CN105357411B - The method and device of detection image quality - Google Patents
The method and device of detection image quality Download PDFInfo
- Publication number
- CN105357411B CN105357411B CN201510719412.5A CN201510719412A CN105357411B CN 105357411 B CN105357411 B CN 105357411B CN 201510719412 A CN201510719412 A CN 201510719412A CN 105357411 B CN105357411 B CN 105357411B
- Authority
- CN
- China
- Prior art keywords
- images
- quality
- feature vector
- tab
- image
- 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.)
- Active
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N1/00—Scanning, transmission or reproduction of documents or the like, e.g. facsimile transmission; Details thereof
- H04N1/00002—Diagnosis, testing or measuring; Detecting, analysing or monitoring not otherwise provided for
- H04N1/00005—Diagnosis, testing or measuring; Detecting, analysing or monitoring not otherwise provided for relating to image data
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- General Health & Medical Sciences (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Image Analysis (AREA)
Abstract
The disclosure is directed to a kind of method and devices of detection image quality, belong to technical field of image processing.This method includes:Multiple pending modeled images are obtained, there are one quality tabs for every modeled images tool;First pretreatment is carried out to multiple modeled images, obtains multiple first images;Fourier transformation is carried out to multiple first images, obtains multiple second images;Second pretreatment is carried out to multiple second images, obtains multiple third images;Multiple feature vectors are extracted from multiple third images;According to multiple feature vectors and corresponding quality tab, quality testing model is built;Based on quality testing model, objective image quality is detected.The disclosure is by building quality testing model, and the quality of target image is detected based on constructed quality testing model, it is participated in due to being not necessarily to user, thus avoids user cognition standard to the interference caused by testing result so that testing result is more objective, accurate.
Description
Technical field
This disclosure relates to technical field of image processing more particularly to a kind of method and device of detection image quality.
Background technology
In technical field of image processing, picture quality generally comprises brightness, contrast, resolution ratio of image etc., image matter
Amount testing result can be used for calibrating image imaging system, optimization algorithm, preferred parameter etc., is widely used in compression of images, passes
During defeated, storage etc..The naked eyes for depending on user at present are detected the quality of image.However it is limited to user's
Cognition standard, when different user carries out quality testing to same image, obtained quality measurements may be different, thus
There is an urgent need for a kind of more objective picture quality detection methods.
Invention content
The disclosure provides a kind of method and device of detection image quality.
According to the first aspect of the embodiments of the present disclosure, a kind of method of detection image quality is provided, the method includes:
Multiple pending modeled images are obtained, there are one quality tab, the quality tab packets for every modeled images tool
Include clear label and fuzzy label;
First pretreatment is carried out to multiple described modeled images, obtains multiple first images;
Fourier transformation is carried out to multiple described first images, obtains multiple second images;
Second pretreatment is carried out to multiple described second images, obtains multiple third images;
Multiple feature vectors are extracted from multiple described third images;
According to the multiple feature vector and corresponding quality tab, quality testing model is built;
Based on the quality testing model, objective image quality is detected.
With reference to first aspect, described that multiple described modelings are schemed in the first possible realization method of first aspect
It is pre-processed as carrying out first, obtains multiple first images, including:
By the size reduction of every modeled images to pre-set dimension, multiple first images are obtained.
With reference to first aspect, in second of possible realization method of first aspect, multiple described second images carry out
Second pretreatment, obtains multiple third images, including:
Multiple described second images are normalized and thresholding is handled, obtain multiple third figures with identical size
Picture.
With reference to first aspect, described according to the multiple feature in the third possible realization method of first aspect
Vectorial and corresponding quality tab builds quality testing model, including:
Using each feature vector as independent variable input, using the corresponding quality tab of each feature vector as dependent variable
During output, the correspondence between feature vector and quality tab is obtained;
It is fitted by the correspondence between described eigenvector and quality tab, obtains quality testing model.
With reference to first aspect, described to be based on the quality testing in the 4th kind of possible realization method of first aspect
Model is detected objective image quality, including:
Obtain the feature vector of the target image;
The feature vector of the target image is input in the quality testing model;
Export the corresponding quality tab of feature vector of the target image.
