orientation detection on a face image to be evaluated by adopting a face orientation detection model to obtain a confidence coefficient of face orientation angle information of a target object, wherein the confidence coefficient is used for representing the accuracy degree of the face orientation angle information of the target object; and taking the confidence as the image quality of the facial image to be evaluated. It can be known from the above description that, in the embodiment of the present invention, the facial orientation detection model is used to evaluate the image quality of the facial image to be evaluated, the training of the facial orientation detection model is simple, only the facial orientation angle of the training object in the training sample image needs to be labeled, the labeling cost is low, in addition, the parameter of the facial orientation angle is more objective and global, when the facial orientation angle information output by the model is more accurate (i.e., the confidence is higher), it indicates that the more effective information the model acquires from the facial image to be evaluated is, i.e., the better the quality of the facial image to be evaluated is, i.e., the image quality of the facial image to be evaluated is represented more objectively by the confidence of the facial orientation angle information, the accuracy is good, the reference value is high, and the problems that the labeling cost is high, and the labeling cost, The labeled sample is not comprehensive and objective enough, and the finally determined face image quality is poor in accuracy.
The above-mentioned contents briefly describe the method for evaluating the quality of a face image of the present invention, and a specific application thereof is described below.
In an optional embodiment of the present invention, after obtaining the image quality of the face image to be evaluated, the method further includes the following steps (1) and (2):
(1) acquiring a preset confidence level threshold;
(2) and screening the facial image to be evaluated based on a preset reliability threshold value and the image quality of the facial image to be evaluated to obtain a high-quality facial image to be evaluated, wherein the image quality of the high-quality facial image to be evaluated is greater than the preset reliability threshold value.
In an embodiment of the present invention, the face orientation detection model includes: a face orientation branch and a confidence branch; the face orientation branch is used for detecting face orientation angle information of a target object in the face image to be evaluated according to the face image to be evaluated; the confidence branch is used for detecting the confidence of the face orientation angle information of the target object in the facial image to be evaluated according to the facial image to be evaluated.
The training process of the face orientation detection model is described in detail below.
In an alternative embodiment of the present invention, referring to fig. 3, the training process of the face orientation detection model includes the following steps:
step S301, acquiring a plurality of training sample images containing faces of training objects and face orientation angle information of the training objects in each training sample image;
specifically, the face orientation angle information of a training object in a training sample image is labeled in a manual labeling mode, and the labeled face orientation angle information (including pitch, face pitch angle, yaw, roll, and face roll angle) is used as a supervision signal to train a regression model of the face orientation.
Step S302, training the original model of the face orientation detection model through the face orientation angle information of the training objects in the plurality of training sample images and each training sample image to obtain the face orientation detection model.
In an embodiment of the present invention, the original model of the face orientation detection model includes: the face orientation branch to be trained and the confidence degree branch to be trained, wherein the face orientation branch to be trained comprises: the convolution layers are connected with the convolution layers in the first preset number, and the full connection layers in the second preset number are connected with the convolution layers in the first preset number; the confidence branches to be trained include: the convolution layer of the first preset number, and the full link layer of the third preset number connected with the convolution layer of the first preset number. That is, the face orientation branch to be trained and the confidence branch to be trained share a basic convolutional network structure, and the basic convolutional network structure includes the first preset number of convolutional layers described above.
Accordingly, the face-directing branch comprises: the convolution layers are connected with the convolution layers in the first preset number, and the full connection layers in the second preset number are connected with the convolution layers in the first preset number; the confidence branch includes: the convolution layer of the first preset number, and the full link layer of the third preset number connected with the convolution layer of the first preset number.
The first predetermined number may be 5 layers, the second predetermined number may be 2 layers, and the third predetermined number may be 1 layer, and the number is not particularly limited in the embodiment of the present invention.
Specifically, training the original model of the face orientation detection model by using the face orientation angle information of the training target in the plurality of training sample images and each training sample image includes:
1) training the face orientation branch to be trained through the plurality of training sample images and the face orientation angle information of the training object in each training sample image to obtain the face orientation branch of the face orientation detection model;
in implementation, the face orientation angle information (i.e. pitch, yaw and roll) of the training object in the training sample images and the training sample images are input into a network of face orientation branches to be trained, predicted face orientation angle information is output, the predicted face orientation angle information is subtracted from the labeled face orientation angle information to obtain three face orientation angle losses (i.e. pitch _ loss, yaw _ loss and roll _ loss), the three face orientation angle losses are added to serve as a loss function of the face orientation branch network, the loss function is propagated in reverse, iteration is repeated until the network converges, and the face orientation branches of the face orientation detection model are obtained through training.
