CN109978881A - A kind of method and apparatus of saliency processing - Google Patents
A kind of method and apparatus of saliency processing Download PDFInfo
- Publication number
- CN109978881A CN109978881A CN201910278429.XA CN201910278429A CN109978881A CN 109978881 A CN109978881 A CN 109978881A CN 201910278429 A CN201910278429 A CN 201910278429A CN 109978881 A CN109978881 A CN 109978881A
- Authority
- CN
- China
- Prior art keywords
- image
- saliency
- value
- group
- bval
- 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.)
- Granted
Links
- 238000012545 processing Methods 0.000 title claims abstract description 56
- 238000000034 method Methods 0.000 title claims abstract description 50
- KZSNJWFQEVHDMF-UHFFFAOYSA-N Valine Chemical compound CC(C)C(N)C(O)=O KZSNJWFQEVHDMF-UHFFFAOYSA-N 0.000 claims description 61
- 238000001514 detection method Methods 0.000 claims description 38
- 238000004458 analytical method Methods 0.000 claims description 28
- 230000008569 process Effects 0.000 claims description 23
- 238000012360 testing method Methods 0.000 claims description 2
- 230000008859 change Effects 0.000 description 10
- 238000010586 diagram Methods 0.000 description 3
- 238000003672 processing method Methods 0.000 description 3
- 239000003086 colorant Substances 0.000 description 2
- 238000010835 comparative analysis Methods 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000004321 preservation Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/01—Input arrangements or combined input and output arrangements for interaction between user and computer
- G06F3/011—Arrangements for interaction with the human body, e.g. for user immersion in virtual reality
- G06F3/013—Eye tracking input arrangements
-
- 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
Abstract
The invention discloses a kind of method and apparatus of saliency processing.The method of saliency processing comprises determining that the object of pending saliency processing in image；Image procossing is carried out to the object of the determination, obtains the optimal image of saliency.The method and apparatus handled using saliency of the invention can carry out conspicuousness processing for the object in image, simplify the complexity of saliency processing, meet user to the conspicuousness demand of image.
Description
Technical field
The present invention relates to field of image processing more particularly to a kind of saliency treating method and apparatus.
Background technique
Vision significance (Visual Attention Mechanism, VA, i.e. vision noticing mechanism) refers in face of one
When scene, the mankind automatically handle area-of-interest and selectively ignore region of loseing interest in, these people sense is emerging
Interesting region is referred to as salient region.With the propagation of internet bring big data quantity and the high speed development of artificial intelligence,
People increasingly pay attention to significance analysis, more and more significance analysis algorithms occur.These significance analysis algorithm master
It is used to calculate the conspicuousness of image, calculates complicated height, there is different adaptability to different images.Currently, people are urgent
The conspicuousness by changing some object in image is wanted to reach attracting purpose, this is required to carry out image
Conspicuousness processing.However, since the immature and not perfect of significance analysis algorithm causes to realize the conspicuousness processing of image
Complexity, and it is unable to reach expected conspicuousness effect.
Summary of the invention
The main purpose of the present invention is to provide a kind of method and apparatus of saliency processing, can be in image
Object carry out conspicuousness processing, simplify the complexity of saliency processing, meeting user needs the conspicuousness of image
It asks.
In order to solve the above-mentioned technical problems, the present invention provides a kind of methods of saliency processing, comprising:
Determine the object of pending saliency processing in image；
Image procossing is carried out to the object of the determination, obtains the optimal image of saliency.
In an exemplary embodiment, the above method also has the characteristics that following:
The object of the described pair of determination carries out image procossing, and obtaining the optimal image of saliency includes step S0-S3.
Step S0: brightness value bVal, contrast value cVal, tone value hVal, intensity value sVal are initialized as respectively
bVal0、cVal0、hVal0And sVal0, by bVal0、cVal0、hVal0And sVal0As current bVal, cVal, hVal, sVal
Value.
Step S1: image procossing is carried out according to object of current bVal, cVal, hVal, sVal value to the determination；And it saves
Image before processing is first group of image, saves step S0 treated that image is second group of image, and to described second group
Image carries out dynamic image pro cess, and the image after saving dynamic image pro cess is third group image.
Step S2: image detection analysis is carried out for first group of image, second group of image and third group image respectively.
Step S3: the result of the image detection analysis of above-mentioned three groups of images is compared and analyzed, saliency is obtained
Data, it is optimal whether the image saliency data judged according to default rule meets saliency, if being unsatisfactory for figure
Picture conspicuousness is optimal, then readjusts brightness value bVal, contrast value cVal, tone value according to obtained image saliency data
HVal, intensity value sVal, and by brightness value bVal, contrast value cVal, tone value hVal, the intensity value sVal after adjusting
As current bVal, cVal, hVal, sVal value, step S1 is continued to execute；If it is optimal to meet saliency, image is obtained
The optimal image of conspicuousness terminates process.
In an exemplary embodiment, the above method also has the characteristics that following:
The step S2: image detection is carried out for first group of image, second group of image and third group image respectively
Analysis includes:
The most object of saliency is detected in first group of image, second group of image and third group image respectively, with
And detect human eye residence time and the regularity of distribution on the object.
In an exemplary embodiment, the above method also has the characteristics that following:
If it is optimal to be unsatisfactory for saliency according to the image saliency data that default rule judges, in basis
Obtained image saliency data readjust brightness value bVal, contrast value cVal, tone value hVal, intensity value sVal it
Before, the method also includes:
Judge the times N that the step S1 is performed, if the number that step S1 is performed is more than or equal to preset threshold value,
Optimal image saliency data is searched in the image saliency data of obtained n times, and the image of best performance is shown
The work property corresponding image of the data image optimal as saliency.
In an exemplary embodiment, the above method also has the characteristics that following:
The regularity of distribution that stops on the object of detection human eye includes:
The region that stops on the object of detection human eye, on the area residence time, and the region that stops it is bright
Angle value, contrast value, tone value and intensity value obtain the regularity of distribution that human eye stops on the object.
To solve the above-mentioned problems, the present invention also provides a kind of saliency processing units, comprising: memory and place
Manage device；Wherein:
The memory, for storing the program for being used for saliency processing；
The processor executes the program for saliency processing for reading, performs the following operations:
Determine the object of pending saliency processing in image；
Image procossing is carried out to the object of the determination, obtains the optimal image of saliency.
In an exemplary embodiment, the processor is used to read the journey for executing the processing of described image conspicuousness
Sequence executes and carries out image procossing to the object of the determination, and obtaining the optimal image of saliency includes step S0-S3.
Step S0: brightness value bVal, contrast value cVal, tone value hVal, intensity value sVal are initialized as respectively
bVal0、cVal0、hVal0And sVal0, by bVal0、cVal0、hVal0And sVal0As current bVal, cVal, hVal, sVal
Value.
Step S1: image procossing is carried out according to object of current bVal, cVal, hVal, sVal value to the determination；And it saves
Image before processing is first group of image, saves step S0 treated that image is second group of image, and to described second group
Image carries out dynamic image pro cess, and the image after saving dynamic image pro cess is third group image.
Step S2: image detection analysis is carried out for first group of image, second group of image and third group image respectively.
Step S3: the result of the image detection analysis of above-mentioned three groups of images is compared and analyzed, saliency is obtained
Data, it is optimal whether the image saliency data judged according to default rule meets saliency, if being unsatisfactory for figure
Picture conspicuousness is optimal, then readjusts brightness value bVal, contrast value cVal, tone value according to obtained image saliency data
HVal, intensity value sVal, and by brightness value bVal, contrast value cVal, tone value hVal, the intensity value sVal after adjusting
As current bVal, cVal, hVal, sVal value, step S1 is continued to execute；If it is optimal to meet saliency, image is obtained
The optimal image of conspicuousness terminates process.
In an exemplary embodiment, the processor is used to read the journey for executing the processing of described image conspicuousness
The step S2 of execution: sequence carries out image detection for first group of image, second group of image and third group image respectively
Analysis includes:
The most object of saliency is detected in first group of image, second group of image and third group image respectively, with
And detect human eye residence time and the regularity of distribution on the object.
In an exemplary embodiment, the processor is used to read the journey for executing the processing of described image conspicuousness
Sequence also performs the following operations:
If it is optimal to be unsatisfactory for saliency according to the image saliency data that default rule judges, in basis
Obtained image saliency data readjust brightness value bVal, contrast value cVal, tone value hVal, intensity value sVal it
Before,
Judge the times N that the step S1 is performed, if the number that step S1 is performed is more than or equal to preset threshold value,
Optimal image saliency data is searched in the image saliency data of obtained n times, and the image of best performance is shown
The work property corresponding image of the data image optimal as saliency.
In an exemplary embodiment, the regularity of distribution that stops on the object of detection human eye includes:
The region that stops on the object of detection human eye, on the area residence time, and the region that stops it is bright
Angle value, contrast value, tone value and intensity value obtain the regularity of distribution that human eye stops on the object.
To sum up, the method for the saliency processing of the application determines pending saliency processing in image first
Then object carries out image procossing to the object of the determination, obtain the optimal image of saliency.The image of the application is significant
Property processing method and apparatus, compared with the existing technology, simplify saliency processing complexity, meet user to figure
The conspicuousness demand of picture.
Detailed description of the invention
Fig. 1 is the flow chart for the method that the saliency of the embodiment of the present invention is handled.
Fig. 2 is the specific flow chart for obtaining the optimal image of saliency of the embodiment of the present invention.
Fig. 3 is the schematic diagram of the saliency processing unit of the embodiment of the present invention.
Specific embodiment
To make the objectives, technical solutions, and advantages of the present invention clearer, below in conjunction with attached drawing to the present invention
Embodiment be described in detail.It should be noted that in the absence of conflict, in the embodiment and embodiment in the application
Feature can mutual any combination.
Fig. 1 is the flow chart of the method for the saliency processing of the embodiment of the present invention.According to the flow chart of Fig. 1, this reality
Apply example saliency processing method include:
Step A: the object of pending saliency processing in image is determined；
Step B: image procossing is carried out to the object of the determination, obtains the optimal image of saliency.
In an exemplary embodiment, step B: image procossing is carried out to the object of the determination, it is significant to obtain image
The optimal image of property includes step S0-S3.
Step S0: brightness value bVal, contrast value cVal, tone value hVal, intensity value sVal are initialized as respectively
bVal0、cVal0、hVal0And sVal0, by bVal0、cVal0、hVal0And sVal0As current bVal, cVal, hVal, sVal
Value.
Step S1: image procossing is carried out according to object of current bVal, cVal, hVal, sVal value to the determination；And it saves
Image before processing is first group of image, saves step S0 treated that image is second group of image, and to described second group
Image carries out dynamic image pro cess, and the image after saving dynamic image pro cess is third group image.
Step S2: image detection analysis is carried out for first group of image, second group of image and third group image respectively.
Step S3: the result of the image detection analysis of above-mentioned three groups of images is compared and analyzed, saliency is obtained
Data, it is optimal whether the image saliency data judged according to default rule meets saliency, if being unsatisfactory for figure
Picture conspicuousness is optimal, then readjusts brightness value bVal, contrast value cVal, tone value according to obtained image saliency data
HVal, intensity value sVal, and by brightness value bVal, contrast value cVal, tone value hVal, the intensity value sVal after adjusting
As current bVal, cVal, hVal, sVal value, step S1 is continued to execute；If it is optimal to meet saliency, image is obtained
The optimal image of conspicuousness terminates process.
As shown in Fig. 2, step S3 includes sub-step S30, S31, S32 and S33.
Wherein, sub-step S30: the result of the image detection analysis of above-mentioned three groups of images is compared and analyzed, figure is obtained
As saliency data.
Sub-step S31: whether the image saliency data judged according to default rule meets saliency most
It is excellent；If it is optimal to meet saliency, sub-step S32 is executed；If it is optimal to be unsatisfactory for saliency, sub-step is executed
S33。
Sub-step S32: second group after will be currently processed and third group image it is optimal as the saliency obtained
Image, and terminate the process of saliency processing.
Sub-step S33: brightness value bVal, contrast value cVal, color are readjusted according to obtained image saliency data
Tone pitch hVal and intensity value sVal, and by brightness value bVal, contrast value cVal, tone value hVal and the saturation degree after adjusting
Value sVal continues to execute step S1 as current bVal, cVal, hVal and sVal value.
In an exemplary embodiment, according to current bVal, cVal, hVal, sVal value to the object of the determination into
Row image procossing may include:
It is public according to the brightness-formula of image, contrast formula, tone respectively according to current bVal, cVal, hVal, sVal value
Formula and saturation degree formula carry out image procossing to the object of the determination.
Wherein, the brightness-formula (1) of image is as follows:
The contrast formula (2) of image and (3) are as follows:
CVal=Average+ (RGB-Average) * (1+percentage) (2)
The tone formula (4) of image is as follows:
The saturation degree formula (5) of image is as follows:
Wherein, MaxV=max (R, G, B), MinV=min (R, G, B), nRGB indicate three primary colors color value adjusted,
RGB expression three primary colors color value, Average expression average brightness, percentage expression adjustment percentage, bVal, cVal,
HVal and sVal respectively indicates brightness value, contrast value, tone value and intensity value.
In the embodiment of another exemplary, due to certain restrictive condition, brightness value, contrast value, the tone value of image
Need to meet certain constraint condition with the one or more in intensity value, for example, tone value requires to be a fixed value, or
Intensity value requires within the scope of some, or while tone value requires to be a fixed value, it is desirable that intensity value is at certain
In a range, at this time in step sl image is carried out to the object of the determination according to current bVal, cVal, hVal and sVal value
Processing may include:
Under the premise of meeting constraint condition, according to one or more of current bVal, cVal, hVal and sVal value pair
The object of the determination carries out image procossing.
In an exemplary embodiment, it step S2: is directed to first group of image, second group of image and third respectively
Group image carries out image detection analysis
The most object of saliency is detected in first group of image, second group of image and third group image respectively, with
And detect human eye residence time and the regularity of distribution on the object.
It, can be using eye tracker to first group of image, second group of image and third in another exemplary embodiment
Group image carries out image detection analysis.
In an exemplary embodiment, whether the image saliency data in above-mentioned steps S3 for judging to obtain is full
The optimal default rule of sufficient saliency can be what user was arranged according to demand, for example, it may be human eye is at second group
The residence time on the object of saliency that is desired to have in image is greater than T seconds, expectation of the human eye in third group image
Residence time is not less than T seconds on object with saliency, or can be expectation of the human eye in second group of image
Residence time is the object that is desired to have saliency of the human eye in first group of image on object with saliency
2 times (or other numerical value) etc. of residence time on body.Certainly, which can also set based on experience value
It sets.The application to the particular content of default rule without limitation.
In an exemplary embodiment, if being unsatisfactory for according to the image saliency data that default rule judges
Saliency is optimal, then is readjusting brightness value bVal, contrast value cVal, color according to obtained image saliency data
Before tone pitch hVal, intensity value sVal, the method also includes:
Judge the times N that the step S1 is performed, if the number that step S1 is performed is more than or equal to preset threshold value,
Optimal image saliency data is searched in the image saliency data of obtained n times, and the image of best performance is shown
The work property corresponding image of the data image optimal as saliency.
In another exemplary embodiment, the regularity of distribution that detection human eye stops on the object includes:
The region that stops on the object of detection human eye, on the area residence time, and the region that stops it is bright
Angle value, contrast value, tone value and intensity value obtain the regularity of distribution that human eye stops on the object.
In an exemplary embodiment, in step s3, to the result of the image detection analysis of above-mentioned three groups of images
It compares and analyzes and includes:
The object for the most saliency in every group of image analyzed according to the image detection of above-mentioned three groups of images
Body is analyzed after being handled image, and most whether the object of saliency is changed；If being changed
Become, judge change after most saliency object whether be desired most saliency object, according to
Parameter differences between three groups of images compare the variation of the most object of saliency；
If no change has taken place, continue the object for analyzing most saliency of the obtained human eye in every group of image
Upper residence time and the regularity of distribution compare the change of residence time and the regularity of distribution according to the parameter differences between three groups of images
Change；
Record above-mentioned comparative analysis as a result, obtaining image saliency data.
In another exemplary embodiment, it can use unified template and save each image saliency data,
In order to analyze image saliency data.It, can be with alternatively, when the object of most saliency does not change
The form of curve saves brightness value bVal, contrast value cVal, tone value hVal, intensity value sVal and human eye residence time
Relationship, so as to the more convenient relationship intuitively observed between them, the application is not specifically limited in this embodiment.
To sum up, the method for the saliency processing of the application determines pending saliency processing in image first
Then object carries out image procossing to the object of the determination, obtain the optimal image of saliency.The image of the application is significant
Property processing method and apparatus, compared with the existing technology, simplify saliency processing complexity, meet user to figure
The conspicuousness demand of picture.
Fig. 3 is the schematic diagram of the saliency processing unit of the embodiment of the present invention.Schematic diagram according to Fig.3, this
The saliency processing unit of embodiment includes memory and processor.Wherein:
The memory 100, for storing the program for being used for saliency processing；
The processor 200 executes the program for saliency processing for reading, performs the following operations:
Determine the object of pending saliency processing in image；
Image procossing is carried out to the object of the determination, obtains the optimal image of saliency.
In an exemplary embodiment, processor 200 is used to read the program for executing the processing of described image conspicuousness,
It executes and image procossing is carried out to the object of the determination, obtaining the optimal image of saliency includes:
Step S0: brightness value bVal, contrast value cVal, tone value hVal, intensity value sVal are initialized as respectively
bVal0、cVal0、hVal0And sVal0, by bVal0、cVal0、hVal0And sVal0As current bVal, cVal, hVal, sVal
Value.
Step S1: image procossing is carried out according to object of current bVal, cVal, hVal, sVal value to the determination；And it saves
Image before processing is first group of image, saves step S0 treated that image is second group of image, and to described second group
Image carries out dynamic image pro cess, and the image after saving dynamic image pro cess is third group image；
Step S2: image detection analysis is carried out for first group of image, second group of image and third group image respectively；
Step S3: the result of the image detection analysis of above-mentioned three groups of images is compared and analyzed, saliency is obtained
Data, it is optimal whether the image saliency data judged according to default rule meets saliency, if being unsatisfactory for figure
Picture conspicuousness is optimal, then readjusts brightness value bVal, contrast value cVal, tone value according to obtained image saliency data
HVal, intensity value sVal, and by brightness value bVal, contrast value cVal, tone value hVal, the intensity value sVal after adjusting
As current bVal, cVal, hVal, sVal value, step S1 is continued to execute；If it is optimal to meet saliency, image is obtained
The optimal image of conspicuousness terminates process.
In an exemplary embodiment, according to current bVal, cVal, hVal, sVal value to the object of the determination into
Row image procossing may include:
It is public according to the brightness-formula of image, contrast formula, tone respectively according to current bVal, cVal, hVal, sVal value
Formula and saturation degree formula carry out image procossing to the object of the determination.
Wherein, brightness-formula, contrast formula, tone formula and saturation degree formula are specifically shown in formula above-mentioned (1)-(5).
In the embodiment of another exemplary, due to certain restrictive condition, brightness value, contrast value, the tone value of image
Need to meet certain constraint condition with the one or more in intensity value, for example, tone value requires to be a fixed value, or
Intensity value requires within the scope of some, or while tone value requires to be a fixed value, it is desirable that intensity value is at certain
In a range, at this time in step sl image is carried out to the object of the determination according to current bVal, cVal, hVal, sVal value
Processing may include:
Under the premise of meeting constraint condition, according to one or more of current bVal, cVal, hVal, sVal value pair
The object of the determination carries out image procossing.
In an exemplary embodiment, processor 200 is used to read the program for executing the processing of described image conspicuousness,
The step S2 executed: image detection point is carried out for first group of image, second group of image and third group image respectively
Analysis includes:
The most object of saliency is detected in first group of image, second group of image and third group image respectively, with
And detect human eye residence time and the regularity of distribution on the object.
It, can be using eye tracker to first group of image, second group of image and third in another exemplary embodiment
Group image carries out image detection analysis.
In an exemplary embodiment, whether the image saliency data in above-mentioned steps S3 for judging to obtain is full
The optimal default rule of sufficient saliency can be what user was arranged according to demand, for example, it may be human eye is at second group
The residence time on the object of saliency that is desired to have in image is greater than T seconds, expectation of the human eye in third group image
Residence time is not less than T seconds on object with saliency, or can be expectation of the human eye in second group of image
Residence time is the object that is desired to have saliency of the human eye in first group of image on object with saliency
2 times (or other numerical value) etc. of residence time on body.Certainly, which can also set based on experience value
It sets.The application to the particular content of default rule without limitation.
In an exemplary embodiment, processor 200 is used to read the program for executing the processing of described image conspicuousness,
Also perform the following operations:
If it is optimal to be unsatisfactory for saliency according to the image saliency data that default rule judges, in basis
Obtained image saliency data readjust brightness value bVal, contrast value cVal, tone value hVal, intensity value sVal it
Before,
Judge the times N that the step S1 is performed, if the number that step S1 is performed is more than or equal to preset threshold value,
Optimal image saliency data is searched in the image saliency data of obtained n times, and the image of best performance is shown
The work property corresponding image of the data image optimal as saliency.
In another exemplary embodiment, the regularity of distribution that detection human eye stops on the object includes:
The region that stops on the object of detection human eye, on the area residence time, and the region that stops it is bright
Angle value, contrast value, tone value and intensity value obtain the regularity of distribution that human eye stops on the object.
In an exemplary embodiment, in step s3, to the result of the image detection analysis of above-mentioned three groups of images
It compares and analyzes and includes:
The object for the most saliency in every group of image analyzed according to the image detection of above-mentioned three groups of images
Body is analyzed after being handled image, and most whether the object of saliency is changed；If being changed
Become, judge change after most saliency object whether be desired most saliency object, according to
Parameter differences between three groups of images compare the variation of the most object of saliency；
If no change has taken place, continue the object for analyzing most saliency of the obtained human eye in every group of image
Upper residence time and the regularity of distribution compare the change of residence time and the regularity of distribution according to the parameter differences between three groups of images
Change；
Record above-mentioned comparative analysis as a result, obtaining image saliency data.
In another exemplary embodiment, it can use unified template and save each image saliency data,
In order to analyze image saliency data.It, can be with alternatively, when the object of most saliency does not change
The form of curve saves brightness value bVal, contrast value cVal, tone value hVal, intensity value sVal and human eye residence time
Relationship, so as to the more convenient relationship intuitively observed between them, the application is not specifically limited in this embodiment.
It is further described below in the method that concrete application example handles saliency.
Step 1: initialization.The initialization includes the initialization of parameter, further includes the initialization of rule.Wherein, parameter
Initialization includes that brightness value bVal, contrast value cVal, tone value hVal, intensity value sVal are initialized as bVal respectively0、
cVal0、hVal0And sVal0.The initialization of rule includes the initialization of default rule, preset threshold value etc..
Step 2: the object of pending saliency processing in image is determined.
Step 3: it is carried out according to formula (1)-(5) according to object of current bVal, cVal, hVal, sVal value to the determination
Image procossing.
Step 4: the image before preservation processing is first group of image, and saving step 3 treated image is the second group picture
Picture, and dynamic image pro cess is carried out to second group of image, the image after saving dynamic image pro cess is third group image；It will be real
The person of testing is divided into three groups, respectively corresponds three groups of images.
Step 5: experimenter wears eye tracker, clicks the object for oneself thinking to attract him, and eye tracker detects in the process
Experimenter residence time and regularity of distribution on object obtain the result of image detection analysis corresponding with three groups of images.
Step 6: the result of the image detection analysis of above-mentioned three groups of images is compared and analyzed, saliency is obtained
Data, it is optimal whether the image saliency data judged according to default rule meets saliency, if being unsatisfactory for figure
Picture conspicuousness is optimal, then readjusts brightness value bVal, contrast value cVal, tone value according to obtained image saliency data
HVal, intensity value sVal, and by brightness value bVal, contrast value cVal, tone value hVal, the intensity value sVal after adjusting
As current bVal, cVal, hVal, sVal value, step 3 is continued to execute；If it is optimal to meet saliency, image is obtained
The optimal image of conspicuousness terminates process.
Between each step in above-mentioned concrete application example and there is no stringent to execute sequence, those skilled in the art
Can according to need carry out each step of reasonable arrangement execute sequence.
Those of ordinary skill in the art will appreciate that all or part of the steps in the above method can be instructed by program
Related hardware is completed, and described program can store in computer readable storage medium, such as read-only memory, disk or CD
Deng.Optionally, one or more integrated circuits can be used also to realize in all or part of the steps of above-described embodiment.Accordingly
Ground, each module/unit in above-described embodiment can take the form of hardware realization, can also use the shape of software function module
Formula is realized.The present invention is not limited to the combinations of the hardware and software of any particular form.
The above is only a preferred embodiment of the present invention, and certainly, the invention may also have other embodiments, without departing substantially from this
In the case where spirit and its essence, those skilled in the art make various corresponding changes in accordance with the present invention
And deformation, but these corresponding changes and modifications all should fall within the scope of protection of the appended claims of the present invention.
Claims (10)
1. a kind of method of saliency processing, comprising:
Determine the object of pending saliency processing in image；
Image procossing is carried out to the object of the determination, obtains the optimal image of saliency.
2. the method as described in claim 1, which is characterized in that the object of the described pair of determination carries out image procossing, is schemed
As the optimal image of conspicuousness includes:
Step S0: brightness value bVal, contrast value cVal, tone value hVal, intensity value sVal are initialized as bVal respectively0、
cVal0、hVal0And sVal0, by bVal0、cVal0、hVal0And sVal0As current bVal, cVal, hVal, sVal value；
Step S1: image procossing is carried out according to object of current bVal, cVal, hVal, sVal value to the determination；And save processing
Preceding image is first group of image, and saving step S0 treated image is second group of image, and to second group of image
Dynamic image pro cess is carried out, the image after saving dynamic image pro cess is third group image；
Step S2: image detection analysis is carried out for first group of image, second group of image and third group image respectively；
Step S3: comparing and analyzing the result of the image detection analysis of above-mentioned three groups of images, obtain image saliency data,
It is optimal whether the image saliency data judged according to default rule meets saliency, if it is significant to be unsatisfactory for image
Property it is optimal, then according to obtained image saliency data readjust brightness value bVal, contrast value cVal, tone value hVal,
Intensity value sVal, and using after adjusting brightness value bVal, contrast value cVal, tone value hVal, intensity value sVal as
Current bVal, cVal, hVal, sVal value, continues to execute step S1；If it is optimal to meet saliency, it is significant to obtain image
Property optimal image, terminate process.
3. method according to claim 2, which is characterized in that the step S2: being directed to first group of image, second respectively
Group image and third group image carry out image detection analysis
The most object of saliency, Yi Jijian are detected in first group of image, second group of image and third group image respectively
Survey human eye residence time and the regularity of distribution on the object.
4. method as claimed in claim 2 or claim 3, which is characterized in that if significant according to the image that default rule judges
Property data be unsatisfactory for that saliency is optimal, then brightness value bVal, right is being readjusted according to obtained image saliency data
Before angle value cVal, tone value hVal, intensity value sVal, the method also includes:
The times N that the step S1 is performed is judged, if the number that step S1 is performed is more than or equal to preset threshold value,
Search optimal image saliency data in the image saliency data of obtained n times, and by the saliency of best performance
The corresponding image of the data image optimal as saliency.
5. method as claimed in claim 3, which is characterized in that the regularity of distribution packet that the detection human eye stops on the object
It includes:
The brightness value in the region that stops on the object of detection human eye, on the area residence time, and the region that stops,
Contrast value, tone value and intensity value obtain the regularity of distribution that human eye stops on the object.
6. a kind of saliency processing unit, comprising: memory and processor；Wherein:
The memory, for storing the program for being used for saliency processing；
The processor executes the program for saliency processing for reading, performs the following operations:
Determine the object of pending saliency processing in image；
Image procossing is carried out to the object of the determination, obtains the optimal image of saliency.
7. device as claimed in claim 6, which is characterized in that the processor executes at described image conspicuousness for reading
The program of reason executes and carries out image procossing to the object of the determination, obtains the optimal image of saliency and includes:
Step S0: brightness value bVal, contrast value cVal, tone value hVal, intensity value sVal are initialized as bVal respectively0、
cVal0、hVal0And sVal0, by bVal0、cVal0、hVal0And sVal0As current bVal, cVal, hVal, sVal value；
Step S1: image procossing is carried out according to object of current bVal, cVal, hVal, sVal value to the determination；And save processing
Preceding image is first group of image, and saving step S0 treated image is second group of image, and to second group of image
Dynamic image pro cess is carried out, the image after saving dynamic image pro cess is third group image；
Step S2: image detection analysis is carried out for first group of image, second group of image and third group image respectively；
Step S3: comparing and analyzing the result of the image detection analysis of above-mentioned three groups of images, obtain image saliency data,
It is optimal whether the image saliency data judged according to default rule meets saliency, if it is significant to be unsatisfactory for image
Property it is optimal, then according to obtained image saliency data readjust brightness value bVal, contrast value cVal, tone value hVal,
Intensity value sVal, and using after adjusting brightness value bVal, contrast value cVal, tone value hVal, intensity value sVal as
Current bVal, cVal, hVal, sVal value, continues to execute step S1；If it is optimal to meet saliency, it is significant to obtain image
Property optimal image, terminate process.
8. device as claimed in claim 7, which is characterized in that the processor executes at described image conspicuousness for reading
The step S2 of execution: the program of reason carries out figure for first group of image, second group of image and third group image respectively
Include: as testing and analyzing
The most object of saliency, Yi Jijian are detected in first group of image, second group of image and third group image respectively
Survey human eye residence time and the regularity of distribution on the object.
9. device as claimed in claim 7 or 8, which is characterized in that the processor is significant for reading execution described image
Property processing program, also perform the following operations:
If it is optimal to be unsatisfactory for saliency according to the image saliency data that default rule judges, according to obtaining
Image saliency data readjust brightness value bVal, contrast value cVal, tone value hVal, intensity value sVal before,
The times N that the step S1 is performed is judged, if the number that step S1 is performed is more than or equal to preset threshold value,
Search optimal image saliency data in the image saliency data of obtained n times, and by the saliency of best performance
The corresponding image of the data image optimal as saliency.
10. device as claimed in claim 8, which is characterized in that the regularity of distribution that the detection human eye stops on the object
Include:
The brightness value in the region that stops on the object of detection human eye, on the area residence time, and the region that stops,
Contrast value, tone value and intensity value obtain the regularity of distribution that human eye stops on the object.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910278429.XA CN109978881B (en) | 2019-04-09 | 2019-04-09 | Image saliency processing method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910278429.XA CN109978881B (en) | 2019-04-09 | 2019-04-09 | Image saliency processing method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109978881A true CN109978881A (en) | 2019-07-05 |
| CN109978881B CN109978881B (en) | 2021-11-26 |
Family
ID=67083504
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910278429.XA Active CN109978881B (en) | 2019-04-09 | 2019-04-09 | Image saliency processing method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109978881B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2021190433A1 (en) * | 2020-03-24 | 2021-09-30 | 华为技术有限公司 | Method and device for updating object recognition model |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101098489A (en) * | 2006-06-27 | 2008-01-02 | 三星电子株式会社 | Image processing apparatus and method of enhancing visibility of displayed image |
| CN101980248A (en) * | 2010-11-09 | 2011-02-23 | 西安电子科技大学 | Improved visual attention model-based method of natural scene object detection |
| CN102999908A (en) * | 2012-11-19 | 2013-03-27 | 西安电子科技大学 | Synthetic aperture radar (SAR) airport segmentation method based on improved visual attention model |
| CN105023253A (en) * | 2015-07-16 | 2015-11-04 | 上海理工大学 | Visual underlying feature-based image enhancement method |
| CN105574866A (en) * | 2015-12-15 | 2016-05-11 | 努比亚技术有限公司 | Image processing method and apparatus |
| CN106775527A (en) * | 2016-12-07 | 2017-05-31 | 中国联合网络通信集团有限公司 | Adjust method, device and the display device of the display parameters of display panel |
| US20170206426A1 (en) * | 2016-01-15 | 2017-07-20 | Ford Global Technologies, Llc | Pedestrian Detection With Saliency Maps |
| CN107909553A (en) * | 2017-11-02 | 2018-04-13 | 青岛海信电器股份有限公司 | A kind of image processing method and equipment |
| CN108647695A (en) * | 2018-05-02 | 2018-10-12 | 武汉科技大学 | Soft image conspicuousness detection method based on covariance convolutional neural networks |
| CN108696732A (en) * | 2017-02-17 | 2018-10-23 | 北京三星通信技术研究有限公司 | Wear the method for adjusting resolution and equipment of display equipment |
-
2019
- 2019-04-09 CN CN201910278429.XA patent/CN109978881B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101098489A (en) * | 2006-06-27 | 2008-01-02 | 三星电子株式会社 | Image processing apparatus and method of enhancing visibility of displayed image |
| CN101980248A (en) * | 2010-11-09 | 2011-02-23 | 西安电子科技大学 | Improved visual attention model-based method of natural scene object detection |
| CN102999908A (en) * | 2012-11-19 | 2013-03-27 | 西安电子科技大学 | Synthetic aperture radar (SAR) airport segmentation method based on improved visual attention model |
| CN105023253A (en) * | 2015-07-16 | 2015-11-04 | 上海理工大学 | Visual underlying feature-based image enhancement method |
| CN105574866A (en) * | 2015-12-15 | 2016-05-11 | 努比亚技术有限公司 | Image processing method and apparatus |
| US20170206426A1 (en) * | 2016-01-15 | 2017-07-20 | Ford Global Technologies, Llc | Pedestrian Detection With Saliency Maps |
| CN106775527A (en) * | 2016-12-07 | 2017-05-31 | 中国联合网络通信集团有限公司 | Adjust method, device and the display device of the display parameters of display panel |
| CN108696732A (en) * | 2017-02-17 | 2018-10-23 | 北京三星通信技术研究有限公司 | Wear the method for adjusting resolution and equipment of display equipment |
| CN107909553A (en) * | 2017-11-02 | 2018-04-13 | 青岛海信电器股份有限公司 | A kind of image processing method and equipment |
| CN108647695A (en) * | 2018-05-02 | 2018-10-12 | 武汉科技大学 | Soft image conspicuousness detection method based on covariance convolutional neural networks |
Non-Patent Citations (4)
| Title |
|---|
| HOOMAN M.等: "Image Saliency Detection By Residual And Inception-like CNNs", 《2018 6TH RSI INTERNATIONAL CONFERENCE ON ROBOTICS AND MECHATRONICS (ICROM)》 * |
| 何鹏超: "改进的卷积神经网络模型及其应用研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 李晶 著: "《均衡认知负荷的人机界面信息编码和研究》", 30 September 2018, 南京：东南大学出版社 * |
| 杨杰 等: "《视频目标检测和跟踪及其应用》", 31 December 2012 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2021190433A1 (en) * | 2020-03-24 | 2021-09-30 | 华为技术有限公司 | Method and device for updating object recognition model |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109978881B (en) | 2021-11-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11410277B2 (en) | Method and device for blurring image background, storage medium and electronic apparatus | |
| US9483835B2 (en) | Depth value restoration method and system | |
| WO2019100282A1 (en) | Face skin color recognition method, device and intelligent terminal | |
| US11610433B2 (en) | Skin tone assisted digital image color matching | |
| US9639965B2 (en) | Adjusting color attribute of an image in a non-uniform way | |
| WO2017088365A1 (en) | Skin-colour detection method and apparatus | |
| US8792728B2 (en) | Near-duplicate image detection | |
| CN108038466B (en) | Multi-channel human eye closure recognition method based on convolutional neural network | |
| CN105404846B (en) | A kind of image processing method and device | |
| US20190228231A1 (en) | Video segmentation using predictive models trained to provide aesthetic scores | |
| CN107507144B (en) | Skin color enhancement processing method and device and image processing device | |
| US9881202B2 (en) | Providing visual effects for images | |
| WO2012122682A1 (en) | Method for calculating image visual saliency based on color histogram and overall contrast | |
| CN108292359B (en) | Eye feature recognition | |
| CN106210522B (en) | A kind of image processing method, device, mobile terminal and readable storage medium storing program for executing | |
| KR20180109658A (en) | Apparatus and method for image processing | |
| US20040120599A1 (en) | Detection and enhancement of backlit images | |
| CN111476849A (en) | Object color recognition method and device, electronic equipment and storage medium | |
| WO2017177616A1 (en) | Face recognition method and device, and picture displaying method and device | |
| Ciesla et al. | Eye pupil location using webcam | |
| CN109949248A (en) | Modify method, apparatus, equipment and the medium of the color of vehicle in the picture | |
| Abeln et al. | Preference for well-balanced saliency in details cropped from photographs | |
| CN109978881A (en) | A kind of method and apparatus of saliency processing | |
| Lou et al. | Smoke root detection from video sequences based on multi-feature fusion | |
| CN114187201A (en) | Model training method, image processing method, device, 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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |