CN111210402A - Face image quality scoring method and device, computer equipment and storage medium - Google Patents
Face image quality scoring method and device, computer equipment and storage medium Download PDFInfo
- Publication number
- CN111210402A CN111210402A CN201911221143.4A CN201911221143A CN111210402A CN 111210402 A CN111210402 A CN 111210402A CN 201911221143 A CN201911221143 A CN 201911221143A CN 111210402 A CN111210402 A CN 111210402A
- Authority
- CN
- China
- Prior art keywords
- face image
- training
- scores
- decision tree
- tree model
- 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
- 238000013077 scoring method Methods 0.000 title claims abstract description 14
- 238000012549 training Methods 0.000 claims abstract description 108
- 238000003066 decision tree Methods 0.000 claims abstract description 74
- 238000011156 evaluation Methods 0.000 claims abstract description 71
- 238000000034 method Methods 0.000 claims abstract description 23
- 230000001815 facial effect Effects 0.000 claims description 7
- 238000002372 labelling Methods 0.000 claims description 6
- 238000004590 computer program Methods 0.000 claims description 5
- 238000010586 diagram Methods 0.000 description 6
- 238000010606 normalization Methods 0.000 description 5
- 238000012545 processing Methods 0.000 description 4
- 238000000605 extraction Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 238000013527 convolutional neural network Methods 0.000 description 2
- 238000012937 correction Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000013441 quality evaluation Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
-
- 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
- G06V40/165—Detection; Localisation; Normalisation using facial parts and geometric relationships
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30168—Image quality inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30196—Human being; Person
- G06T2207/30201—Face
Abstract
The invention discloses a face image quality scoring method, a face image quality scoring device, computer equipment and a storage medium. In the invention, a face image to be scored is obtained; evaluating the face image based on a plurality of evaluation dimensions respectively to obtain a plurality of evaluation scores; and inputting the plurality of evaluation scores into an image quality grading decision tree model obtained by pre-training so as to obtain the comprehensive grade of the human face image quality. The face image quality grading method comprehensively grades the face image quality based on a plurality of evaluation dimensions, so that the evaluation result of the face image quality is more accurate.
Description
Technical Field
The invention relates to the field of face recognition, in particular to a face image quality scoring method, a face image quality scoring device, computer equipment and a storage medium.
Background
For the face recognition technology, the face features acquired by a camera device need to be compared with a face feature template stored in a database, the face feature template is obtained by feature extraction according to a pre-acquired user face image, and the quality of the face image directly influences the accuracy of the face feature template and further influences the face recognition accuracy. In the prior art, the quality of a face image is evaluated and scored, and when the score is larger than a certain value, feature extraction is performed on the basis of the face image and the extracted face feature is used as a face feature template.
When the quality of a face image is evaluated and graded in the prior art, the quality of the face image is generally graded based on a single index, for example, the quality of the face image is graded based on one index of brightness, definition or contrast, and the like.
Disclosure of Invention
The invention mainly provides a face image quality evaluation method which can solve the problem that the existing face quality grading method based on single index can only evaluate the single index of a face image and cannot comprehensively evaluate the face image.
In order to solve the technical problems, the invention adopts a technical scheme that: the method for scoring the quality of the face image is applied to a server and comprises the following steps:
acquiring a face image to be scored;
evaluating the face image based on a plurality of evaluation dimensions respectively to obtain a plurality of evaluation scores;
and inputting a plurality of evaluation scores into an image quality scoring decision tree model obtained by pre-training so as to obtain a comprehensive score of the human face image quality.
Preferably, the plurality of evaluation dimensions comprises: the human face area accounts for the proportion, brightness, definition, contrast, human face posture, human face exaggerated expression and human face shielding of the human face image.
Preferably, the image quality scoring decision tree model is obtained by training through the following steps:
acquiring a plurality of sample face images to form a training set, and respectively labeling comprehensive quality scores to the plurality of sample face images;
dividing the plurality of sample face images into N training subsets according to the comprehensive quality scores marked by each sample face image, wherein the number of the comprehensive quality scores of the sample face images in each training subset is the same as that of the comprehensive quality scores of the plurality of sample face images in the training subset; wherein N is an integer greater than or equal to 1;
and training and obtaining the image quality scoring decision tree model based on the N training subsets.
Preferably, the step of labeling the multiple sample face images with the comprehensive quality scores respectively specifically includes:
for each sample face image, evaluating the sample face image based on a plurality of evaluation dimensions to obtain a plurality of dimension scores;
and marking the comprehensive quality score of each sample face image according to the multiple dimension scores of each sample face image.
Preferably, the step of dividing the plurality of sample face images into N training subsets according to the comprehensive quality score labeled for each sample face image specifically includes:
dividing the plurality of sample face images into a plurality of component groups according to the comprehensive quality scores marked by each sample face image, wherein the comprehensive quality scores of the plurality of sample face images in each component group are the same;
and extracting one or more sample face images from each component group to form a training subset, and extracting the sample face images into N training subsets.
Preferably, the step of obtaining the image quality score decision tree model based on the training of the N training subsets specifically includes:
training based on the first training subset to obtain a first decision tree model;
correcting the first decision tree model according to a second training subset to obtain a second decision tree model;
correcting the (i-1) th decision tree model according to the (i) th training subset to obtain an (i) th decision tree model; wherein i is more than or equal to 3 and less than or equal to N, and i is an integer;
and correcting the (N-1) th decision tree model according to the (N) th training subset to obtain an (N) th decision tree model, wherein the (N) th decision tree model is the image quality scoring decision tree model.
Preferably, the step of rectifying the first decision tree model according to the second training subset to obtain a second decision tree model specifically includes:
respectively obtaining image quality prediction scores of a plurality of sample face images in the second training subset according to the first decision tree model;
sequentially judging whether the predicted scores of the plurality of sample face images are the same as the scores of the marked comprehensive quality scores;
if not, correcting the marked comprehensive quality score according to the prediction score of the sample face image, and adding a plurality of corrected sample face images into the first training subset to obtain an updated first training subset;
training the updated first training subset to obtain the second decision tree model.
In order to solve the technical problem, the invention adopts another technical scheme that: there is provided a face image quality scoring apparatus, comprising:
the face image acquisition module is used for acquiring a face image to be scored;
the multi-dimensional evaluation module is used for evaluating the face image based on a plurality of evaluation dimensions respectively to obtain a plurality of evaluation scores;
and the comprehensive grading module is used for inputting the plurality of evaluation scores into an image quality grading decision tree model obtained by pre-training so as to obtain the comprehensive grading of the human face image quality.
In order to solve the technical problem, the invention adopts another technical scheme that: there is provided a computer device comprising a processor and a memory, the processor being coupled to the memory and the processor being operable to execute instructions to implement the above-described face image quality scoring method.
In order to solve the technical problem, the invention adopts another technical scheme that: there is provided a storage medium having stored thereon a computer program for execution by a processor to implement the above-described face image quality scoring method.
The invention has the beneficial effects that: different from the situation of the prior art, the method and the device obtain the face image to be scored; evaluating the face image based on a plurality of evaluation dimensions respectively to obtain a plurality of evaluation scores; and inputting the plurality of evaluation scores into an image quality grading decision tree model obtained by pre-training so as to obtain the comprehensive grade of the human face image quality. The face image quality grading method comprehensively grades the face image quality based on a plurality of evaluation dimensions, so that the evaluation result of the face image quality is more accurate.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings needed to be used in the description of the embodiments are briefly introduced below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without inventive efforts, wherein:
fig. 1 is a flowchart of a method for scoring a face image quality according to an embodiment of the present invention;
FIG. 2 is a training flow of an image quality scoring decision tree model according to another embodiment of the present invention
A flow chart;
FIG. 3 is a flow chart of a second decision tree model according to another embodiment of the present invention;
fig. 4 is a schematic structural diagram of a face image quality scoring apparatus according to another embodiment of the present invention;
FIG. 5 is a schematic structural diagram of a computer device according to another embodiment of the present invention;
fig. 6 is a schematic structural diagram of a storage medium according to another embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. 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 one
Fig. 1 is a flowchart of a face image quality scoring method according to an embodiment of the present invention, where the face image quality scoring method is applied to a server, and the face image quality scoring method includes step S100, step S200, and step S300.
Step S100: and acquiring a face image to be scored.
Step S200: and evaluating the face image based on the evaluation dimensions respectively to obtain a plurality of evaluation scores.
Specifically, a face frame and a face key point can be detected from a face image through a face detector, and then the face image is evaluated from a plurality of evaluation dimensions based on the face image, the face frame and the face key point. The plurality of evaluation dimensions includes: the proportion, brightness, definition, contrast, face posture, face exaggerated expression and face shielding of the face area in the face image. Evaluating the face image by calculating the proportion of the face region to the face image to obtain an evaluation score; evaluating the face image by calculating the brightness of the face image to obtain an evaluation score; by analogy, the face image is evaluated from multiple evaluation dimensions to obtain multiple evaluation scores.
Further specifically, the proportion of the face area to the face image, the brightness, the definition, the contrast and the face posture of the face image are calculated by means of the prior art, the exaggerated expression and the face shielding of the face are recognized by a convolutional neural network, normalization processing is carried out on each calculation result and each recognition result, the normalization is carried out to a value range of 0-10 points, and the value after the normalization is an evaluation value.
Step S300: and inputting the plurality of evaluation scores into an image quality grading decision tree model obtained by pre-training so as to obtain the comprehensive grade of the human face image quality.
Specifically, an image quality scoring decision tree model is trained in advance, and a plurality of evaluation scores obtained from evaluation of a plurality of evaluation dimensions are input into the scoring decision tree model to obtain a comprehensive score of the face image quality. The comprehensive score is obtained based on evaluation of 7 evaluation dimensions of proportion, brightness, definition, contrast, face posture, face exaggerated expression and face shielding of a face region in a face image, and compared with the score based on a single evaluation dimension, the comprehensive score is more accurate in evaluation of the quality of the face image.
In the embodiment of the invention, a face image to be scored is obtained; evaluating the face image based on a plurality of evaluation dimensions respectively to obtain a plurality of evaluation scores; and inputting the plurality of evaluation scores into an image quality grading decision tree model obtained by pre-training so as to obtain the comprehensive grade of the human face image quality. The face image quality grading method comprehensively grades the face image quality based on a plurality of evaluation dimensions, so that the evaluation result of the face image quality is more accurate.
Example two
Fig. 2 is a flowchart illustrating a training process of an image quality scoring decision tree model according to another embodiment of the present invention, where the training process of the image quality scoring decision tree model includes the following steps S400, S500, and S600.
Step S400: and acquiring a plurality of sample face images to form a training set, and respectively labeling the comprehensive quality scores of the plurality of sample face images.
Specifically, a plurality of sample face images are obtained, and the plurality of sample face images form training data in a training set.
Specifically, the step of respectively labeling the comprehensive quality scores of the face images of the multiple samples specifically comprises: for each sample face image, evaluating the sample face image based on a plurality of evaluation dimensions to obtain a plurality of dimension scores; and marking the comprehensive quality score of each sample face image according to the multiple dimension scores of each sample face image. Wherein the plurality of evaluation dimensions include: the proportion, brightness, definition, contrast, face posture, face exaggerated expression and face shielding of the face area in the face image. For each sample face image, calculating the proportion of a face region in the face image, the brightness, the definition, the contrast and the face posture of the face image by the prior art means, identifying the exaggerated expression and the face shielding of the face by a convolutional neural network, and carrying out normalization processing on each calculation result and each identification result to a value range of 0-10 minutes, wherein the value after normalization is dimension score. And for each sample face image, marking the comprehensive quality score of the sample face image according to a plurality of dimension scores obtained by a plurality of evaluation dimensions, wherein the comprehensive quality score can be set to be in a score range of 1-10 points.
Step S500: dividing the multiple sample face images into N training subsets according to the comprehensive quality scores marked by each sample face image, wherein the number of the scores of the comprehensive quality scores of the sample face images in each training subset is the same as that of the scores of the comprehensive quality scores of the multiple sample face images in the training subset; wherein N is an integer greater than or equal to 1.
Specifically, the number of training subsets can be freely set, for example, N may be set to 5. If the scores of the comprehensive quality scores of all the sample face images in the training set include 1 score, 2 scores, 3 scores, 4 scores and 5 scores, and the number of the scores is 5, the scores of the comprehensive quality scores of the plurality of sample face images in each training subset also include 1 score, 2 scores, 3 scores, 4 scores and 5 scores.
More specifically, the step of dividing the plurality of sample face images into N training subsets according to the comprehensive quality score labeled on each sample face image specifically includes: dividing the plurality of sample face images into a plurality of component groups according to the comprehensive quality scores marked by each sample face image, wherein the comprehensive quality scores of the plurality of sample face images in each component group are the same; and extracting one or more sample face images from each component group to form a training subset, and extracting the plurality of sample face images into N training subsets. For example: if the scores of the comprehensive quality scores of the sample face images in the training set comprise 1 score, 2 scores, 3 scores, 4 scores and 5 scores, dividing a plurality of sample face images in the training set into 5 score groups, wherein the 5 score groups comprise 1 score group, 2 score group, 3 score group, 4 score group and 5 score group, the comprehensive quality scores of all sample face images in the 1 score group are 1 score, the comprehensive quality scores of all sample face images in the 2 score group are 2 scores, and the rest can be analogized to other score groups. Then, one or more sample facial images are extracted from each score group to form a training subset, for example, one sample facial image can be extracted from 1 score group, 2 score groups, 3 score groups, 4 score groups and 5 score groups to form a training subset, and the training subset comprises 5 sample facial images with 5 scores. The training set is extracted and divided into 5(N is 5) training subsets in the extraction mode, and the sample face images in each training subset are different.
Step S600: and training and obtaining an image quality scoring decision tree model based on the N training subsets.
Specifically, a first decision tree model is obtained based on a first training subset training; correcting the first decision tree model according to the second training subset to obtain a second decision tree model; correcting the (i-1) th decision tree model according to the (i) th training subset to obtain an (i) th decision tree model; wherein i is more than or equal to 3 and less than or equal to N, and i is an integer; and correcting the (N-1) th decision tree model according to the (N) th training subset to obtain an (N) th decision tree model, wherein the (N) th decision tree model is an image quality scoring decision tree model.
More specifically, as shown in fig. 3, the step of correcting the first decision tree model according to the second training subset to obtain the second decision tree model specifically includes step S601, step S602, step S603, and step S604:
and S601, respectively obtaining image quality prediction scores of a plurality of sample face images in the second training subset according to the first decision tree model.
Specifically, for each sample face image in the second training subset, the multiple dimensionality scores of the sample face image are input into the first decision tree model, and the image quality prediction score of the sample face image is obtained.
And step S602, sequentially judging whether the prediction scores of the face images of the multiple samples are the same as the scores of the marked comprehensive quality scores.
Specifically, for each sample face image in the second training subset, whether the prediction score of the sample face image is the same as the score of the labeled comprehensive quality score is judged, and the score judgment is performed on each sample face image in the second training subset.
And step S603, if not, correcting the marked comprehensive quality score according to the prediction score of the sample face image, and adding the corrected multiple sample face images into the first training subset to obtain an updated first training subset.
Specifically, for each sample face image in the second training subset, if the predicted score of the sample face image is different from the score of the comprehensive quality score marked by the sample face image, the marked comprehensive quality score is corrected, all sample face images with different scores in the second training subset are corrected, and all sample face images in the corrected second training subset are added into the first training subset to obtain an updated first training subset. Wherein, the scheme for correcting the marked comprehensive quality score can be as follows: and taking the average value of the prediction score of the sample face image and the marked comprehensive quality score, taking the average value as the marked comprehensive quality score after correction, if the average value is a non-integer, rounding down, and if the average value is 6.5, taking 6 as the marked comprehensive quality score after correction.
Step S604, the updated first training subset is trained to obtain a second decision tree model.
Specifically, the sample face images in the corrected second training subset are added to the first training subset to obtain an updated first training subset, and the updated first training subset is trained to obtain a second decision tree model.
In this embodiment, the process of correcting the i-1 st decision tree model according to the i (i is greater than or equal to 3 and less than or equal to N) th training subset to obtain the i-th decision tree model is the same as the process of correcting the first decision tree model according to the second training subset to obtain the second decision tree model, and the process of correcting the i-1 st decision tree model according to the i (i is greater than or equal to 3 and less than or equal to N) th training subset to obtain the i-th decision tree model is not described herein one by one.
In the embodiment of the invention, the image quality grading decision tree model is obtained by training according to the data of a plurality of sample face images, the accuracy is high, and meanwhile, the image quality grading decision tree model is obtained by training 7 parameters, namely the proportion of the face region of the sample face images in the face images, the brightness, the definition, the contrast, the face posture, the exaggerated expression of the face and the face shielding, so that the comprehensive grade of the face images obtained by evaluating based on the image quality grading decision tree model is also obtained based on the 7 parameters, the evaluation dimensionality is large, and the evaluation result is more accurate.
EXAMPLE III
Fig. 4 is a schematic structural diagram of a face image quality scoring device according to another embodiment of the present invention, where the face image quality scoring device includes a face image obtaining module 100, a multi-dimensional evaluation module 200, and a comprehensive scoring module 300.
The face image acquisition module 100 is used for acquiring a face image to be scored;
the multi-dimensional evaluation module 200 is configured to evaluate the face image based on a plurality of evaluation dimensions, respectively, to obtain a plurality of evaluation scores;
the comprehensive scoring module 300 is configured to input a plurality of evaluation scores into a pre-trained image quality scoring decision tree model to obtain a comprehensive score of the face image quality.
Preferably, the plurality of evaluation dimensions comprises: the proportion, brightness, definition, contrast, face posture, face exaggerated expression and face shielding of the face area in the face image.
The specific implementation of the face image quality scoring device provided by the embodiment of the present invention is the same as the specific implementation of the face image quality scoring method, and the specific implementation of the face image quality scoring device may refer to the description of the first embodiment and the second embodiment, which is not described herein again.
In the embodiment of the invention, a face image to be scored is obtained; evaluating the face image based on a plurality of evaluation dimensions respectively to obtain a plurality of evaluation scores; and inputting the plurality of evaluation scores into an image quality grading decision tree model obtained by pre-training so as to obtain the comprehensive grade of the human face image quality. The face image quality grading method comprehensively grades the face image quality based on a plurality of evaluation dimensions, so that the evaluation result of the face image quality is more accurate.
Example four
Fig. 5 is a schematic structural diagram of a computer device according to another embodiment of the present invention, the computer device includes a processor 400 and a memory 500, the processor 400 is coupled to the memory 500, and the processor 400 executes instructions to implement the facial image quality scoring method in any of the above embodiments when operating.
The processor 400 may also be referred to as a CPU (Central Processing Unit). Processor 400 may be an integrated circuit chip having signal processing capabilities. Processor 400 may also be a general purpose processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), an off-the-shelf programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components. A general purpose processor may be a microprocessor or the processor may be any conventional processor, but is not limited thereto.
EXAMPLE five
Referring to fig. 6, fig. 6 is a schematic diagram of a storage medium according to another embodiment of the present invention, in which a computer program 600 is stored, and the computer program 600 can be executed by the processor 400 to implement the facial image quality scoring method in any of the above embodiments.
Alternatively, the readable storage medium may be various media that can store program codes, such as a usb disk, a removable hard disk, a Read-only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, or may be a terminal device such as a computer, a server, a mobile phone, or a tablet.
In the invention, a face image to be scored is obtained; evaluating the face image based on a plurality of evaluation dimensions respectively to obtain a plurality of evaluation scores; and inputting the plurality of evaluation scores into an image quality grading decision tree model obtained by pre-training so as to obtain the comprehensive grade of the human face image quality. The face image quality grading method comprehensively grades the face image quality based on a plurality of evaluation dimensions, so that the evaluation result of the face image quality is more accurate.
The above description is only an embodiment of the present invention, and not intended to limit the scope of the present invention, and all modifications of equivalent structures and equivalent processes performed by the present specification and drawings, or directly or indirectly applied to other related technical fields, are included in the scope of the present invention.
Claims (10)
1. The face image quality scoring method is applied to a server, and is characterized by comprising the following steps:
acquiring a face image to be scored;
evaluating the face image based on a plurality of evaluation dimensions respectively to obtain a plurality of evaluation scores;
and inputting a plurality of evaluation scores into an image quality scoring decision tree model obtained by pre-training so as to obtain a comprehensive score of the human face image quality.
2. The method of claim 1, wherein the plurality of evaluation dimensions comprise: the human face area accounts for the proportion, brightness, definition, contrast, human face posture, human face exaggerated expression and human face shielding of the human face image.
3. The method for scoring the facial image quality as claimed in claim 2, wherein the image quality scoring decision tree model is obtained by training through the following steps:
acquiring a plurality of sample face images to form a training set, and respectively labeling comprehensive quality scores to the plurality of sample face images;
dividing the plurality of sample face images into N training subsets according to the comprehensive quality scores marked by each sample face image, wherein the number of the comprehensive quality scores of the sample face images in each training subset is the same as that of the comprehensive quality scores of the plurality of sample face images in the training subset; wherein N is an integer greater than or equal to 1;
and training and obtaining the image quality scoring decision tree model based on the N training subsets.
4. The method for scoring the quality of the face image according to claim 3, wherein the step of labeling the comprehensive quality scores of the plurality of sample face images respectively specifically comprises the steps of:
for each sample face image, evaluating the sample face image based on a plurality of evaluation dimensions to obtain a plurality of dimension scores;
and marking the comprehensive quality score of each sample face image according to the multiple dimension scores of each sample face image.
5. The method for scoring the quality of the face images according to claim 4, wherein the step of dividing the plurality of sample face images into N training subsets according to the comprehensive quality score labeled on each sample face image specifically comprises:
dividing the plurality of sample face images into a plurality of component groups according to the comprehensive quality scores marked by each sample face image, wherein the comprehensive quality scores of the plurality of sample face images in each component group are the same;
and extracting one or more sample face images from each component group to form a training subset, and extracting the sample face images into N training subsets.
6. The method for scoring the facial image quality as recited in claim 4, wherein the step of obtaining the image quality scoring decision tree model based on the training of the N training subsets specifically comprises:
training based on the first training subset to obtain a first decision tree model;
correcting the first decision tree model according to a second training subset to obtain a second decision tree model;
correcting the (i-1) th decision tree model according to the (i) th training subset to obtain an (i) th decision tree model; wherein i is more than or equal to 3 and less than or equal to N, and i is an integer;
and correcting the (N-1) th decision tree model according to the (N) th training subset to obtain an (N) th decision tree model, wherein the (N) th decision tree model is the image quality scoring decision tree model.
7. The method for scoring the quality of a human face image according to claim 6, wherein the step of correcting the first decision tree model according to the second training subset to obtain the second decision tree model specifically comprises:
respectively obtaining image quality prediction scores of a plurality of sample face images in the second training subset according to the first decision tree model;
sequentially judging whether the predicted scores of the plurality of sample face images are the same as the scores of the marked comprehensive quality scores;
if not, correcting the marked comprehensive quality score according to the prediction score of the sample face image, and adding a plurality of corrected sample face images into the first training subset to obtain an updated first training subset;
training the updated first training subset to obtain the second decision tree model.
8. Face image quality scoring device, characterized in that face image quality scoring device includes:
the face image acquisition module is used for acquiring a face image to be scored;
the multi-dimensional evaluation module is used for evaluating the face image based on a plurality of evaluation dimensions respectively to obtain a plurality of evaluation scores;
and the comprehensive grading module is used for inputting the plurality of evaluation scores into an image quality grading decision tree model obtained by pre-training so as to obtain the comprehensive grading of the human face image quality.
9. A computer device, characterized in that the computer device comprises a processor and a memory, the processor is coupled with the memory, and the processor executes instructions to realize the human face image quality scoring method according to any one of claims 1-7 when in work.
10. A storage medium having stored thereon a computer program, wherein the computer program is executed by a processor to implement the method for scoring a human face image quality according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911221143.4A CN111210402A (en) | 2019-12-03 | 2019-12-03 | Face image quality scoring method and device, computer equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911221143.4A CN111210402A (en) | 2019-12-03 | 2019-12-03 | Face image quality scoring method and device, computer equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111210402A true CN111210402A (en) | 2020-05-29 |
Family
ID=70786222
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911221143.4A Pending CN111210402A (en) | 2019-12-03 | 2019-12-03 | Face image quality scoring method and device, computer equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111210402A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112132074A (en) * | 2020-09-28 | 2020-12-25 | 平安养老保险股份有限公司 | Face image verification method and device, computer equipment and storage medium |
| CN112329638A (en) * | 2020-11-06 | 2021-02-05 | 上海优扬新媒信息技术有限公司 | Image scoring method, device and system |
| CN112507985A (en) * | 2021-02-03 | 2021-03-16 | 成都新希望金融信息有限公司 | Face image screening method and device, electronic equipment and storage medium |
| CN112950579A (en) * | 2021-02-26 | 2021-06-11 | 北京金山云网络技术有限公司 | Image quality evaluation method and device and electronic equipment |
| CN113536991A (en) * | 2021-06-29 | 2021-10-22 | 北京百度网讯科技有限公司 | Training set generation method, human face image processing method, device and electronic equipment |
| CN113936320A (en) * | 2021-10-21 | 2022-01-14 | 北京的卢深视科技有限公司 | Face image quality evaluation method, electronic device and storage medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108288027A (en) * | 2017-12-28 | 2018-07-17 | 新智数字科技有限公司 | A kind of detection method of picture quality, device and equipment |
| CN109146856A (en) * | 2018-08-02 | 2019-01-04 | 深圳市华付信息技术有限公司 | Picture quality assessment method, device, computer equipment and storage medium |
| CN109285149A (en) * | 2018-09-04 | 2019-01-29 | 杭州比智科技有限公司 | Appraisal procedure, device and the calculating equipment of quality of human face image |
| CN110046652A (en) * | 2019-03-18 | 2019-07-23 | 深圳神目信息技术有限公司 | Face method for evaluating quality, device, terminal and readable medium |
| CN110458792A (en) * | 2018-05-04 | 2019-11-15 | 北京眼神科技有限公司 | Method and device for evaluating quality of face image |
-
2019
- 2019-12-03 CN CN201911221143.4A patent/CN111210402A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108288027A (en) * | 2017-12-28 | 2018-07-17 | 新智数字科技有限公司 | A kind of detection method of picture quality, device and equipment |
| CN110458792A (en) * | 2018-05-04 | 2019-11-15 | 北京眼神科技有限公司 | Method and device for evaluating quality of face image |
| CN109146856A (en) * | 2018-08-02 | 2019-01-04 | 深圳市华付信息技术有限公司 | Picture quality assessment method, device, computer equipment and storage medium |
| CN109285149A (en) * | 2018-09-04 | 2019-01-29 | 杭州比智科技有限公司 | Appraisal procedure, device and the calculating equipment of quality of human face image |
| CN110046652A (en) * | 2019-03-18 | 2019-07-23 | 深圳神目信息技术有限公司 | Face method for evaluating quality, device, terminal and readable medium |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112132074A (en) * | 2020-09-28 | 2020-12-25 | 平安养老保险股份有限公司 | Face image verification method and device, computer equipment and storage medium |
| CN112329638A (en) * | 2020-11-06 | 2021-02-05 | 上海优扬新媒信息技术有限公司 | Image scoring method, device and system |
| CN112507985A (en) * | 2021-02-03 | 2021-03-16 | 成都新希望金融信息有限公司 | Face image screening method and device, electronic equipment and storage medium |
| CN112950579A (en) * | 2021-02-26 | 2021-06-11 | 北京金山云网络技术有限公司 | Image quality evaluation method and device and electronic equipment |
| CN113536991A (en) * | 2021-06-29 | 2021-10-22 | 北京百度网讯科技有限公司 | Training set generation method, human face image processing method, device and electronic equipment |
| CN113536991B (en) * | 2021-06-29 | 2023-06-23 | 北京百度网讯科技有限公司 | Training set generation method, face image processing method, device and electronic equipment |
| CN113936320A (en) * | 2021-10-21 | 2022-01-14 | 北京的卢深视科技有限公司 | Face image quality evaluation method, electronic device and storage medium |
| CN113936320B (en) * | 2021-10-21 | 2022-03-25 | 北京的卢深视科技有限公司 | Face image quality evaluation method, electronic device and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107944020B (en) | Face image searching method and device, computer device and storage medium | |
| CN111210402A (en) | Face image quality scoring method and device, computer equipment and storage medium | |
| CN111046133A (en) | Question-answering method, question-answering equipment, storage medium and device based on atlas knowledge base | |
| CN109919252B (en) | Method for generating classifier by using few labeled images | |
| CN108038208B (en) | Training method and device of context information recognition model and storage medium | |
| CN111199186A (en) | Image quality scoring model training method, device, equipment and storage medium | |
| CN110807314A (en) | Text emotion analysis model training method, device and equipment and readable storage medium | |
| CN110503143B (en) | Threshold selection method, device, storage medium and device based on intention recognition | |
| CN108959474B (en) | Entity relation extraction method | |
| CN110263326B (en) | User behavior prediction method, prediction device, storage medium and terminal equipment | |
| JP2022512065A (en) | Image classification model training method, image processing method and equipment | |
| US11062120B2 (en) | High speed reference point independent database filtering for fingerprint identification | |
| CN110717554A (en) | Image recognition method, electronic device, and storage medium | |
| CN104142912A (en) | Accurate corpus category marking method and device | |
| CN106372624A (en) | Human face recognition method and human face recognition system | |
| CN111784665B (en) | OCT image quality evaluation method, system and device based on Fourier transform | |
| CN112307820A (en) | Text recognition method, device, equipment and computer readable medium | |
| CN109213886B (en) | Image retrieval method and system based on image segmentation and fuzzy pattern recognition | |
| CN112036304A (en) | Medical bill layout identification method and device and computer equipment | |
| CN113192028B (en) | Quality evaluation method and device for face image, electronic equipment and storage medium | |
| CN111783812A (en) | Method and device for identifying forbidden images and computer readable storage medium | |
| CN108229565B (en) | Cognition-based image understanding method | |
| CN114943674A (en) | Defect detection method, electronic device and storage medium | |
| CN109389173B (en) | M-CNN-based test paper score automatic statistical analysis method and device | |
| CN112070793A (en) | Target extraction method and device |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20200529 |
|
| WD01 | Invention patent application deemed withdrawn after publication |