2) Training confidence degree branches to be trained based on the multiple training sample images, the face orientation angle information of the training object in each training sample image and the face orientation branches of the face orientation detection model to obtain confidence degree branches of the face orientation detection model, and obtaining the face orientation detection model according to the face orientation branches of the face orientation detection model and the confidence degree branches of the face orientation detection model.
After the training of the face orientation branch is completed, the confidence level branch is continuously trained, a plurality of training sample images and face orientation angle information of a training object in each training sample image are input into a first preset number of convolutional layers, a third preset number of full-connection layer outputs pitch uncertainty, yaw uncertainty and roll uncertainty connected with the face-oriented branch output module, the output pitch uncertainty, yaw uncertainty and roll uncertainty are weighted with the face-oriented branch output pitch _ loss, yaw _ loss and roll _ loss respectively, the network adaptively adjusts the magnitudes of the three uncertainties according to the image difficulty, thereby learning uncertainty of the face orientation angles, adding up the learned uncertainty of the three face orientation angles, the uncertainty of the face orientation angle information output by the model is obtained, and the sum of the uncertainties of the three face orientation angles subtracted by 1 is the confidence coefficient of the face orientation angle information.
The face orientation branch is trained firstly, and then the confidence branch is trained, so that the trained face orientation detection model has better performance.
The method for evaluating the quality of a face image according to the present invention will be described in its entirety with reference to fig. 4.
Referring to fig. 4, an original image is obtained first, a face orientation angle label is obtained by labeling a face orientation angle of the original image, a face orientation branch is trained through the face orientation angle label and the original image, and after training is completed, face orientation angle loss output by the original image, the face orientation angle label and the face orientation branch is continuously input to a confidence branch to train the confidence branch, so that a face orientation detection model is obtained.
When the face orientation detection method is applied, the face image to be evaluated of the target object is input to the face orientation detection model, and then the face image to be evaluated is screened based on the preset confidence level threshold value, so that the high-quality face image to be evaluated and the low-quality face image to be evaluated are obtained.
The face orientation detection model only needs a face orientation angle label when being trained, and does not need additional labels (such as blurring, shielding, non-human faces and the like); the image quality output by the model is not artificially defined any more, but the image quality considered by the model, when the downstream model is identified by screening the images, the images more beneficial to model identification can be screened, and the overall identification precision of the downstream model is improved.
Example 3:
the embodiment of the present invention further provides an evaluation apparatus for facial image quality, which is mainly used for executing the evaluation method for facial image quality provided by the above-mentioned content of the embodiment of the present invention, and the evaluation apparatus for facial image quality provided by the embodiment of the present invention is specifically described below.
Fig. 5 is a schematic diagram of an evaluation apparatus of a face image quality according to an embodiment of the present invention, as shown in fig. 5, the evaluation apparatus of a face image quality mainly includes: an acquisition unit 10, a detection unit 20 and a setting unit 30, wherein:
an acquisition unit configured to acquire a face image to be evaluated of a target object;
the face orientation detection unit is used for detecting the face orientation of the face image to be evaluated by adopting a face orientation detection model to obtain the confidence coefficient of the face orientation angle information of the target object, wherein the confidence coefficient is used for representing the accuracy degree of the face orientation angle information of the target object;
and the setting unit is used for taking the confidence coefficient as the image quality of the facial image to be evaluated.