According to the second aspect of the embodiment of the present disclosure, a kind of device of detection image quality is provided, described device includes:
Acquisition module, for obtaining multiple pending modeled images, there are one quality tab, institutes for every modeled images tool
It includes clear label and fuzzy label to state quality tab;
First processing module obtains multiple first images for carrying out the first pretreatment to multiple described modeled images;
Conversion module obtains multiple second images for carrying out Fourier transformation to multiple described first images;
Second processing module obtains multiple third images for carrying out the second pretreatment to multiple described second images;
Extraction module, for extracting multiple feature vectors from multiple described third images;
Module is built, for according to the multiple feature vector and corresponding quality tab, building quality testing model;
Detection module is detected objective image quality for being based on the quality testing model.
In conjunction with second aspect, in the first possible realization method of second aspect, the first processing module is used for
By the size reduction of every modeled images to pre-set dimension, multiple first images are obtained.
In conjunction with second aspect, in second of possible realization method of second aspect, the Second processing module is used for
Multiple described second images are normalized and thresholding is handled, obtain multiple third images with identical size.
In conjunction with second aspect, in the third possible realization method of second aspect, the structure module, for inciting somebody to action
The process that each feature vector inputs as independent variable, exports the corresponding quality tab of each feature vector as dependent variable
In, obtain the correspondence between feature vector and quality tab;Pass through pair between described eigenvector and quality tab
It should be related to and be fitted, obtain quality testing model.
In conjunction with second aspect, in the 4th kind of possible realization method of second aspect, the detection module, for obtaining
The feature vector of the target image;The feature vector of the target image is input in the quality testing model;Output
The corresponding quality tab of feature vector of the target image.
According to the third aspect of the embodiment of the present disclosure, a kind of device of detection image quality is provided, including:
Processor;
Memory for storing the executable instruction of processor;
Wherein, the processor is configured as:
Multiple pending modeled images are obtained, there are one quality tab, the quality tab packets for every modeled images tool
Include clear label and fuzzy label;
First pretreatment is carried out to multiple described modeled images, obtains multiple first images;
Fourier transformation is carried out to multiple described first images, obtains multiple second images;
Second pretreatment is carried out to multiple described second images, obtains multiple third images;
Multiple feature vectors are extracted from multiple described third images;
According to the multiple feature vector and corresponding quality tab, quality testing model is built;
Based on the quality testing model, objective image quality is detected.
The technical scheme provided by this disclosed embodiment can include the following benefits:
By building quality testing model, and the quality of target image is examined based on constructed quality testing model
It surveys, is participated in due to being not necessarily to user, thus avoid cognition standard to the interference caused by testing result so that testing result is more
It is objective, accurate.
It should be understood that above general description and following detailed description is only exemplary and explanatory, not
The disclosure can be limited.
Description of the drawings
The drawings herein are incorporated into the specification and forms part of this specification, and shows the implementation for meeting the disclosure
Example, and together with specification for explaining the principles of this disclosure.
Fig. 1 is a kind of flow chart of the method for detection image quality shown according to an exemplary embodiment.
Fig. 2 is a kind of flow chart of the method for detection image quality shown according to an exemplary embodiment.
Fig. 3 is a kind of structural schematic diagram of the device of detection image quality shown according to an exemplary embodiment.
Fig. 4 is a kind of block diagram of device 400 for detection image quality shown according to an exemplary embodiment.
Specific implementation mode
Example embodiments are described in detail here, and the example is illustrated in the accompanying drawings.Following description is related to
When attached drawing, unless otherwise indicated, the same numbers in different drawings indicate the same or similar elements.Following exemplary embodiment
Described in embodiment do not represent all implementations consistent with this disclosure.On the contrary, they be only with it is such as appended
The example of the consistent device and method of some aspects be described in detail in claims, the disclosure.
Fig. 1 is a kind of flow chart of the method for detection image quality shown according to an exemplary embodiment, such as Fig. 1 institutes
Show, the method for detection image quality is in server, including the following steps.
In step S101, multiple pending modeled images are obtained, there are one quality tabs for every modeled images tool, should
Quality tab includes clear label and fuzzy label.
In step s 102, the first pretreatment is carried out to multiple modeled images, obtains multiple first images.
In step s 103, Fourier transformation is carried out to multiple first images, obtains multiple second images.
In step S104, the second pretreatment is carried out to multiple second images, obtains multiple third images.
In step S105, multiple feature vectors are extracted from multiple third images.
In step s 106, according to multiple feature vectors and corresponding quality tab, quality testing model is built.
In step s 107, it is based on quality testing model, objective image quality is detected.
The method that the embodiment of the present disclosure provides, by building quality testing model, and based on constructed quality testing mould
Type is detected the quality of target image, is participated in due to being not necessarily to user, thus avoids cognition standard and made to testing result
At interference so that testing result is more objective, accurate.
In another embodiment of the disclosure, the first pretreatment is carried out to multiple modeled images, obtains multiple first figures
Picture, including:
By the size reduction of every modeled images to pre-set dimension, multiple first images are obtained.
In another embodiment of the disclosure, multiple second images carry out the second pretreatment, obtain multiple third images,
Including:
Multiple second images are normalized and thresholding is handled, obtain multiple third images with identical size.
In another embodiment of the disclosure, according to multiple feature vectors and corresponding quality tab, structure quality inspection
Model is surveyed, including:
Using each feature vector as independent variable input, using the corresponding quality tab of each feature vector as dependent variable
During output, the correspondence between feature vector and quality tab is obtained;
It is fitted by the correspondence between feature vector and quality tab, obtains quality testing model.
In another embodiment of the disclosure, it is based on quality testing model, objective image quality is detected, wrapped
It includes:
Obtain the feature vector of target image;
The feature vector of target image is input in quality testing model;
Export the corresponding quality tab of feature vector of target image.
The alternative embodiment that any combination forms the disclosure may be used, herein no longer in above-mentioned all optional technical solutions
It repeats one by one.
Fig. 2 is a kind of flow chart of the method for detection image quality shown according to an exemplary embodiment, such as Fig. 2 institutes
Show, the method for detection image quality is in server, including the following steps.
In step s 201, server obtains multiple pending modeled images, and there are one quality for every modeled images tool
Label.
In technical field of image processing, smart mobile phone, smart camera etc. have the imaging system of the mobile terminal of camera function
The factors such as system performance and mobile terminal parameter of taking pictures selected when taking pictures affect the quality of captured image.By right
Captured image be detected, it can be achieved that the calibration of mobile terminal image imaging system, to algorithm in imaging process and and
The optimization of parameter, so that mobile terminal can shoot the image of high quality.
It should be noted that for the ease of to involved in the present embodiment to each pending image distinguish, base
In the different purposes of each pending image in the present embodiment, image to be identified can be divided into modeled images and target figure
Picture.Wherein, modeled images are mainly used for building quality testing model, and target image is mainly used for inputting constructed quality testing
Model obtains quality measurements.
Wherein, every modeled images tool is there are one quality tab, which includes clear label, fuzzy label etc.,
It can be used for reacting the quality condition of modeled images.If the quality tab of any modeled images is clear label, illustrate that this is built
The quality of mould image is preferable;If the quality tab of any modeled images is fuzzy label, illustrate the quality of the modeled images
It is poor.
In the present embodiment, server can be obtained when obtaining the modeled images of multiple pending images from internet
Image of first preset quantity with clear label, and image of the second preset quantity with fuzzy label is obtained, in turn
Using the image with clear label got and the image with fuzzy label as modeled images.Wherein, the first present count
Amount can be 100,200,300 etc., and the second preset quantity can be 200,300,400 etc., the present embodiment
Specific limit is not made to the value of the first preset quantity and the second preset quantity.It should be noted that being chatted for the ease of subsequent
It states, image of the first preset quantity that the present embodiment obtains server with clear label is pre- by second as positive sample
If image of the quantity with fuzzy label is as negative sample.
In step S202, the size reduction of every modeled images to pre-set dimension is obtained multiple first figures by server
Picture.
In order to reduce calculation amount when being calculated multiple modeled images in subsequent step, server is getting multiple
After modeled images, will also every modeled images be carried out with the first pretreatment.For example, server can according to the quantity of modeled images and
The computing capability of server, determines a pre-set dimension, the pre-set dimension can be 1024 pixel *, 480 pixels, 720
360 pixels of pixel * etc., by the way that by the size reduction of every modeled images to pre-set dimension, multiple first images can be obtained.
By taking modeled images A as an example, the original size of modeled images A is long 1024 pixels, wide 720 pixels, is such as set
Pre-set dimension is long 320 pixels, wide 240 pixels can obtain long 320 pictures then by adjusting the length and width of modeled images A
First image A of plain, wide 240 pixels.
In step S203, server carries out Fourier transformation to multiple first images, obtains multiple second images.
In image processing field, time-domain analysis can only the amplitude of signal Analysis change with time situation, remove single-frequency components
Monochromatic wave outside, it is difficult to clearly disclose the frequency composition of signal and each frequency component size, and frequency-region signal can indicate that signal exists
The size of different frequency components is capable of providing more intuitive than time-domain signal, more rich information.
Time-domain signal can be converted into frequency-region signal, and to signal by Fourier transformation as a kind of signal analysis method
Energy is concentrated.For any image, after Fourier transformation, central point neighbouring position region is partially white, generation
Table rather low-frequency signals, marginal position is partially black, represents the signal of relative high frequency.In image processing field, low frequency signal represents flat
Mean value that is to say the profile information of image;High-frequency signal represents mutation, that is to say the information such as details, the texture of image.If will
Image after one width Fourier transformation can get details, texture of image etc. after high frequency filter filters low frequency signal
Information;Correspondingly, it if by the image after a width Fourier transformation after low-frequency filter filters high-frequency signal, can get
The profile information of image.
For any image after Fourier transformation, if marginal portion amplitude is higher, edge than more visible and
Comparison is relatively strong, then can determine that the image is more clear;If the amplitude of marginal portion is relatively low, edge is relatively fuzzyyer,
Then can determine that the image is relatively fuzzyyer, such as have ghost image etc..
After multiple first images are carried out Fourier transformation, you can obtain multiple second images.
In step S204, multiple second images are normalized server and thresholding processing, obtains multiple having
The third image of identical size.
In order to improve the accuracy of constructed quality testing model, server will also carry out threshold value to multiple second images
Change is handled, and is rejected and is higher than a pixel value or the image less than a pixel value in image.In addition, strong in different illumination to make
Captured image is with uniformity under the image-forming conditions such as degree, direction, distance, posture, and server passes through to multiple facial images
It translated, rotated, being scaled, the dimension normalizations processing such as standard cutting, to reduce in subsequent step to the second image
The calculation amount of reason.After above-mentioned normalization and thresholding processing, multiple thirds with identical size can be obtained in server
Image.The identical size can be 240 pixel *, 180 pixels etc., and the present embodiment does not have the size of identical size work
The restriction of body.
In step S205, server extracts multiple feature vectors from multiple third images.
For the image after Fourier transformation, multiple features, each feature available one are all had in every image
A column vector indicates.For the ease of, according to feature construction quality testing model included in third image, being taken in subsequent step
The feature that business device can be included according to every third image extracts the corresponding column vector of this feature from every third image, and
By the Column vector groups extracted from every third image at a matrix, the corresponding feature vector of every third image is obtained.
In step S206, server builds quality testing model according to multiple feature vectors and corresponding quality tab.
Since there are one quality tabs for each modeled images tool, and each feature vector is handled to modeled images
It is extracted in the third image obtained afterwards, thus each feature vector should also correspond to a quality tab.With feature vector
For in independent variable, the function that the corresponding quality tab of feature vector is dependent variable, by regarding each feature vector as independent variable
During exporting the corresponding quality tab of each feature vector as dependent variable, feature vector and quality can be obtained in input
Correspondence between label.It is fitted by the correspondence between feature vector and quality tab, can be obtained one
Model, the model are quality testing model.
In step S207, it is based on the quality testing model, server is detected objective image quality.
Based on constructed quality testing model, server can to it is any need carry out quality testing target image into
Row detection.Specific detection process is as follows:
First, server obtains the feature vector of target image.
About server obtain target image feature vector mode, reference can be made to above-mentioned steps S201 to step S205 from
The process of feature vector is extracted in modeled images, this embodiment is not repeated.
Secondly, the feature vector of target image is input in quality testing model by server, and exports target image
The corresponding quality tab of feature vector.
After getting the feature vector of target image, server is inputted the feature vector got as independent variable
Into the quality testing model built in advance, the quality testing model that builds in advance is according to the feature vector of input, by this feature
The corresponding quality tab output of vector, the quality tab of the output is the testing result to target image.
For example, the feature vector got is input in the quality testing model built in advance, if the matter built in advance
Feature vector of the detection model according to input is measured, the quality tab of output is clear label, then can determine the matter of the target image
Amount is clear.
The method that the embodiment of the present disclosure provides, by building quality testing model, and based on constructed quality testing mould
Type is detected the quality of target image, is participated in due to being not necessarily to user, thus avoids cognition standard and made to testing result
At interference so that testing result is more objective, accurate.
Fig. 3 is a kind of structural schematic diagram of the device of detection image quality shown according to an exemplary embodiment.Reference
Fig. 3, the device include:Acquisition module 301, first processing module 302, conversion module 303, Second processing module 304, extraction mould
Block 305, structure module 306, detection module 307.
The acquisition module 301 is configured as obtaining multiple pending modeled images, and there are one matter for every modeled images tool
Label is measured, which includes clear label and fuzzy label;
The first processing module 302 is configured as carrying out the first pretreatment to multiple modeled images, obtains multiple first figures
Picture;
The conversion module 303 is configured as carrying out Fourier transformation to multiple first images, obtains multiple second images;
The Second processing module 304 is configured as carrying out the second pretreatment to multiple second images, obtains multiple third figures
Picture;
The extraction module 305 is configured as extracting multiple feature vectors from multiple third images;
The structure module 306 is configured as, according to multiple feature vectors and corresponding quality tab, building quality testing mould
Type;
The detection module 307 is configured as being based on quality testing model, is detected to objective image quality.
In another embodiment of the disclosure, which is configured as the ruler of every modeled images
It is very little to be contracted to pre-set dimension, obtain multiple first images.
In another embodiment of the disclosure, which is configured as carrying out multiple the second images
Normalization and thresholding processing, obtain multiple third images with identical size.
In another embodiment of the disclosure, which is configured as using each feature vector as certainly
Variable input during exporting the corresponding quality tab of each feature vector as dependent variable, obtains feature vector and matter
Measure the correspondence between label;It is fitted by the correspondence between feature vector and quality tab, obtains quality
Detection model.
In another embodiment of the disclosure, which is configured as obtaining the feature vector of target image;
The feature vector of target image is input in quality testing model;Export the corresponding quality mark of feature vector of target image
Label.
The device that the embodiment of the present disclosure provides, by building quality testing model, and based on constructed quality testing mould
Type is detected the quality of target image, is participated in due to being not necessarily to user, thus avoids cognition standard and made to testing result
At interference so that testing result is more objective, accurate.
About the device in above-described embodiment, wherein modules execute the concrete mode of operation in related this method
Embodiment in be described in detail, explanation will be not set forth in detail herein.
Fig. 4 is a kind of block diagram of device 400 for detection image quality shown according to an exemplary embodiment.Example
Such as, device 400 may be provided as a server.With reference to Fig. 4, device 400 includes processing component 422, further comprises one
A or multiple processors, and by the memory resource representated by memory 432, it can holding by processing component 422 for storing
Capable instruction, such as application program.The application program stored in memory 432 may include it is one or more each
Corresponding to the module of one group of instruction.In addition, processing component 422 is configured as executing instruction, to execute above-mentioned detection image quality
Method.
Multiple pending modeled images are obtained, there are one quality tab, the quality tab packets for every modeled images tool
Include clear label and fuzzy label;
First pretreatment is carried out to multiple modeled images, obtains multiple first images;
Fourier transformation is carried out to multiple first images, obtains multiple second images;
Second pretreatment is carried out to multiple second images, obtains multiple third images;
Multiple feature vectors are extracted from multiple third images;
According to multiple feature vectors and corresponding quality tab, quality testing model is built;
Based on quality testing model, objective image quality is detected.
In another embodiment of the disclosure, the first pretreatment is carried out to multiple modeled images, obtains multiple first figures
Picture, including:
By the size reduction of every modeled images to pre-set dimension, multiple first images are obtained.
In another embodiment of the disclosure, multiple second images carry out the second pretreatment, obtain multiple third images,
Including:
Multiple second images are normalized and thresholding is handled, obtain multiple third images with identical size.
In another embodiment of the disclosure, according to multiple feature vectors and corresponding quality tab, structure quality inspection
Model is surveyed, including:
Using each feature vector as independent variable input, using the corresponding quality tab of each feature vector as dependent variable
During output, the correspondence between feature vector and quality tab is obtained;
It is fitted by the correspondence between feature vector and quality tab, obtains quality testing model.
In another embodiment of the disclosure, it is based on quality testing model, objective image quality is detected, wrapped
It includes:
Obtain the feature vector of target image;
The feature vector of target image is input in quality testing model;
Export the corresponding quality tab of feature vector of target image.
Device 400 can also include the power management that a power supply module 426 is configured as executive device 400, and one has
Line or radio network interface 450 are configured as device 400 being connected to network and input and output (I/O) interface 458.Dress
Setting 400 can operate based on the operating system for being stored in memory 432, such as Windows ServerTM, Mac OS XTM,
UnixTM,LinuxTM, FreeBSDTMOr it is similar.
The device that the embodiment of the present disclosure provides, by building quality testing model, and based on constructed quality testing mould
Type is detected the quality of target image, is participated in due to being not necessarily to user, thus avoids cognition standard and made to testing result
At interference so that testing result is more objective, accurate.
Those skilled in the art will readily occur to its of the disclosure after considering specification and putting into practice disclosure disclosed herein
Its embodiment.This application is intended to cover any variations, uses, or adaptations of the disclosure, these modifications, purposes or
Person's adaptive change follows the general principles of this disclosure and includes the undocumented common knowledge in the art of the disclosure
Or conventional techniques.The description and examples are only to be considered as illustrative, and the true scope and spirit of the disclosure are by following
Claim is pointed out.
It should be understood that the present disclosure is not limited to the precise structures that have been described above and shown in the drawings, and
And various modifications and changes may be made without departing from the scope thereof.The scope of the present disclosure is only limited by the accompanying claims.
Claims (7)
1. a kind of method of detection image quality, which is characterized in that the method includes:
Multiple pending modeled images are obtained, for every modeled images tool there are one quality tab, the quality tab includes clear
Clear label and fuzzy label;
By the size reduction of every modeled images to pre-set dimension, multiple first images are obtained;
Fourier transformation is carried out to multiple described first images, obtains multiple second images;
Multiple described second images are normalized and thresholding is handled, obtain multiple third images with identical size;
Multiple feature vectors are extracted from multiple described third images;
According to the multiple feature vector and corresponding quality tab, quality testing model is built;
Based on the quality testing model, objective image quality is detected.
2. according to the method described in claim 1, it is characterized in that, described according to the multiple feature vector and corresponding quality
Label builds quality testing model, including:
Each feature vector is being exported as independent variable input, using the corresponding quality tab of each feature vector as dependent variable
During, obtain the correspondence between feature vector and quality tab;
It is fitted by the correspondence between described eigenvector and quality tab, obtains quality testing model.
3. according to the method described in claim 1, it is characterized in that, described be based on the quality testing model, to target image
Quality is detected, including:
Obtain the feature vector of the target image;
The feature vector of the target image is input in the quality testing model;
Export the corresponding quality tab of feature vector of the target image.
4. a kind of device of detection image quality, which is characterized in that described device includes:
Acquisition module, for obtaining multiple pending modeled images, there are one quality tab, the matter for every modeled images tool
It includes clear label and fuzzy label to measure label;
First processing module, for by the size reduction of every modeled images to pre-set dimension, obtaining multiple first images;
Conversion module obtains multiple second images for carrying out Fourier transformation to multiple described first images;
Second processing module, for multiple described second images are normalized and thresholding processing, obtain it is multiple have phase
With the third image of size;
Extraction module, for extracting multiple feature vectors from multiple described third images;
Module is built, for according to the multiple feature vector and corresponding quality tab, building quality testing model;
Detection module is detected objective image quality for being based on the quality testing model.
5. device according to claim 4, which is characterized in that the structure module, for making by each feature vector
During independent variable input, exporting the corresponding quality tab of each feature vector as dependent variable, feature vector is obtained
Correspondence between quality tab;It is fitted by the correspondence between described eigenvector and quality tab,
Obtain quality testing model.
6. device according to claim 4, which is characterized in that the detection module, for obtaining the target image
Feature vector;The feature vector of the target image is input in the quality testing model;Export the target image
The corresponding quality tab of feature vector.
7. a kind of device of detection image quality, which is characterized in that including:
Processor;
Memory for storing the executable instruction of processor;
Wherein, the processor is configured as:
Multiple pending modeled images are obtained, for every modeled images tool there are one quality tab, the quality tab includes clear
Clear label and fuzzy label;
By the size reduction of every modeled images to pre-set dimension, multiple first images are obtained;
Fourier transformation is carried out to multiple described first images, obtains multiple second images;
Multiple described second images are normalized and thresholding is handled, obtain multiple third images with identical size;
Multiple feature vectors are extracted from multiple described third images;
According to the multiple feature vector and corresponding quality tab, quality testing model is built;
Based on the quality testing model, objective image quality is detected.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510719412.5A CN105357411B (en) | 2015-10-29 | 2015-10-29 | The method and device of detection image quality |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510719412.5A CN105357411B (en) | 2015-10-29 | 2015-10-29 | The method and device of detection image quality |
Publications (2)
Publication Number | Publication Date |
---|---|
CN105357411A CN105357411A (en) | 2016-02-24 |
CN105357411B true CN105357411B (en) | 2018-07-31 |
Family
ID=55333266
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510719412.5A Active CN105357411B (en) | 2015-10-29 | 2015-10-29 | The method and device of detection image quality |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN105357411B (en) |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102915449A (en) * | 2012-09-24 | 2013-02-06 | 华南理工大学 | Photo classification method |
CN103338380A (en) * | 2013-06-06 | 2013-10-02 | 宁波大学 | Adaptive image quality objective evaluation method |
CN104282028A (en) * | 2014-10-30 | 2015-01-14 | 电子科技大学 | Blurred image detection method fusing frequency spectrum information and cepstrum information |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO1998052011A2 (en) * | 1997-05-14 | 1998-11-19 | Emory University | Systems and methods for analyzing phantom images |
-
2015
- 2015-10-29 CN CN201510719412.5A patent/CN105357411B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102915449A (en) * | 2012-09-24 | 2013-02-06 | 华南理工大学 | Photo classification method |
CN103338380A (en) * | 2013-06-06 | 2013-10-02 | 宁波大学 | Adaptive image quality objective evaluation method |
CN104282028A (en) * | 2014-10-30 | 2015-01-14 | 电子科技大学 | Blurred image detection method fusing frequency spectrum information and cepstrum information |
Non-Patent Citations (1)
Title |
---|
personal identification based on iris texture analysis;Li Ma,et al;《IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE》;20031231;第25卷(第12期);第1519-1533页 * |
Also Published As
Publication number | Publication date |
---|---|
CN105357411A (en) | 2016-02-24 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108152278B (en) | Urine detection method and device | |
US7747080B2 (en) | System and method for scanning edges of a workpiece | |
Guo et al. | An improved Hough transform voting scheme utilizing surround suppression | |
CN110766679A (en) | Lens contamination detection method and device and terminal equipment | |
CN109886918A (en) | Wire drawing burr detection method, electronic equipment and the storage medium of lithium battery pole slice | |
CN104468916A (en) | Test method based on cell phone response time | |
CN105005972A (en) | Shooting distance based distortion correction method and mobile terminal | |
CN110207592A (en) | Building cracks measurement method, device, computer equipment and storage medium | |
CN110378227B (en) | Method, device and equipment for correcting sample labeling data and storage medium | |
CN111257329A (en) | Smartphone camera defect detection method and detection system | |
CN115908415B (en) | Edge-based defect detection method, device, equipment and storage medium | |
WO2020068362A1 (en) | Hybrid depth processing | |
CN112730251B (en) | Device and method for detecting screen color defects | |
CN111915541B (en) | Image enhancement processing method, device, equipment and medium based on artificial intelligence | |
CN111126147B (en) | Image processing method, device and electronic system | |
CN115587552A (en) | Grid optimization method and device, terminal equipment and storage medium | |
CN107016324A (en) | A kind of fingerprint image processing method and fingerprint detection equipment | |
CN111050027B (en) | Lens distortion compensation method, device, equipment and storage medium | |
CN105357411B (en) | The method and device of detection image quality | |
CN111639708B (en) | Image processing method, device, storage medium and equipment | |
CN115690747B (en) | Vehicle blind area detection model test method and device, electronic equipment and storage medium | |
CN112150413A (en) | Pipeline vibration detection method, device, equipment and storage medium | |
CN107231553A (en) | Corner location acquisition methods and device | |
CN116109962A (en) | Vibration frequency measurement method and device based on video phase, electronic equipment and storage medium | |
CN108734175A (en) | A kind of extracting method of characteristics of image, device and electronic equipment |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |