CN107146198A - A kind of intelligent method of cutting out of photo and device - Google Patents
A kind of intelligent method of cutting out of photo and device Download PDFInfo
- Publication number
- CN107146198A CN107146198A CN201710256861.XA CN201710256861A CN107146198A CN 107146198 A CN107146198 A CN 107146198A CN 201710256861 A CN201710256861 A CN 201710256861A CN 107146198 A CN107146198 A CN 107146198A
- Authority
- CN
- China
- Prior art keywords
- sliding window
- aesthetic feeling
- aesthetics
- clipping region
- original scale
- 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
- 238000000034 method Methods 0.000 title claims abstract description 37
- 238000013210 evaluation model Methods 0.000 claims description 34
- 230000008569 process Effects 0.000 claims description 4
- 238000011156 evaluation Methods 0.000 abstract description 4
- 238000013527 convolutional neural network Methods 0.000 description 8
- 238000012549 training Methods 0.000 description 6
- 238000012545 processing Methods 0.000 description 3
- 230000006870 function Effects 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 239000011521 glass Substances 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 230000035800 maturation Effects 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 230000001537 neural effect Effects 0.000 description 1
- 238000012805 post-processing Methods 0.000 description 1
- 230000008439 repair process Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Classifications
-
- G06T3/04—
Landscapes
- Processing Or Creating Images (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of intelligent method of cutting out of photo and device.This method includes：Obtain smart machine photograph image；Some subregions are generated in photograph image, the aesthetics scoring of some subregions is calculated, optimal aesthetic feeling clipping region is obtained according to the aesthetics scoring；Photograph image is cut according to the coordinate of optimal aesthetic feeling clipping region.The intelligent method of cutting out of the photo of the embodiment of the present invention and device realize that the intelligence of photo is cut by image aesthetic evaluation technology so that the photo for more conforming to application scenarios is cut out from aesthstic angle；The applicable photo type of the present invention is more universal and easy to operate, efficiency high, so, the present invention effectively overcomes various shortcoming of the prior art and has high industrial utilization.
Description
Technical field
The present invention relates to technical field of image processing, the more particularly to a kind of intelligent method of cutting out of photo and device.
Background technology
As intelligent terminal is taken pictures the continuous enhancing of performance, people are keen to record the beautiful wink in life conveniently
Between, and share into social networks.But the photo that generally we have shot is not appropriate for directly being uploaded to internet, it is necessary to enter
Some post-processings of row.For example, finger imprudence has blocked lens edge when taking pictures or lens edge someone is strayed into；Merely desire to interception
Part picture in photo is used as highlighting；There is a very beautiful rectangle from taking pictures, and wechat head portrait requires it is square.
Above-mentioned to usually require to carry out photo cutting in the case of these, the current main flow photo that the application of figure class provides of taking pictures/repair cuts work(
Can be by freedom/fixed proportion manual editing's sectional drawing, but due to domestic consumer do not have generally composition etc. specialty photography
Knowledge, thus the photo aesthetic quality that cuts out manually it cannot be guaranteed that, in addition, cutting out the ratio of photo manually may not be inconsistent yet
Close application requirement.Application No. CN201410681308.7 invention carries out Face datection to the image of shooting and obtains Face datection
Frame；Using Face datection frame central point as origin, according to area and the area ratio 10 of Face datection frame：7 generation length-width ratios are 4：3
Clipping rectangle frame；Part beyond clipping rectangle frame described in image is cropped.Angle pair of the invention from Face datection
The faces such as certificate photo have better effects for the cutting of main photo, but can not realize intelligent sanction for substantial amounts of common photo
Cut.
The content of the invention
In view of the above problems, the invention provides a kind of intelligent method of cutting out of photo and device.
The intelligent method of cutting out of photo that the present invention is provided, comprises the following steps：
Obtain smart machine photograph image；
Some subregions are generated in the photograph image, the aesthetics scoring of some subregions are calculated, according to described
Aesthetics scoring obtains optimal aesthetic feeling clipping region；
The photograph image is cut according to the coordinate of the optimal aesthetic feeling clipping region.
Present invention also offers a kind of intelligent Scissoring device of photo, including：Image collection module, computing module, cutting mould
Block：
Described image acquisition module, for obtaining smart machine photograph image；
The computing module, for generating some subregions in the photograph image, calculates some subregions
Aesthetics scoring, optimal aesthetic feeling clipping region is obtained according to the aesthetics scoring；
The cutting module, for cutting the photograph image according to the coordinate of the optimal aesthetic feeling clipping region.
The present invention has the beneficial effect that：
The intelligent method of cutting out of the photo of the embodiment of the present invention and device realize the intelligence of photo by image aesthetic evaluation technology
It can cut so that the photo for more conforming to application scenarios is cut out from aesthstic angle；The applicable photo type of the present invention is more general
Time, and easy to operate, efficiency high.So, the present invention effectively overcomes various shortcoming of the prior art and has high industrial profit
With value.
Brief description of the drawings
Fig. 1 is the flow chart of the intelligent method of cutting out of photo of the inventive method embodiment；
Fig. 2 is the structural representation of the intelligent Scissoring device of photo of apparatus of the present invention embodiment；
Fig. 3 is the flow chart of the intelligent method of cutting out of photo of present example 1；
Fig. 4 is the flow chart of the intelligent method of cutting out of photo of present example 2.
Embodiment
The exemplary embodiment of the disclosure is more fully described below with reference to accompanying drawings.Although showing the disclosure in accompanying drawing
Exemplary embodiment, it being understood, however, that may be realized in various forms the disclosure without should be by embodiments set forth here
Limited.On the contrary, these embodiments are provided to facilitate a more thoroughly understanding of the present invention, and can be by the scope of the present disclosure
Complete conveys to those skilled in the art.
In order to solve photo aesthetic quality that prior art cuts it cannot be guaranteed that or in addition to the photo based on face it is substantial amounts of
Common photo can not but realize the problem of intelligence is cut, the invention provides a kind of intelligent method of cutting out of photo and device, below
With reference to accompanying drawing and embodiment, the present invention will be described in further detail.It should be appreciated that specific embodiment described herein
Only to explain the present invention, the present invention is not limited.
The method according to the invention embodiment is there is provided a kind of intelligent method of cutting out of photo, and Fig. 1 is that the inventive method is implemented
The flow chart of the intelligent method of cutting out of the photo of example, as shown in figure 1, according to the intelligent method of cutting out of the photo of the inventive method embodiment
Including following processing：
Step 101, smart machine photograph image is obtained.
Specifically, smart machine include PC, smart mobile phone, tablet personal computer, intelligent glasses, and other have take the photograph
As the equipment of head and data processing function.
Step 102, some subregions are generated in the photograph image, the aesthetics scoring of some subregions is calculated,
Optimal aesthetic feeling clipping region is obtained according to the aesthetics scoring.Specifically, some subregions are generated in photograph image includes use
Default sliding window travels through the photograph image, obtains some sub-regions.
Wherein, step 102 includes following two technical schemes.The first technical scheme comprises the following steps：Use first
Sliding window travels through the photograph image, and first slip during traversal is obtained using default aesthetic evaluation model
Several corresponding aesthetics scorings of window, choose highest aesthetics scoring, and most Jiamei is used as using the corresponding position of highest aesthetics scoring
Feel clipping region；Wherein, the one or more during the selection of the first sliding window is following：Original scale sliding window, standard proportional
Sliding window, preset ratio sliding window.
More specific, it is described to use when first sliding window includes the preset ratio sliding window
One sliding window travels through the photograph image, and first cunning during traversal is obtained using default aesthetic evaluation model
Several the corresponding aesthetics scorings of dynamic window, comprise the following steps：
The photograph image is traveled through using preset ratio sliding window, is obtained using default aesthetic evaluation model in traversal
During first sliding window it is corresponding several aesthetics scoring.
More specific, when first sliding window includes original scale sliding window and standard proportional sliding window
When, the photograph image is traveled through using the first sliding window, is obtained using default aesthetic evaluation model during traversal
Several corresponding aesthetics scorings of first sliding window, choose highest aesthetics scoring, including following scheme A and option b.
Specifically, option A comprises the following steps：
The photograph image is traveled through using original scale sliding window, is obtained using default aesthetic evaluation model in traversal
During the original scale sliding window it is corresponding several aesthetics scoring；
The photograph image is traveled through using standard proportional sliding window, is obtained using default aesthetic evaluation model in traversal
During the standard proportional sliding window it is corresponding several aesthetics scoring；
In several corresponding aesthetics scorings of the original scale sliding window and standard proportional sliding window correspondence
Several aesthetics scoring in choose highest aesthetics scoring.
Specifically, option b comprises the following steps：
The photograph image is traveled through using original scale sliding window, is obtained using default aesthetic evaluation model in traversal
During the original scale sliding window it is corresponding several aesthetics scoring；
Highest aesthetics scoring in several the corresponding aesthetics scorings of original scale sliding window is chosen, it is beautiful according to highest
The coordinate of the corresponding original scale sliding window of scoring obtains the optimal aesthetic feeling clipping region under original scale；
The standard proportional sliding window is calculated to hand over simultaneously with the maximum of the optimal aesthetic feeling clipping region under the original scale
Than the position that will produce maximum friendship and ratio is used as the optimal aesthetic feeling clipping region under standard proportional；
According to the optimal aesthetic feeling clipping region under instruction selection original scale or the optimal aesthetic feeling Crop Area under standard proportional
One in domain is used as the optimal aesthetic feeling clipping region of image.
More specific, in option A and option b, original scale sliding window and standard proportional sliding window can select
For several yardsticks.For example when original scale sliding window is from 3 yardsticks, the pantograph ratio of each yardstick is 1.2, and smallest dimension is slided
When the horizontal direction length of side of window is the 1/4 of the long side of original image horizontal direction, that is, represent from the original of following three kinds of yardsticks
Scaled image：Size is 0.25 times of original image size, size is original image size 0.25 × 1.2 times, size be original
0.25 × 1.2 × 1.2 times of beginning image size.
Second of technical scheme of step 102 comprises the following steps：The photograph image is traveled through using the second sliding window,
Several corresponding aesthetics scorings of second sliding window during traversal are obtained, are made with several described aesthetics scorings
For the aesthetic quality value of several the second sliding window center point coordinates in ergodic process, generated according to several aesthetic quality values
Aesthetic quality collection of illustrative plates；Wherein, second sliding window is photograph image original scale, and the second sliding window is less than or equal to
First sliding window, the second sliding window step-length is less than or equal to the first sliding window；
The energy highest zone in the aesthetic quality collection of illustrative plates is obtained using default clustering algorithm, with the energy highest
The minimum enclosed rectangle in region is used as the optimal aesthetic feeling region of image；
The maximum in the window of cutting ratio needed for calculating and the optimal aesthetic feeling region of described image is handed over and compared, and will produce maximum hand over
And the position of ratio is used as optimal aesthetic feeling clipping region.
Specifically, when first sliding window includes original scale sliding window and standard proportional sliding window, meter
Calculate the maximum friendship of the first sliding window and the optimal aesthetic feeling region of described image and compare, the position of maximum friendship and ratio will be produced as most
Clipping region is felt by Jiamei, comprises the following steps：
Calculate the maximum friendship of original scale sliding window and the optimal aesthetic feeling region of described image and compare, maximum hand over simultaneously will be produced
The position of ratio is used as the optimal aesthetic feeling clipping region under original scale；
Calculate the maximum friendship of standard proportional sliding window and the optimal aesthetic feeling region of described image and compare, maximum hand over simultaneously will be produced
The position of ratio is used as the optimal aesthetic feeling clipping region under standard proportional；
Selected in the optimal aesthetic feeling clipping region under optimal aesthetic feeling clipping region and standard proportional under the original scale
Nearest one of center and the optimal aesthetic feeling regional center of described image is taken as the optimal aesthetic feeling clipping region of image.
It is further comprising the steps of as another embodiment of the inventive method embodiment：Obtain the target Crop Area specified
Domain seed point；Further, it is described to travel through the photograph image using default sliding window, obtain the institute during traversal
Several the corresponding aesthetics scorings of default sliding window are stated, are comprised the following steps：
The target clipping region seed vertex neighborhood is traveled through using default sliding window, the institute during traversal is obtained
State several the corresponding aesthetics scorings of default sliding window.
Step 103, the photograph image is cut according to the coordinate of the optimal aesthetic feeling clipping region, obtains the photo figure
The optimum clipped region of picture.
The intelligent method of cutting out of the photo of the embodiment of the present invention realizes that the intelligence of photo is cut by image aesthetic evaluation technology,
So that cutting out the photo for more conforming to application scenarios from aesthstic angle；The applicable photo type of the present invention is more universal, and behaviour
Make convenient, efficiency high.So, the present invention effectively overcomes various shortcoming of the prior art and has high industrial utilization.
Embodiment of the method with the present invention is corresponding, and present invention also offers a kind of intelligent Scissoring device of photo, Fig. 2 is this
The structural representation of the intelligent Scissoring device of the photo of invention device embodiment, as shown in Fig. 2 according to apparatus of the present invention embodiment
Photo intelligence Scissoring device includes：Image collection module 20, computing module 22, cutting module 24, below to the embodiment of the present invention
Modules be described in detail.
Specifically, described image acquisition module 20, for obtaining smart machine photograph image.
The computing module 22, for generating some subregions in the photograph image, calculates some subregions
Aesthetics scoring, according to it is described aesthetics scoring obtain optimal aesthetic feeling clipping region.
Specifically, some subregions are generated in photograph image to be included traveling through the photo figure with default sliding window
Picture, obtains some sub-regions.
The computing module specifically for：The photograph image is traveled through using the first sliding window, default aesthetics is utilized
Evaluation model obtains several corresponding aesthetics scorings of first sliding window during traversal, chooses highest aesthetics and comments
Point, optimal aesthetic feeling clipping region is used as using the corresponding position of highest aesthetics scoring；Wherein, during the selection of the first sliding window is following
It is one or more of：Original scale sliding window, standard proportional sliding window, preset ratio sliding window；
Or, for traveling through the photograph image using the second sliding window, obtain described second during traversal
Several corresponding aesthetics scorings of sliding window, several second slips in ergodic process are used as using several described aesthetics scorings
The aesthetic quality value of window center point coordinates, aesthetic quality collection of illustrative plates is generated according to several aesthetic quality values；Wherein, described second
Sliding window is photograph image original scale, and the second sliding window is less than or equal to the first sliding window, the second sliding window
Step-length is less than or equal to the first sliding window；
The energy highest zone in the aesthetic quality collection of illustrative plates is obtained using default clustering algorithm, with the energy highest
The minimum enclosed rectangle in region is used as the optimal aesthetic feeling region of image；
The maximum in the window of cutting ratio needed for calculating and the optimal aesthetic feeling region of described image is handed over and compared, and will produce maximum hand over
And the position of ratio is used as optimal aesthetic feeling clipping region.
More specific, when first sliding window includes original scale sliding window and standard proportional sliding window
When, the computing module specifically for：
The photograph image is traveled through using original scale sliding window, is obtained using default aesthetic evaluation model in traversal
During the original scale sliding window it is corresponding several aesthetics scoring；
The photograph image is traveled through using standard proportional sliding window, is obtained using default aesthetic evaluation model in traversal
During the standard proportional sliding window it is corresponding several aesthetics scoring；
In several corresponding aesthetics scorings of the original scale sliding window and standard proportional sliding window correspondence
Several aesthetics scoring in choose highest aesthetics scoring；
Or the photograph image is traveled through using original scale sliding window, obtained using default aesthetic evaluation model
Several corresponding aesthetics scorings of the original scale sliding window during traversal；
Highest aesthetics scoring in several the corresponding aesthetics scorings of original scale sliding window is chosen, it is beautiful according to highest
The coordinate of the corresponding original scale sliding window of scoring obtains the optimal aesthetic feeling clipping region under original scale；
The standard proportional sliding window is calculated to hand over simultaneously with the maximum of the optimal aesthetic feeling clipping region under the original scale
Than the position that will produce maximum friendship and ratio is used as the optimal aesthetic feeling clipping region under standard proportional；
According to the optimal aesthetic feeling clipping region under instruction selection original scale or the optimal aesthetic feeling Crop Area under standard proportional
One in domain is used as the optimal aesthetic feeling clipping region of image；
Or calculate the maximum friendship of the first sliding window and the optimal aesthetic feeling region of described image and compare, maximum hand over simultaneously will be produced
The position of ratio comprises the following steps as optimal aesthetic feeling clipping region：
Calculate the maximum friendship of original scale sliding window and the optimal aesthetic feeling region of described image and compare, maximum hand over simultaneously will be produced
The position of ratio is used as the optimal aesthetic feeling clipping region under original scale；
Calculate the maximum friendship of standard proportional sliding window and the optimal aesthetic feeling region of described image and compare, maximum hand over simultaneously will be produced
The position of ratio is used as the optimal aesthetic feeling clipping region under standard proportional；
Selected in the optimal aesthetic feeling clipping region under optimal aesthetic feeling clipping region and standard proportional under the original scale
Nearest one of center and the optimal aesthetic feeling regional center of described image is taken as the optimal aesthetic feeling clipping region of image.
The cutting module 24, for cutting the photograph image according to the coordinate of the optimal aesthetic feeling clipping region, is obtained
To the optimum clipped region of the photograph image.
The default aesthetic evaluation model is obtained by training, and the machine learning method used includes convolutional Neural net
Network, limited Boltzmann machine, depth confidence network etc..Further, apparatus of the present invention embodiment also includes model training mould
Block, the model training module is used for：
Obtain several images of known aesthetics scoring；
Default aesthetic evaluation model is trained using several images of known aesthetics scoring, after being trained
Aesthetic evaluation model.
The intelligent method of cutting out of the photo of the present invention and system, have the advantages that：
1st, the present invention can cut out the photo of standard proportional from optimal aesthstic angle intelligent, solve manually
The off-gauge situation of photographic scale that photo is likely to occur low aesthetic feeling and cut out is cut, to meet practical application request.
2nd, interactive function is supported, optimal aesthstic photo can be cut out in the range of neighborhood of a point being manually specified, with more
The customization for meeting user well cuts demand.
For more detailed explanation embodiments of the invention, example 1 and example 2 are provided.
As shown in figure 3, example 1 comprises the following steps：
Step S1：Photograph image is obtained from smart machine photograph album.
Step S2：System judges whether to be manually specified a certain coordinate in image as target clipping region seed point.
Step S3：If the target clipping region seed point not being manually specified, uses multiple dimensioned original scale and standard
Ratio (1:1、5:3、4:3、5:4、16:9、3:2、7:5) sliding window traversing graph picture, and use convolutional neural networks aesthetic evaluation
Model calculates the aesthetics scoring obtained in each window.The highest aesthetics scoring for choosing original scale and standard proportional window is corresponding
Position, is used as optimal aesthetic feeling clipping region under original scale and standard proportional；
If in the presence of the target clipping region seed point being manually specified, using multiple dimensioned original scale and standard proportional (1:1、
5:3、4:3、5:4、16:9、3:2、7:5) sliding window traversal seed vertex neighborhood, it is ensured that each sliding window includes seed point, and makes
The aesthetics scoring obtained in each window is calculated with convolutional neural networks aesthetic evaluation model.Choose original scale and standard proportional window
The corresponding position of highest aesthetics scoring of mouth, is used as optimal aesthetic feeling clipping region under original scale and standard proportional.
Step S4：Smart machine is cut according to the coordinate of optimal aesthetic feeling clipping region under original scale and standard proportional respectively
Original image, obtains original scale and standard proportional hypograph optimum clipped region.Many scaled images cut result and selected for user
Select.If optimal aesthetic feeling clipping region coordinate is coordinates of original image coordinates, without cutting.
In step sl, photograph image is obtained from smart machine photograph album.
In step s 2, system judges whether to be manually specified a certain coordinate in image as target clipping region seed
Point.
In step s3, if a certain coordinate not being manually specified in image is used as target clipping region seed point
Multiple dimensioned original scale and standard proportional (1:1、5:3、4:3、5:4、16:9、3:2、7:5) sliding window is with a fixed step size time
Image is gone through, and the aesthetics scoring (1-10 points) obtained in each window is calculated using convolutional neural networks aesthetic evaluation model.Choose
Original scale and the corresponding position of the highest of standard proportional window aesthetics scoring, are used as original scale and standard proportional Xia Jiameis
Feel clipping region.
It should be noted that aesthetic evaluation model is obtained by training, it is to avoid the subjectivity and piece of artificial rule model
Face property.Detect original scale and multiple standards ratio (1:1、5:3、4:3、16:9、16:9、3:2、7:5) optimal aesthetic feeling is cut under
During region, remove outside using the sliding window traversing graph picture under all proportions, also only can once be schemed with original scale window traversal
As obtaining optimal aesthetic feeling clipping region under original scale, multiple standards ratio window and optimal aesthetic feeling under original scale are then calculated
The maximum of clipping region is handed over and compared, and the position that will produce maximum friendship and ratio is used as the optimal aesthetic feeling clipping region under standard proportional.
Preferably, sliding window totally 7 yardsticks, each scaling ratio is 1.2, the horizontal direction of smallest dimension sliding window
The length of side is set to the 1/4 of the long side of original image horizontal direction, and the horizontal direction step-length of sliding window is set to original image horizontal direction
The 1/16 of the length of side, the vertical direction step-length of sliding window is set to the 1/16 of the original image vertical direction length of side.
In the step S3, if a certain coordinate being manually specified in image is used as target clipping region seed point
Multiple dimensioned original scale and standard proportional (1:1、5:3、4:3、5:4、16:9、3:2、7:5) sliding window traversal seed vertex neighborhood,
And ensure that each sliding window includes seed point.The aesthetics obtained in each window is calculated using convolutional neural networks aesthetic evaluation model
Scoring (1-10 points), chooses the corresponding position of the highest aesthetics scoring of original scale and standard proportional window, is used as original scale
And optimal aesthetic feeling clipping region under standard proportional.
It should be noted that aesthetic evaluation model is obtained by training, it is to avoid the subjectivity and piece of artificial rule model
Face property.Detect original scale and multiple standards ratio (1:1、5:3、4:3、16:9、16:9、3:2、7:5) optimal aesthetic feeling is cut under
During region, remove outside using the sliding window traversing graph picture under all proportions, also only can once be schemed with original scale window traversal
As obtaining optimal aesthetic feeling clipping region under original scale, multiple standards ratio window and optimal aesthetic feeling under original scale are then calculated
The maximum of clipping region is handed over and compared, and the position that will produce maximum friendship and ratio is used as the optimal aesthetic feeling clipping region under standard proportional.
Preferably, sliding window totally 7 yardsticks, each scaling ratio is 1.2, the horizontal direction of smallest dimension sliding window
The length of side is set to the 1/4 of the long side of original image horizontal direction, and the horizontal direction step-length of sliding window is set to original image horizontal direction
The 1/16 of the length of side, the vertical direction step-length of sliding window is set to the 1/16 of the original image vertical direction length of side.Preferably, in seed
Object recognition algorithm is used in vertex neighborhood, it is ensured that seed vertex neighborhood includes the profile of seed point object.Have in the prior art
Many ripe object recognition algorithms, therefore will not be repeated here.
In the step S4, smart machine is respectively according to optimal aesthetic feeling clipping region under original scale and standard proportional
Coordinate cuts original image, obtains original scale and standard proportional hypograph optimum clipped region.Many scaled images cut result
For user's selection.If optimal aesthetic feeling clipping region coordinate is coordinates of original image coordinates, without cutting.
It should be noted that cutting ratio needed for can also being specified by user, this method and system will be specified according to user and cut out
Cut proportional cutting and go out optimized image region.
As shown in figure 4, example 2 comprises the following steps：
Step S1：Photograph image is obtained from smart machine photograph album.
Step S2：The aesthetic quality collection of illustrative plates of photograph image is generated using convolutional neural networks aesthetic evaluation model.
Step S3：System judges whether to be manually specified a certain coordinate in image as target clipping region seed point.
Step S4：If a certain coordinate not being manually specified in image uses cluster as target clipping region seed point
Algorithm obtains energy highest zone in aesthetic quality collection of illustrative plates, obtains the coordinate of the minimum enclosed rectangle of energy highest zone, and its is right
Answer the optimal aesthetic feeling area coordinate of image.Calculate respectively and obtain multiple dimensioned original scale and standard proportional (1:1、5:3、4:3、5:4、
16:9、3:2、7:5) maximum friendship and the coordinate position of ratio of the window with the optimal aesthetic feeling region of image, are used as original scale and standard
Optimal aesthetic feeling clipping region under ratio.
If a certain coordinate being manually specified in image obtains seed as target clipping region seed point using clustering algorithm
Energy highest zone in aesthetic quality collection of illustrative plates in vertex neighborhood, and ensure that energy highest zone scope includes seed point, obtain energy
The coordinate of the minimum enclosed rectangle of highest zone, the optimal aesthetic feeling area coordinate of its correspondence image.Calculate respectively and obtain multiple dimensioned original
Beginning ratio and standard proportional (1:1、5:3、4:3、5:4、16:9、3:2、7:5) window and the maximum friendship in the optimal aesthetic feeling region of image
And the coordinate position of ratio, it is used as optimal aesthetic feeling clipping region under original scale and standard proportional.
Step S5：Smart machine is cut according to the coordinate of optimal aesthetic feeling clipping region under original scale and standard proportional respectively
Original image, obtains original scale and standard proportional hypograph optimum clipped region.Many scaled images cut result and selected for user
Select.If optimal aesthetic feeling clipping region coordinate is coordinates of original image coordinates, without cutting.
In step sl, photograph image is obtained from smart machine photograph album.
In step s 2, the aesthetic quality collection of illustrative plates of photograph image is generated using convolutional neural networks aesthetic evaluation model.With
The sliding window traversing graph picture of a certain size and step-length, uses convolutional neural networks aesthetic evaluation model to calculate each sliding window
Aesthetics scoring (1-10 point) as the aesthetic quality value of sliding window center point coordinate, complete, that is, give birth to as sliding window is traveled through
Into aesthetic quality collection of illustrative plates.
It should be noted that aesthetic evaluation model is obtained by training, it is to avoid the subjectivity and piece of artificial rule model
Face property.
Preferably, the length-width ratio of sliding window is identical with original image, and the horizontal direction length of side of sliding window is set to original
The 1/4 of the long side in image level direction, the vertical direction length of side of sliding window is set to the 1/4 of the long side of original image vertical direction, slides
The horizontal direction step-length of dynamic window is set to the 1/32 of the original image horizontal direction length of side, and the vertical direction step-length of sliding window is set to
The 1/32 of the original image vertical direction length of side.
In step s3, system judges whether to be manually specified a certain coordinate in image as target clipping region seed
Point.
In step s 4, if a certain coordinate not being manually specified in image is used as target clipping region seed point
Clustering algorithm obtains energy highest zone in aesthetic quality collection of illustrative plates, obtains the coordinate of the minimum enclosed rectangle of energy highest zone,
The optimal aesthetic feeling area coordinate of its correspondence image.Calculate respectively and obtain multiple dimensioned original scale and standard proportional (1:1、5:3、4:3、
5:4、16:9、3:2、7:5) window and the optimal aesthetic feeling region of image it is maximum hand over and ratio coordinate position, as original scale and
Optimal aesthetic feeling clipping region under standard proportional.
It should be noted that when multiple dimensioned original scale and standard proportional window are produced most with the optimal aesthetic feeling region of image
When the coordinate position of big friendship and ratio is not unique, selected window center is with the nearest window of the optimal aesthetic feeling regional center of image as most
Clipping region is felt by Jiamei.
In step s 4, if a certain coordinate being manually specified in image uses cluster as target clipping region seed point
Algorithm obtains in seed vertex neighborhood energy highest zone in aesthetic quality collection of illustrative plates, and ensures that energy highest zone scope includes seed
Point, obtains the coordinate of the minimum enclosed rectangle of energy highest zone, the optimal aesthetic feeling area coordinate of its correspondence image.Calculate respectively
To multiple dimensioned original scale and standard proportional (1:1、5:3、4:3、5:4、16:9、3:2、7:5) window and the optimal aesthetic feeling area of image
The maximum friendship in domain and the coordinate position of ratio, are used as optimal aesthetic feeling clipping region under original scale and standard proportional.
It should be noted that when multiple dimensioned original scale and standard proportional window are produced most with the optimal aesthetic feeling region of image
When the coordinate position of big friendship and ratio is not unique, selected window center is with the nearest window of the optimal aesthetic feeling regional center of image as most
Clipping region is felt by Jiamei.Preferably, object recognition algorithm is used in seed vertex neighborhood, it is ensured that seed vertex neighborhood includes seed point
The profile of object.There is the object recognition algorithm of many maturations in the prior art, therefore will not be repeated here.
In the step S5, smart machine is respectively according to optimal aesthetic feeling clipping region under original scale and standard proportional
Coordinate cuts original image, obtains original scale and standard proportional hypograph optimum clipped region.Many scaled images cut result
For user's selection.If optimal aesthetic feeling clipping region coordinate is coordinates of original image coordinates, without cutting.
It should be noted that cutting ratio needed for can also being specified by user, this method and system will be specified according to user and cut out
Cut proportional cutting and go out optimized image region.
The present invention can automatic cutting cut most U.S. region in photo, and provide the mode of operation of user mutual, can cut out
Most U.S. photo near specified location in user, while carrying out scene classification to photograph album photo in advance, lifts aesthetic evaluation accuracy, adopts
With convolutional neural networks aesthetic evaluation model, it is to avoid the subjectivity and one-sidedness of artificial rule feature.This method and system can be with
Applicable scope is wide, can be generally applicable in smart machine.
Embodiments of the invention are the foregoing is only, are not intended to limit the invention, for those skilled in the art
For member, the present invention can have various modifications and variations.Any modification within the spirit and principles of the invention, being made,
Equivalent substitution, improvement etc., should be included within scope of the presently claimed invention.
Claims (10)
1. the intelligent method of cutting out of a kind of photo, it is characterised in that comprise the following steps：
Obtain smart machine photograph image；
Some subregions are generated in the photograph image, the aesthetics scoring of some subregions are calculated, according to the aesthetics
Scoring obtains optimal aesthetic feeling clipping region；
The photograph image is cut according to the coordinate of the optimal aesthetic feeling clipping region.
2. the intelligent method of cutting out of photo as claimed in claim 1, it is characterised in that if described generate in the photograph image
Dry subregion, calculates the aesthetics scoring of some subregions, obtains optimal aesthetic feeling clipping region according to the aesthetics scoring, wraps
Include following steps：
The photograph image is traveled through using the first sliding window, is obtained using default aesthetic evaluation model during traversal
Several corresponding aesthetics of first sliding window score, and choose highest aesthetics scoring, with the corresponding position of highest aesthetics scoring
Put as optimal aesthetic feeling clipping region；
Wherein, the one or more during the selection of the first sliding window is following：Original scale sliding window, standard proportional sliding window
Mouth, preset ratio sliding window.
3. the intelligent method of cutting out of photo as claimed in claim 1, it is characterised in that if described generate in the photograph image
Dry subregion, calculates the aesthetics scoring of some subregions, obtains optimal aesthetic feeling clipping region according to the aesthetics scoring, wraps
Include following steps：
The photograph image is traveled through using the second sliding window, second sliding window during traversal is obtained corresponding
Several aesthetics scorings, several the second sliding window center point coordinates in ergodic process are used as using several described aesthetics scorings
Aesthetic quality value, according to several aesthetic quality values generate aesthetic quality collection of illustrative plates；Wherein, second sliding window is photo
Image original scale, and the second sliding window is less than or equal to the first sliding window, the second sliding window step-length is less than or equal to
First sliding window；
The energy highest zone in the aesthetic quality collection of illustrative plates is obtained using default clustering algorithm, with the energy highest zone
Minimum enclosed rectangle be used as the optimal aesthetic feeling region of image；
Calculate the window of required cutting ratio and the maximum of the optimal aesthetic feeling region of described image is handed over and simultaneously compared, maximum friendship will be produced and compared
Position be used as optimal aesthetic feeling clipping region.
4. the intelligent method of cutting out of photo as described in any one of claims 1 to 3, it is characterised in that further comprising the steps of：Obtain
The fixed target clipping region seed point of fetching；
It is described to travel through the photograph image using default sliding window, obtain the default sliding window during traversal
Several corresponding aesthetics scorings of mouth, comprise the following steps：
The target clipping region seed vertex neighborhood is traveled through using default sliding window, obtains described pre- during traversal
If sliding window it is corresponding several aesthetics scoring.
5. the intelligent method of cutting out of photo as claimed in claim 2, it is characterised in that
When first sliding window includes original scale sliding window and standard proportional sliding window；
The photograph image is traveled through using the first sliding window, is obtained using default aesthetic evaluation model during traversal
Several corresponding aesthetics scorings of first sliding window, choose highest aesthetics scoring, comprise the following steps：
The photograph image is traveled through using original scale sliding window, the mistake in traversal is obtained using default aesthetic evaluation model
Several corresponding aesthetics scorings of original scale sliding window described in journey；
The photograph image is traveled through using standard proportional sliding window, the mistake in traversal is obtained using default aesthetic evaluation model
Several corresponding aesthetics scorings of standard proportional sliding window described in journey；
If corresponding in several corresponding aesthetics scorings of the original scale sliding window and the standard proportional sliding window
Highest aesthetics scoring is chosen in dry aesthetics scoring；
Or the photograph image is traveled through using original scale sliding window, obtained using default aesthetic evaluation model in traversal
During the original scale sliding window it is corresponding several aesthetics scoring；
Highest aesthetics scoring in several the corresponding aesthetics scorings of original scale sliding window is chosen, is commented according to highest aesthetics
The coordinate of point corresponding original scale sliding window obtains the optimal aesthetic feeling clipping region under original scale；
Calculate the maximum friendship of the standard proportional sliding window and the optimal aesthetic feeling clipping region under the original scale and compare, will
The position for producing maximum friendship and ratio is used as the optimal aesthetic feeling clipping region under standard proportional；
According in the optimal aesthetic feeling clipping region under instruction selection original scale or the optimal aesthetic feeling clipping region under standard proportional
One be used as the optimal aesthetic feeling clipping region of image.
6. the intelligent method of cutting out of photo as claimed in claim 3, it is characterised in that
When first sliding window includes original scale sliding window and standard proportional sliding window；
Calculate the maximum friendship of the first sliding window and the optimal aesthetic feeling region of described image and compare, the position of maximum friendship and ratio will be produced
As optimal aesthetic feeling clipping region, comprise the following steps：
Maximum hand over for calculating original scale sliding window and the optimal aesthetic feeling region of described image is simultaneously compared, and will produce maximum friendship and ratio
Position is used as the optimal aesthetic feeling clipping region under original scale；
Maximum hand over for calculating standard proportional sliding window and the optimal aesthetic feeling region of described image is simultaneously compared, and will produce maximum friendship and ratio
Position is used as the optimal aesthetic feeling clipping region under standard proportional；
In being chosen in the optimal aesthetic feeling clipping region under optimal aesthetic feeling clipping region and standard proportional under the original scale
Nearest one of the heart and the optimal aesthetic feeling regional center of described image is used as the optimal aesthetic feeling clipping region of image.
7. the intelligent Scissoring device of a kind of photo, it is characterised in that including image collection module, computing module, cut module：
Described image acquisition module, for obtaining smart machine photograph image；
The computing module, for generating some subregions in the photograph image, calculates the aesthetics of some subregions
Scoring, optimal aesthetic feeling clipping region is obtained according to the aesthetics scoring；
The cutting module, for cutting the photograph image according to the coordinate of the optimal aesthetic feeling clipping region.
8. the intelligent Scissoring device of photo as claimed in claim 7, it is characterised in that
The computing module, for traveling through the photograph image using the first sliding window, utilizes default aesthetic evaluation model
Several corresponding aesthetics scorings of first sliding window during traversal are obtained, highest aesthetics scoring are chosen, with most
High aesthstic corresponding position of scoring is used as optimal aesthetic feeling clipping region；Wherein, the first sliding window choose it is following in one kind or
It is several：Original scale sliding window, standard proportional sliding window, preset ratio sliding window；
Or, for traveling through the photograph image using the second sliding window, obtain second slip during traversal
Several corresponding aesthetics scorings of window, several second sliding windows in ergodic process are used as using several described aesthetics scorings
The aesthetic quality value of center point coordinate, aesthetic quality collection of illustrative plates is generated according to several aesthetic quality values；Wherein, described second slide
Window is photograph image original scale, and the second sliding window is less than or equal to the first sliding window, the second sliding window step-length
Less than or equal to the first sliding window；
The energy highest zone in the aesthetic quality collection of illustrative plates is obtained using default clustering algorithm, with the energy highest zone
Minimum enclosed rectangle be used as the optimal aesthetic feeling region of image；
Calculate the window of required cutting ratio and the maximum of the optimal aesthetic feeling region of described image is handed over and simultaneously compared, maximum friendship will be produced and compared
Position be used as optimal aesthetic feeling clipping region.
9. the intelligent Scissoring device of photo as claimed in claim 7 or 8, it is characterised in that also including seed point acquisition module：
The seed point acquisition module, for obtaining the target clipping region seed point specified；
Further, the computing module is used for：
The target clipping region seed vertex neighborhood is traveled through using default sliding window, obtains described pre- during traversal
If sliding window it is corresponding several aesthetics scoring.
10. the intelligent Scissoring device of photo as claimed in claim 8, it is characterised in that
When first sliding window includes original scale sliding window and standard proportional sliding window, the computing module is used
In：
The photograph image is traveled through using original scale sliding window, the mistake in traversal is obtained using default aesthetic evaluation model
Several corresponding aesthetics scorings of original scale sliding window described in journey；
The photograph image is traveled through using standard proportional sliding window, the mistake in traversal is obtained using default aesthetic evaluation model
Several corresponding aesthetics scorings of standard proportional sliding window described in journey；
If corresponding in several corresponding aesthetics scorings of the original scale sliding window and the standard proportional sliding window
Highest aesthetics scoring is chosen in dry aesthetics scoring；
Or the photograph image is traveled through using original scale sliding window, obtained using default aesthetic evaluation model in traversal
During the original scale sliding window it is corresponding several aesthetics scoring；
Highest aesthetics scoring in several the corresponding aesthetics scorings of original scale sliding window is chosen, is commented according to highest aesthetics
The coordinate of point corresponding original scale sliding window obtains the optimal aesthetic feeling clipping region under original scale；
Calculate the maximum friendship of the standard proportional sliding window and the optimal aesthetic feeling clipping region under the original scale and compare, will
The position for producing maximum friendship and ratio is used as the optimal aesthetic feeling clipping region under standard proportional；
According in the optimal aesthetic feeling clipping region under instruction selection original scale or the optimal aesthetic feeling clipping region under standard proportional
One be used as the optimal aesthetic feeling clipping region of image；
Or calculate the maximum of the first sliding window and the optimal aesthetic feeling region of described image and hand over and compare, maximum friendship and ratio will be produced
Position comprises the following steps as optimal aesthetic feeling clipping region：
Maximum hand over for calculating original scale sliding window and the optimal aesthetic feeling region of described image is simultaneously compared, and will produce maximum friendship and ratio
Position is used as the optimal aesthetic feeling clipping region under original scale；
Maximum hand over for calculating standard proportional sliding window and the optimal aesthetic feeling region of described image is simultaneously compared, and will produce maximum friendship and ratio
Position is used as the optimal aesthetic feeling clipping region under standard proportional；
In being chosen in the optimal aesthetic feeling clipping region under optimal aesthetic feeling clipping region and standard proportional under the original scale
Nearest one of the heart and the optimal aesthetic feeling regional center of described image is used as the optimal aesthetic feeling clipping region of image.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710256861.XA CN107146198B (en) | 2017-04-19 | 2017-04-19 | Intelligent photo cutting method and device |
| PCT/CN2017/116058 WO2018192244A1 (en) | 2017-04-19 | 2017-12-14 | Shooting guidance method for intelligent device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710256861.XA CN107146198B (en) | 2017-04-19 | 2017-04-19 | Intelligent photo cutting method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107146198A true CN107146198A (en) | 2017-09-08 |
| CN107146198B CN107146198B (en) | 2022-08-16 |
Family
ID=59774749
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710256861.XA Active CN107146198B (en) | 2017-04-19 | 2017-04-19 | Intelligent photo cutting method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107146198B (en) |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107520881A (en) * | 2017-10-19 | 2017-12-29 | 国海芝 | A kind of photo automatic cutting machine people |
| CN108062739A (en) * | 2017-11-02 | 2018-05-22 | 广东数相智能科技有限公司 | A kind of picture intelligence method of cutting out and device based on body position |
| CN108062755A (en) * | 2017-11-02 | 2018-05-22 | 广东数相智能科技有限公司 | A kind of picture intelligence method of cutting out and device |
| CN109146892A (en) * | 2018-07-23 | 2019-01-04 | 北京邮电大学 | A kind of image cropping method and device based on aesthetics |
| CN109523503A (en) * | 2018-09-11 | 2019-03-26 | 北京三快在线科技有限公司 | A kind of method and apparatus of image cropping |
| CN110147833A (en) * | 2019-05-09 | 2019-08-20 | 北京迈格威科技有限公司 | Facial image processing method, apparatus, system and readable storage medium storing program for executing |
| CN110188652A (en) * | 2019-05-24 | 2019-08-30 | 北京字节跳动网络技术有限公司 | Processing method, device, terminal and the storage medium of facial image |
| CN111095293A (en) * | 2017-12-15 | 2020-05-01 | 华为技术有限公司 | Image aesthetic processing method and electronic equipment |
| CN111460929A (en) * | 2020-03-17 | 2020-07-28 | 稿定（厦门）科技有限公司 | Method and device for filling unspecified frame image |
| CN111696112A (en) * | 2020-06-15 | 2020-09-22 | 携程计算机技术（上海）有限公司 | Automatic image cutting method and system, electronic equipment and storage medium |
| CN112839167A (en) * | 2020-12-30 | 2021-05-25 | Oppo（重庆）智能科技有限公司 | Image processing method, image processing device, electronic equipment and computer readable medium |
| CN112884781A (en) * | 2021-02-24 | 2021-06-01 | 北京小米松果电子有限公司 | Image cropping method and device, electronic equipment and storage medium |
| CN113066008A (en) * | 2020-01-02 | 2021-07-02 | 杭州喔影网络科技有限公司 | Jigsaw generating method and equipment |
| CN113205522A (en) * | 2021-04-28 | 2021-08-03 | 华中科技大学 | Intelligent image clipping method and system based on antithetical domain adaptation |
| WO2021254510A1 (en) * | 2020-06-20 | 2021-12-23 | 华为技术有限公司 | Method for determining screenshot area, and related apparatus |
| CN114911551A (en) * | 2021-02-08 | 2022-08-16 | 花瓣云科技有限公司 | Display method and electronic equipment |
| CN116030280A (en) * | 2023-02-22 | 2023-04-28 | 青岛创新奇智科技集团股份有限公司 | Template matching method, device, storage medium and equipment |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103996186A (en) * | 2014-04-29 | 2014-08-20 | 小米科技有限责任公司 | Image cutting method and image cutting device |
| CN104717413A (en) * | 2013-12-12 | 2015-06-17 | 北京三星通信技术研究有限公司 | Shooting assistance method and equipment |
| CN105528757A (en) * | 2015-12-08 | 2016-04-27 | 华南理工大学 | Content-based image aesthetic quality improvement method |
| CN105787966A (en) * | 2016-03-21 | 2016-07-20 | 复旦大学 | An aesthetic evaluation method for computer pictures |
-
2017
- 2017-04-19 CN CN201710256861.XA patent/CN107146198B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104717413A (en) * | 2013-12-12 | 2015-06-17 | 北京三星通信技术研究有限公司 | Shooting assistance method and equipment |
| CN103996186A (en) * | 2014-04-29 | 2014-08-20 | 小米科技有限责任公司 | Image cutting method and image cutting device |
| CN105528757A (en) * | 2015-12-08 | 2016-04-27 | 华南理工大学 | Content-based image aesthetic quality improvement method |
| CN105787966A (en) * | 2016-03-21 | 2016-07-20 | 复旦大学 | An aesthetic evaluation method for computer pictures |
Non-Patent Citations (3)
| Title |
|---|
| MASASHI NISHIYAMA 等: "Sensation-based Photo Cropping", 《PROCEEDINGS OF THE 17TH ACM INTERNATIONAL CONFERENCE ON MULTIMEDIA》 * |
| YONG JIN 等: "Aesthetic photo composition by optimal crop-and-warp", 《COMPUTERS & GRAPHICS》 * |
| YUN LIANG 等: "Optimised image retargeting using aesthetic-based cropping and scaling", 《IET IMAGE PROCESSING》 * |
Cited By (25)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107520881A (en) * | 2017-10-19 | 2017-12-29 | 国海芝 | A kind of photo automatic cutting machine people |
| CN108062755B (en) * | 2017-11-02 | 2020-10-02 | 广东数相智能科技有限公司 | Intelligent picture clipping method and device |
| CN108062739A (en) * | 2017-11-02 | 2018-05-22 | 广东数相智能科技有限公司 | A kind of picture intelligence method of cutting out and device based on body position |
| CN108062755A (en) * | 2017-11-02 | 2018-05-22 | 广东数相智能科技有限公司 | A kind of picture intelligence method of cutting out and device |
| CN111095293A (en) * | 2017-12-15 | 2020-05-01 | 华为技术有限公司 | Image aesthetic processing method and electronic equipment |
| CN111095293B (en) * | 2017-12-15 | 2023-09-12 | 华为技术有限公司 | Image aesthetic processing method and electronic equipment |
| CN109146892A (en) * | 2018-07-23 | 2019-01-04 | 北京邮电大学 | A kind of image cropping method and device based on aesthetics |
| CN109146892B (en) * | 2018-07-23 | 2020-06-19 | 北京邮电大学 | Image clipping method and device based on aesthetics |
| CN109523503A (en) * | 2018-09-11 | 2019-03-26 | 北京三快在线科技有限公司 | A kind of method and apparatus of image cropping |
| WO2020052523A1 (en) * | 2018-09-11 | 2020-03-19 | 北京三快在线科技有限公司 | Method and apparatus for cropping image |
| US20210201445A1 (en) * | 2018-09-11 | 2021-07-01 | Beijing Sankuai Online Technology Co., Ltd | Image cropping method |
| CN110147833A (en) * | 2019-05-09 | 2019-08-20 | 北京迈格威科技有限公司 | Facial image processing method, apparatus, system and readable storage medium storing program for executing |
| CN110188652A (en) * | 2019-05-24 | 2019-08-30 | 北京字节跳动网络技术有限公司 | Processing method, device, terminal and the storage medium of facial image |
| CN113066008A (en) * | 2020-01-02 | 2021-07-02 | 杭州喔影网络科技有限公司 | Jigsaw generating method and equipment |
| CN111460929B (en) * | 2020-03-17 | 2022-07-19 | 稿定（厦门）科技有限公司 | Method and device for filling unspecified frame image |
| CN111460929A (en) * | 2020-03-17 | 2020-07-28 | 稿定（厦门）科技有限公司 | Method and device for filling unspecified frame image |
| CN111696112A (en) * | 2020-06-15 | 2020-09-22 | 携程计算机技术（上海）有限公司 | Automatic image cutting method and system, electronic equipment and storage medium |
| CN111696112B (en) * | 2020-06-15 | 2023-04-07 | 携程计算机技术（上海）有限公司 | Automatic image cutting method and system, electronic equipment and storage medium |
| WO2021254510A1 (en) * | 2020-06-20 | 2021-12-23 | 华为技术有限公司 | Method for determining screenshot area, and related apparatus |
| CN112839167A (en) * | 2020-12-30 | 2021-05-25 | Oppo（重庆）智能科技有限公司 | Image processing method, image processing device, electronic equipment and computer readable medium |
| CN114911551A (en) * | 2021-02-08 | 2022-08-16 | 花瓣云科技有限公司 | Display method and electronic equipment |
| CN112884781A (en) * | 2021-02-24 | 2021-06-01 | 北京小米松果电子有限公司 | Image cropping method and device, electronic equipment and storage medium |
| CN113205522A (en) * | 2021-04-28 | 2021-08-03 | 华中科技大学 | Intelligent image clipping method and system based on antithetical domain adaptation |
| CN113205522B (en) * | 2021-04-28 | 2022-05-13 | 华中科技大学 | Intelligent image clipping method and system based on antithetical domain adaptation |
| CN116030280A (en) * | 2023-02-22 | 2023-04-28 | 青岛创新奇智科技集团股份有限公司 | Template matching method, device, storage medium and equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107146198B (en) | 2022-08-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107146198A (en) | A kind of intelligent method of cutting out of photo and device | |
| US11321385B2 (en) | Visualization of image themes based on image content | |
| KR102195922B1 (en) | Internet-based facial beautification system | |
| EP3338217B1 (en) | Feature detection and masking in images based on color distributions | |
| US8692830B2 (en) | Automatic avatar creation | |
| US20190164248A1 (en) | Image grid with selectively prominent images | |
| EP3077898B1 (en) | Editing options for image regions | |
| KR101605983B1 (en) | Image recomposition using face detection | |
| US9256950B1 (en) | Detecting and modifying facial features of persons in images | |
| CN106326823B (en) | Method and system for obtaining head portrait in picture | |
| CN111222433B (en) | Automatic face auditing method, system, equipment and readable storage medium | |
| US20130108170A1 (en) | Image Recomposition From Face Detection And Facial Features | |
| TW201526604A (en) | Automatic photographing method and system thereof | |
| CN108319937A (en) | Method for detecting human face and device | |
| WO2021051605A1 (en) | Virtual video synthesis method and apparatus based on expression driving, and storage medium | |
| CN113254491A (en) | Information recommendation method and device, computer equipment and storage medium | |
| US20220207819A1 (en) | Light estimation using neural networks | |
| US20220101596A1 (en) | Object relighting using neural networks | |
| CN113781408A (en) | Intelligent guidance system and method for image shooting | |
| CN108537165A (en) | Method and apparatus for determining information | |
| WO2018192244A1 (en) | Shooting guidance method for intelligent device | |
| CN112839167B (en) | Image processing method, device, electronic equipment and computer readable medium | |
| CN113240573A (en) | Local and global parallel learning-based style transformation method and system for ten-million-level pixel digital image | |
| CN111080667A (en) | Automatic composition cutting method and system for rapid portrait photo | |
| USRE49044E1 (en) | Automatic avatar creation |
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 |