In the embodiment of the invention, firstly, a facial image to be evaluated of a target object is obtained; performing face orientation detection on a face image to be evaluated by adopting a face orientation detection model to obtain a confidence coefficient of face orientation angle information of a target object, wherein the confidence coefficient is used for representing the accuracy degree of the face orientation angle information of the target object; and taking the confidence as the image quality of the facial image to be evaluated. It can be known from the above description that, in the embodiment of the present invention, the facial orientation detection model is used to evaluate the image quality of the facial image to be evaluated, the training of the facial orientation detection model is simple, only the facial orientation angle of the training object in the training sample image needs to be labeled, the labeling cost is low, in addition, the parameter of the facial orientation angle is more objective and global, when the facial orientation angle information output by the model is more accurate (i.e., the confidence is higher), it indicates that the more effective information the model acquires from the facial image to be evaluated is, i.e., the better the quality of the facial image to be evaluated is, i.e., the image quality of the facial image to be evaluated is represented more objectively by the confidence of the facial orientation angle information, the accuracy is good, the reference value is high, and the problems that the labeling cost is high, and the labeling cost, The labeled sample is not comprehensive and objective enough, and the finally determined face image quality is poor in accuracy.
Optionally, the apparatus is further configured to: acquiring a preset confidence level threshold; and screening the facial image to be evaluated based on a preset reliability threshold value and the image quality of the facial image to be evaluated to obtain a high-quality facial image to be evaluated, wherein the image quality of the high-quality facial image to be evaluated is greater than the preset reliability threshold value.
Optionally, the face orientation detection model comprises: a face orientation branch and a confidence branch; the face orientation branch is used for detecting face orientation angle information of a target object in the face image to be evaluated according to the face image to be evaluated; the confidence branch is used for detecting the confidence of the face orientation angle information of the target object in the facial image to be evaluated according to the facial image to be evaluated.
Optionally, the apparatus is further configured to: acquiring a plurality of training sample images containing faces of training objects and face orientation angle information of the training objects in each training sample image; and training the original model of the face orientation detection model through the face orientation angle information of the training objects in the training sample images and the training sample images to obtain the face orientation detection model.
Optionally, the original model of the face orientation detection model comprises: the face to be trained is branched towards and the confidence to be trained is branched; the apparatus is also configured to: training the face orientation branch to be trained through the plurality of training sample images and the face orientation angle information of the training object in each training sample image to obtain the face orientation branch of the face orientation detection model; training confidence degree branches to be trained based on the multiple training sample images, the face orientation angle information of the training object in each training sample image and the face orientation branches of the face orientation detection model to obtain confidence degree branches of the face orientation detection model, and obtaining the face orientation detection model according to the face orientation branches of the face orientation detection model and the confidence degree branches of the face orientation detection model.
Optionally, the face-facing branch comprises: the convolution layers are connected with the convolution layers in the first preset number, and the full connection layers in the second preset number are connected with the convolution layers in the first preset number; the confidence branch includes: the convolution layer of the first preset number, and the full link layer of the third preset number connected with the convolution layer of the first preset number.
Optionally, the face orientation angle information includes: a face pitch angle of the target object, a face yaw angle of the target object, and a face roll angle of the target object.
Optionally, the target object comprises any one of: human, animal.
The device provided by the embodiment of the present invention has the same implementation principle and technical effect as the method embodiments, and for the sake of brief description, reference may be made to the corresponding contents in the method embodiments without reference to the device embodiments.
In another embodiment of the present invention, a computer storage medium is also provided, on which a computer program is stored, which when executed by a computer performs the steps of the method of the above-described method embodiment.
In another embodiment of the present invention, a computer program is also provided, which may be stored on a storage medium in the cloud or in the local. When being executed by a computer or a processor, for performing the respective steps of the method of an embodiment of the invention and for implementing the respective modules in the management device of camera distribution according to an embodiment of the invention.
In addition, in the description of the embodiments of the present invention, unless otherwise explicitly specified or limited, the terms "mounted," "connected," and "connected" are to be construed broadly, e.g., as meaning either a fixed connection, a removable connection, or an integral connection; can be mechanically or electrically connected; they may be connected directly or indirectly through intervening media, or they may be interconnected between two elements. The specific meanings of the above terms in the present invention can be understood in specific cases to those skilled in the art.
In the description of the present invention, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc., indicate orientations or positional relationships based on the orientations or positional relationships shown in the drawings, and are only for convenience of description and simplicity of description, but do not indicate or imply that the device or element being referred to must have a particular orientation, be constructed and operated in a particular orientation, and thus, should not be construed as limiting the present invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance.
It is clear to those skilled in the art that, for convenience and brevity of description, the specific working processes of the above-described systems, apparatuses and units may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again.
In the several embodiments provided in the present application, it should be understood that the disclosed system, apparatus and method may be implemented in other ways. The above-described embodiments of the apparatus are merely illustrative, and for example, the division of the units is only one logical division, and there may be other divisions when actually implemented, and for example, a plurality of units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection of devices or units through some communication interfaces, and may be in an electrical, mechanical or other form.
The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.
In addition, each functional unit in each embodiment of the present invention may be integrated into one analysis unit, or each unit may exist alone physically, or two or more units are integrated into one unit.
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 non-volatile computer readable storage medium executable by the analyzer. 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, a server, or a network device) to execute 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-mentioned embodiments are only specific embodiments of the present invention, which are used for illustrating the technical solutions of the present invention and not for limiting the same, and the protection scope of the present invention is not limited thereto, although the present invention is described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: any person skilled in the art can modify or easily conceive the technical solutions described in the foregoing embodiments or equivalent substitutes for some technical features within the technical scope of the present disclosure; such modifications, changes or substitutions do not depart from the spirit and scope of the embodiments of the present invention, and they should be construed as being included therein. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims (11)
1. A method of evaluating the quality of a facial image, comprising:
acquiring a facial image to be evaluated of a target object;
performing face orientation detection on the face image to be evaluated by adopting a face orientation detection model to obtain a confidence coefficient of the face orientation angle information of the target object, wherein the confidence coefficient is used for representing the accuracy degree of the face orientation angle information of the target object;
and taking the confidence as the image quality of the facial image to be evaluated.
2. The method according to claim 1, wherein after obtaining the image quality of the facial image to be evaluated, the method further comprises:
acquiring a preset confidence level threshold;
and screening the facial image to be evaluated based on the preset confidence level threshold value and the image quality of the facial image to be evaluated to obtain a high-quality facial image to be evaluated, wherein the image quality of the high-quality facial image to be evaluated is greater than the preset confidence level threshold value.
3. The method of claim 1, wherein the face orientation detection model comprises: a face orientation branch and a confidence branch;
the face orientation branch is used for detecting face orientation angle information of a target object in the face image to be evaluated according to the face image to be evaluated;
the confidence branch is used for detecting the confidence of the face orientation angle information of the target object in the facial image to be evaluated according to the facial image to be evaluated.
4. The method of claim 1, further comprising:
acquiring a plurality of training sample images containing faces of training objects and face orientation angle information of the training objects in each training sample image;
and training an original model of the face orientation detection model according to the plurality of training sample images and the face orientation angle information of the training object in each training sample image to obtain the face orientation detection model.
5. The method of claim 4, wherein the original model of the face orientation detection model comprises: the face to be trained is branched towards and the confidence to be trained is branched;
training an original model of the face orientation detection model by the plurality of training sample images and the face orientation angle information of the training subject in each training sample image comprises:
training the face orientation branch to be trained through the plurality of training sample images and the face orientation angle information of the training object in each training sample image to obtain the face orientation branch of the face orientation detection model;
training the confidence branch to be trained based on the plurality of training sample images, the face orientation angle information of the training object in each training sample image and the face orientation branch of the face orientation detection model to obtain the confidence branch of the face orientation detection model, and obtaining the face orientation detection model according to the face orientation branch of the face orientation detection model and the confidence branch of the face orientation detection model.
6. The method of claim 3, wherein the face-directing bifurcation comprises: the convolution layer structure comprises convolution layers with a first preset number and full connection layers with a second preset number, wherein the full connection layers are connected with the convolution layers with the first preset number; the confidence branch includes: the convolution layers with the first preset number and the full connection layers with the third preset number are connected with the convolution layers with the first preset number.
7. The method according to any one of claims 1 to 6, wherein the face orientation angle information includes: a face pitch angle of the target object, a face yaw angle of the target object, and a face roll angle of the target object.
8. The method according to any one of claims 1 to 6, wherein the target object comprises any one of: human, animal.
9. An evaluation apparatus of a face image quality, characterized by comprising:
an acquisition unit configured to acquire a face image to be evaluated of a target object;
the detection unit is used for detecting the face orientation of the face image to be evaluated by adopting a face orientation detection model to obtain the confidence coefficient of the face orientation angle information of the target object, wherein the confidence coefficient is used for representing the accuracy degree of the face orientation angle information of the target object;
and the setting unit is used for taking the confidence coefficient as the image quality of the facial image to be evaluated.
10. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the steps of the method of any of the preceding claims 1 to 8 are implemented when the computer program is executed by the processor.
11. A computer-readable medium having non-volatile program code executable by a processor, characterized in that the program code causes the processor to perform the steps of the method of any of the preceding claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011037197.8A CN112307900A (en) | 2020-09-27 | 2020-09-27 | Method and device for evaluating facial image quality and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011037197.8A CN112307900A (en) | 2020-09-27 | 2020-09-27 | Method and device for evaluating facial image quality and electronic equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112307900A true CN112307900A (en) | 2021-02-02 |
Family
ID=74488769
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011037197.8A Pending CN112307900A (en) | 2020-09-27 | 2020-09-27 | Method and device for evaluating facial image quality and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112307900A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114299037A (en) * | 2021-12-30 | 2022-04-08 | 广州极飞科技股份有限公司 | Method and device for evaluating quality of object detection result, electronic equipment and computer readable storage medium |
| CN114842464A (en) * | 2022-05-13 | 2022-08-02 | 北京百度网讯科技有限公司 | Image direction recognition method, device, equipment, storage medium and program product |
| WO2024001365A1 (en) * | 2022-06-28 | 2024-01-04 | 魔门塔(苏州)科技有限公司 | Parameter measurement method and device |
-
2020
- 2020-09-27 CN CN202011037197.8A patent/CN112307900A/en active Pending
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114299037A (en) * | 2021-12-30 | 2022-04-08 | 广州极飞科技股份有限公司 | Method and device for evaluating quality of object detection result, electronic equipment and computer readable storage medium |
| CN114299037B (en) * | 2021-12-30 | 2023-09-01 | 广州极飞科技股份有限公司 | Quality evaluation method and device for object detection result, electronic equipment and computer readable storage medium |
| CN114842464A (en) * | 2022-05-13 | 2022-08-02 | 北京百度网讯科技有限公司 | Image direction recognition method, device, equipment, storage medium and program product |
| WO2024001365A1 (en) * | 2022-06-28 | 2024-01-04 | 魔门塔(苏州)科技有限公司 | Parameter measurement method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109344908B (en) | Method and apparatus for generating a model | |
| CN112307900A (en) | Method and device for evaluating facial image quality and electronic equipment | |
| CN108337505B (en) | Information acquisition method and device | |
| CN109086780B (en) | Method and device for detecting electrode plate burrs | |
| CN112183166A (en) | Method and device for determining training sample and electronic equipment | |
| CN111695421A (en) | Image recognition method and device and electronic equipment | |
| CN110956615A (en) | Image quality evaluation model training method and device, electronic equipment and storage medium | |
| CN111008561A (en) | Livestock quantity determination method, terminal and computer storage medium | |
| US11836898B2 (en) | Method and apparatus for generating image, and electronic device | |
| CN112465871A (en) | Method and system for evaluating accuracy of visual tracking algorithm | |
| CN112802076A (en) | Reflection image generation model and training method of reflection removal model | |
| CN112288664A (en) | High dynamic range image fusion method and device and electronic equipment | |
| CN110490058B (en) | Training method, device and system of pedestrian detection model and computer readable medium | |
| CN111325181A (en) | State monitoring method and device, electronic equipment and storage medium | |
| CN111031359A (en) | Video playing method and device, electronic equipment and computer readable storage medium | |
| CN108875901B (en) | Neural network training method and universal object detection method, device and system | |
| CN112580581A (en) | Target detection method and device and electronic equipment | |
| CN113158773A (en) | Training method and training device for living body detection model | |
| CN110717441B (en) | Video target detection method, device, equipment and medium | |
| CN109685015B (en) | Image processing method and device, electronic equipment and computer storage medium | |
| CN111860623A (en) | Method and system for counting tree number based on improved SSD neural network | |
| CN109871814B (en) | Age estimation method and device, electronic equipment and computer storage medium | |
| CN116343007A (en) | Target detection method, device, equipment and storage medium | |
| CN113298122A (en) | Target detection method and device and electronic equipment | |
| CN114170271A (en) | Multi-target tracking method with self-tracking consciousness, equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |