CN109741283A - A kind of method and apparatus for realizing smart filter - Google Patents

Abstract

The invention belongs to field of image processing more particularly to a kind of method and apparatus for realizing smart filter.Extract the primitive character information of original image；Filter is used to original image, obtains the optimization image that user saves, and extract the characteristic information of filter；Primitive character information, filter characteristic information and optimization image are stored in feature database in a manner of one-to-one, construct user preference feature database；User is instructed to the use of filter, to realize smart filter next time using user preference feature database.It solves and the machinery of filter is used in the prior art, the intelligent recommendation and intelligent use of filter are realized, so that filter using effect is more preferably.

Description

A kind of method and apparatus for realizing smart filter

Technical field

The invention belongs to field of image processing more particularly to a kind of method and apparatus for realizing smart filter.

Background technique

Currently, the image for shooting adds filter, carry out beautifying picture, is widely used in intelligent terminal and shooting eventuallyIn end.However, in the prior art, in the software of beautifying picture, being named by the effect to filter, and pass through an exampleImage shows the effect of filter, and user needs to browse all filtering effects figures to search the filter oneself liked, andWill appear selected filter for may not be effect that user wants on present image, cause filtering effects use compared withDifference.

Summary of the invention

The present invention proposes a kind of method and apparatus for realizing smart filter, for solving filter using effect in the prior artBad technical problem.

The present invention proposes a kind of method for realizing smart filter, the specific steps are as follows:

Extract the primitive character information of original image；

Filter is used to original image, obtains the optimization image that user saves, and extract the characteristic information of filter；

Primitive character information, filter characteristic information and optimization image are stored in feature database in a manner of one-to-one, structureBuild user preference feature database；

User is instructed to the use of filter, to realize smart filter next time using user preference feature database.

The primitive character information includes picture contrast, image saturation, image sharpness, color of image histogram, figureAs one of display foreground feature, image background feature, image category or a variety of.

Described image classification includes landscape, personage, cuisines.

The characteristic information for extracting filter includes: filter type and each parameter value of filter, wherein each parameter value of filter withAdjustment percentage is saved.

It further include the favorite image information of user's addition in the user preference feature database, to the favorite image informationAnalyzed, obtain the characteristic information for liking image, as filter characteristic information, like the corresponding filter characteristic information of image withThe actual numerical value of characteristics of image is saved.

It is described to instruct user to the use of filter, to realize smart filter next time using user preference feature database, comprising:

If preserving user preference feature database in detection system, and the data volume of feature database reaches certain threshold value, according toFamily selects " intelligent recommendation " still " liking image effect " to carry out different processing: if user selects " intelligent recommendation ", treatingWhen processing image is handled, the image feature information of image to be processed is extracted, these information and the original being stored in feature databaseBeginning characteristic information is corresponding, and image feature information is compared with the primitive character information in feature database, extracts and characteristics of imageThe most like primitive character information of information, and the corresponding filter characteristic information of the primitive character information is applied to image to be processedOn；If user selects " liking image effect ", listing the favorite image that user is saved will like according to the user's choiceLike that the filter characteristic information of image is applied to image to be processed；

The favorite image listing user and being saved, including the favorite image and image to be processed according to preservationCharacteristics of image is ranked up the display of the favorite image of preservation, wherein characteristics of image includes display foreground feature, imageOne of background characteristics, image category are a variety of；

If not reaching certain threshold value without saving the data volume of user preference feature database or feature database in detection system,Then prompt user currently without associated recommendation.

It is described to extract the primitive character information most like with image feature information and the primitive character information is correspondingFilter characteristic information is applied on image to be processed, further can include: extracts and most like first three of image feature informationA primitive character information, and show user to select the corresponding optimization image of these three primitive character information, according toThe corresponding filter characteristic information of the selected optimization image in family is applied on image to be processed.

The present invention proposes a kind of device for realizing smart filter, specifically includes following module:

Primitive character information extraction modules, for extracting the primitive character information of original image；

Filter characteristic information extracting module, for, using filter, obtaining the optimization image that user saves to original image, andExtract the characteristic information of filter；

Preference profiles library constructs module, for primitive character information, filter characteristic information and optimization image is a pair of with oneThe mode answered is stored in feature database, constructs user preference feature database；

Smart filter realizes module, real for instructing user next time to the use of filter using user preference feature databaseExisting smart filter.

The primitive character information includes picture contrast, image saturation, image sharpness, color of image histogram, figureAs one of display foreground feature, image background feature, image category or a variety of.

Described image classification includes landscape, personage, cuisines.

The characteristic information for extracting filter includes: filter type and each parameter value of filter, wherein each parameter value of filter withAdjustment percentage is saved.

It further include the favorite image information of user's addition in the user preference feature database, to the favorite image informationAnalyzed, obtain the characteristic information for liking image, as filter characteristic information, like the corresponding filter characteristic information of image withThe actual numerical value of characteristics of image is saved.

The smart filter realizes module, is specifically used for:

If preserving user preference feature database in detection system, and the data volume of feature database reaches certain threshold value, according toFamily selects " intelligent recommendation " still " liking image effect " to carry out different processing: if user selects " intelligent recommendation ", treatingWhen processing image is handled, the image feature information of image to be processed is extracted, these information and the original being stored in feature databaseBeginning characteristic information is corresponding, and image feature information is compared with the primitive character information in feature database, extracts and characteristics of imageThe most like primitive character information of information, and the corresponding filter characteristic information of the primitive character information is applied to image to be processedOn；If user selects " liking image effect ", listing the favorite image that user is saved will like according to the user's choiceLike that the filter characteristic information of image is applied to image to be processed；

The favorite image listing user and being saved, including the favorite image and image to be processed according to preservationCharacteristics of image is ranked up the display of the favorite image of preservation, wherein characteristics of image includes display foreground feature, imageOne of background characteristics, image category are a variety of；

If not reaching certain threshold value without saving the data volume of user preference feature database or feature database in detection system,Then prompt user currently without associated recommendation.

It is described to extract the primitive character information most like with image feature information and the primitive character information is correspondingFilter characteristic information is applied on image to be processed, further can include: extracts and most like first three of image feature informationA primitive character information, and show user to select the corresponding optimization image of these three primitive character information, according toThe corresponding filter characteristic information of the selected optimization image in family is applied on image to be processed.

The present invention is based on a kind of method and apparatus for realizing smart filter of proposition, solve in the prior art to filterMachinery uses, and the intelligent recommendation and intelligent use of filter is realized, so that filter using effect is more preferably.

Detailed description of the invention

Fig. 1 is flow chart of the method for the present invention.

Fig. 2 is the device of the invention structure chart.

Specific embodiment

With reference to the accompanying drawing, it elaborates to embodiment.

Flow chart of the method for the present invention as shown in Figure 1:

The present invention proposes a kind of method for realizing smart filter, the specific steps are as follows:

Extract the primitive character information of original image；

Filter is used to original image, obtains the optimization image that user saves, and extract the characteristic information of filter；

Primitive character information, filter characteristic information and optimization image are stored in feature database in a manner of one-to-one, structureBuild user preference feature database；

User is instructed to the use of filter, to realize smart filter next time using user preference feature database.

The primitive character information includes picture contrast, image saturation, image sharpness, color of image histogram, figureAs one of foreground features, image background feature, image category or a variety of.

Described image classification includes landscape, personage, cuisines.

The characteristic information for extracting filter includes: filter type and each parameter value of filter, wherein each parameter value of filter withAdjustment percentage is saved.For example, to the saturation boost of image 10%, then saturation degree is saved as ,+10%.

It further include the favorite image information of user's addition in the user preference feature database, to the favorite image informationAnalyzed, obtain the characteristic information for liking image, as filter characteristic information, like the corresponding filter characteristic information of image withThe actual numerical value of characteristics of image is saved.User favorite image can be from network after see in circle of friends think to compareThe image preferred, the image are common images, do not save relevant filter information.

It is described to instruct user to the use of filter, to realize smart filter next time using user preference feature database, comprising:

If preserving user preference feature database in detection system, and the data volume of feature database reaches certain threshold value, according toFamily selects " intelligent recommendation " still " liking image effect " to carry out different processing: if user selects " intelligent recommendation ", treatingWhen processing image is handled, the image feature information of image to be processed is extracted, these information and the original being stored in feature databaseBeginning characteristic information is corresponding, and image feature information is compared with the primitive character information in feature database, during the comparison process, optionalSelect picture contrast, image saturation, image sharpness, color of image histogram, display foreground feature, image background feature, figureAs one of classification or it is a variety of be weighted comparison, for example, special using picture contrast, display foreground feature and image backgroundSign is when comparing, and the contrast difference between movement images assigns score value W1 with the difference, movement images prospect or background are specialBefore sign, such as similitude between display foreground or background object profile, display foreground or background object color similarity, imageScape or background object classification similitude etc. weight imparting prospect score value W2, background score value W3 according to these similitudes, then willW1, W2, W3 are weighted.Above-mentioned specific example is example, and those skilled in the art, which can according to need, suitably to be deformed.

Extract the primitive character information most like with image feature information, and by the corresponding filter of primitive character informationCharacteristic information is applied on image to be processed；If user select " liking image effect ", list user saved it is favoriteImage will like the filter characteristic information of image to be applied to image to be processed according to the user's choice；

The favorite image listing user and being saved, including the favorite image and image to be processed according to preservationCharacteristics of image is ranked up the display of the favorite image of preservation, wherein characteristics of image includes display foreground feature, imageOne of background characteristics, image category are a variety of.For example, being compared using display foreground feature and image background featureWhen, similitude, display foreground or back between movement images prospect or background characteristics, such as display foreground or background object profileScape object color similitude, display foreground or background object classification similitude etc. weight imparting prospect point according to these similitudesW2, W3, are then weighted by value W2, background score value W3.Above-mentioned specific example is example, and those skilled in the art can rootIt is suitably deformed according to needs.

If not reaching certain threshold value without saving the data volume of user preference feature database or feature database in detection system,Then prompt user currently without associated recommendation.

It is described to extract the primitive character information most like with image feature information and the primitive character information is correspondingFilter characteristic information is applied on image to be processed, further can include: extracts and most like first three of image feature informationA primitive character information, and show user to select the corresponding optimization image of these three primitive character information, according toThe corresponding filter characteristic information of the selected optimization image in family is applied on image to be processed.

The device of the invention structure chart as shown in Figure 2, specifically includes following module:

Primitive character information extraction modules, for extracting the primitive character information of original image；

Filter characteristic information extracting module, for, using filter, obtaining the optimization image that user saves to original image, andExtract the characteristic information of filter；

Preference profiles library constructs module, for primitive character information, filter characteristic information and optimization image is a pair of with oneThe mode answered is stored in feature database, constructs user preference feature database；

Smart filter realizes module, real for instructing user next time to the use of filter using user preference feature databaseExisting smart filter.

The primitive character information includes picture contrast, image saturation, image sharpness, color of image histogram, figureAs one of display foreground feature, image background feature, image category or a variety of.

Described image classification includes landscape, personage, cuisines.

The characteristic information for extracting filter includes: filter type and each parameter value of filter, wherein each parameter value of filter withAdjustment percentage is saved.

It further include the favorite image information of user's addition in the user preference feature database, to the favorite image informationAnalyzed, obtain the characteristic information for liking image, as filter characteristic information, like the corresponding filter characteristic information of image withThe actual numerical value of characteristics of image is saved.

The smart filter realizes module, is specifically used for:

If preserving user preference feature database in detection system, and the data volume of feature database reaches certain threshold value, according toFamily selects " intelligent recommendation " still " liking image effect " to carry out different processing: if user selects " intelligent recommendation ", treatingWhen processing image is handled, the image feature information of image to be processed is extracted, these information and the original being stored in feature databaseBeginning characteristic information is corresponding, and image feature information is compared with the primitive character information in feature database, extracts and characteristics of imageThe most like primitive character information of information, and the corresponding filter characteristic information of the primitive character information is applied to image to be processedOn；If user selects " liking image effect ", listing the favorite image that user is saved will like according to the user's choiceLike that the filter characteristic information of image is applied to image to be processed；

The favorite image listing user and being saved, including the favorite image and image to be processed according to preservationCharacteristics of image is ranked up the display of the favorite image of preservation, wherein characteristics of image includes display foreground feature, imageOne of background characteristics, image category are a variety of；

If not reaching certain threshold value without saving the data volume of user preference feature database or feature database in detection system,Then prompt user currently without associated recommendation.

It is described to extract the primitive character information most like with image feature information and the primitive character information is correspondingFilter characteristic information is applied on image to be processed, further can include: extracts and most like first three of image feature informationA primitive character information, and show user to select the corresponding optimization image of these three primitive character information, according toThe corresponding filter characteristic information of the selected optimization image in family is applied on image to be processed.

Above-described embodiment is merely preferred embodiments of the present invention, but protection scope of the present invention is not limited toThis, anyone skilled in the art in the technical scope disclosed by the present invention, the variation that can readily occur in or replacesIt changes, should be covered by the protection scope of the present invention.Therefore, protection scope of the present invention should be with the protection model of claimSubject to enclosing.

Claims (10)

1. a kind of method for realizing smart filter, the specific steps are as follows:

Extract the primitive character information of original image；

Filter is used to original image, obtains the optimization image that user saves, and extract the characteristic information of filter；

Primitive character information, filter characteristic information and optimization image are stored in feature database in a manner of one-to-one, building is usedFamily preference profiles library；

User is instructed to the use of filter, to realize smart filter next time using user preference feature database.

2. the method as described in claim 1, which is characterized in that the primitive character information includes that picture contrast, image are fullWith one of degree, image sharpness, color of image histogram, image image foreground features, image background feature, image category orIt is a variety of；Described image classification includes landscape, personage, cuisines.

3. the method as described in claim 1, which is characterized in that it is described extract filter characteristic information include: filter type andEach parameter value of filter, wherein each parameter value of filter is saved with adjusting percentage.

4. method as claimed in claim 3, which is characterized in that further include the happiness of user's addition in the user preference feature databaseThe image information of love analyzes the favorite image information, obtains the characteristic information for liking image, believes as filter featureBreath, likes that the corresponding filter characteristic information of image is saved with the actual numerical value of characteristics of image.

5. method as claimed in claim 4, which is characterized in that described to instruct user next time right using user preference feature databaseSmart filter is realized in the use of filter, comprising:

If preserving user preference feature database in detection system, and the data volume of feature database reaches certain threshold value, selects according to userIt selects " intelligent recommendation " and still " liking image effect " carries out different processing: if user selects " intelligent recommendation ", to be processedWhen image is handled, the image feature information of image to be processed is extracted, these information and the original spy being stored in feature databaseReference breath corresponds to, and image feature information compares with the primitive character information in feature database, extracts and image feature informationMost like primitive character information, and the corresponding filter characteristic information of the primitive character information is applied on image to be processed；If user selects " liking image effect ", figure will be liked according to the user's choice by listing the favorite image that user is savedThe filter characteristic information of picture is applied to image to be processed；

The favorite image listing user and being saved, including according to the favorite image of preservation and the image of image to be processedFeature is ranked up the display of the favorite image of preservation, wherein characteristics of image includes display foreground feature, image backgroundOne of feature, image category are a variety of；

If not reaching certain threshold value without saving the data volume of user preference feature database or feature database in detection system, mentionShow user currently without associated recommendation；

It is described to extract the primitive character information most like with image feature information, and by the corresponding filter of primitive character informationCharacteristic information is applied on image to be processed, further can include: extracts first three original most like with image feature informationBeginning characteristic information, and show user to select the corresponding optimization image of these three primitive character information, according to user instituteThe corresponding filter characteristic information of the optimization image of selection is applied on image to be processed.

6. a kind of device for realizing smart filter, which is characterized in that specifically include following module:

Primitive character information extraction modules, for extracting the primitive character information of original image；

Filter characteristic information extracting module obtains the optimization image that user saves, and extract for using filter to original imageThe characteristic information of filter；

Preference profiles library constructs module, is used for primitive character information, filter characteristic information and optimization image with one-to-oneMode is stored in feature database, constructs user preference feature database；

Smart filter realizes module, for instructing user to the use of filter, to realize intelligence next time using user preference feature databaseIt can filter.

7. device as claimed in claim 6, which is characterized in that the primitive character information includes that picture contrast, image are fullWith one of degree, image sharpness, color of image histogram, image image foreground features, image background feature, image category orIt is a variety of；Described image classification includes landscape, personage, cuisines.

8. device as claimed in claim 6, which is characterized in that it is described extract filter characteristic information include: filter type andEach parameter value of filter, wherein each parameter value of filter is saved with adjusting percentage.

9. device as claimed in claim 8, which is characterized in that further include the happiness of user's addition in the user preference feature databaseThe image information of love analyzes the favorite image information, obtains the characteristic information for liking image, believes as filter featureBreath, likes that the corresponding filter characteristic information of image is saved with the actual numerical value of characteristics of image.

10. device as claimed in claim 9, which is characterized in that the smart filter realizes module, is specifically used for: if detectionUser preference feature database is preserved in system, and the data volume of feature database reaches certain threshold value, " intelligently pushed away according to user's selectionRecommend " the still different processing of " liking image effect " progress: if user selects " intelligent recommendation ", to image to be processedWhen reason, the image feature information of image to be processed is extracted, these information are corresponding with the primitive character information being stored in feature database,Image feature information is compared with the primitive character information in feature database, is extracted and most like original of image feature informationCharacteristic information, and the corresponding filter characteristic information of the primitive character information is applied on image to be processed；If user's selection " happinessLike image effect ", then the favorite image that user is saved is listed, according to the user's choice, will like the filter feature of imageInformation application is in image to be processed；

The favorite image listing user and being saved, including according to the favorite image of preservation and the image of image to be processedFeature is ranked up the display of the favorite image of preservation, wherein characteristics of image includes display foreground feature, image backgroundOne of feature, image category are a variety of；

If not reaching certain threshold value without saving the data volume of user preference feature database or feature database in detection system, mentionShow user currently without associated recommendation；

It is described to extract the primitive character information most like with image feature information, and by the corresponding filter of primitive character informationCharacteristic information is applied on image to be processed, further can include: extracts first three original most like with image feature informationBeginning characteristic information, and show user to select the corresponding optimization image of these three primitive character information, according to user instituteThe corresponding filter characteristic information of the optimization image of selection is applied on image to be processed.