CN106650737A - Image automatic cutting method - Google Patents
Image automatic cutting method Download PDFInfo
- Publication number
- CN106650737A CN106650737A CN201611041091.9A CN201611041091A CN106650737A CN 106650737 A CN106650737 A CN 106650737A CN 201611041091 A CN201611041091 A CN 201611041091A CN 106650737 A CN106650737 A CN 106650737A
- Authority
- CN
- China
- Prior art keywords
- image
- cutting
- candidate
- aesthetic feeling
- cutting image
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 39
- 230000004044 response Effects 0.000 claims abstract description 54
- 239000000203 mixture Substances 0.000 claims abstract description 37
- 238000010586 diagram Methods 0.000 claims description 41
- 238000012549 training Methods 0.000 claims description 26
- 238000013527 convolutional neural network Methods 0.000 claims description 24
- 238000000605 extraction Methods 0.000 claims description 13
- 230000014759 maintenance of location Effects 0.000 claims description 10
- 238000013507 mapping Methods 0.000 claims description 4
- 230000001537 neural effect Effects 0.000 claims 1
- 230000007547 defect Effects 0.000 abstract description 3
- 238000012216 screening Methods 0.000 abstract description 2
- 230000000694 effects Effects 0.000 description 3
- 238000001914 filtration Methods 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 239000000523 sample Substances 0.000 description 3
- 230000008859 change Effects 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000012706 support-vector machine Methods 0.000 description 2
- FGUUSXIOTUKUDN-IBGZPJMESA-N C1(=CC=CC=C1)N1C2=C(NC([C@H](C1)NC=1OC(=NN=1)C1=CC=CC=C1)=O)C=CC=C2 Chemical compound C1(=CC=CC=C1)N1C2=C(NC([C@H](C1)NC=1OC(=NN=1)C1=CC=CC=C1)=O)C=CC=C2 FGUUSXIOTUKUDN-IBGZPJMESA-N 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000007935 neutral effect Effects 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 238000000059 patterning Methods 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
- G06V10/267—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion by performing operations on regions, e.g. growing, shrinking or watersheds
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Biophysics (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Biomedical Technology (AREA)
- Health & Medical Sciences (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to an image automatic cutting method. The method comprises the following steps: extracting an aesthetic response graph and a gradient energy graph of an image to be cut; intensively extracting candidate cutting images from the image to be cut; based on the aesthetic response graph, screening the candidate cutting images; and based on the aesthetic response graph and the gradient energy graph, estimating composition fractions of screened candidate cutting images and determining that a candidate cutting image with the highest score is a cutting image. According to the scheme, an aesthetic influence area of a picture is studied by use of the aesthetic response graph, an aesthetic reservation portion is determined by use of the aesthetic response graph, high aesthetic quality of the cutting image is better reserved to be greatest extent, at the same time, a gradient distribution rule is analyzed by use of the gradient energy graph, and the composition fraction of the cutting image is evaluated based on the aesthetic response graph and the gradient energy graph. The image automatic cutting method provided by the embodiment of the invention makes up for the defect of image composition expression and solves the problem of how to improve the robustness and precision of image automatic cutting.
Description
Technical field
The present invention relates to pattern-recognition, machine learning and technical field of computer vision, more particularly to a kind of image is automatic
Method of cutting out.
Background technology
With the fast development of computer technology and digital media technology, people are to computer vision, artificial intelligence, machine
The demand in the fields such as perception and expectation also more and more higher.One during the automatic cutting of image is edited automatically as image weighs very much
Also to get growing concern for and develop with common task.Image automatic cutting technology be exactly want to remove it is unnecessary
Region, area-of-interest is emphasized, so as to improve the overall composition and aesthetic qualities of image.A kind of effective and automatic image
Method of cutting out can not only be such that the mankind free from loaded down with trivial details work, and can also provide to some layman
The suggestion of the picture editting of specialty.
Because image cropping is the task of a unusual subjectivity, existing rule is difficult to consider all influence factors.Pass
The image automatic cutting region of system Saliency maps are usually used to recognize image in main region or area-of-interest, while logical
Cross some rules formulated to come computation energy function minimum or Study strategies and methods to find clipping region.What but these were formulated
Rule is to the task of image cropping this subjectivity and not comprehensive enough, and precision also is difficult to reach user's request.
In view of this, it is special to propose the present invention.
The content of the invention
In order to solve the problems referred to above of the prior art, be solve how to improve image automatic cutting robustness and
The technical problem of precision and a kind of automatic image cutting out method is provided.
To achieve these goals, there is provided technical scheme below:
A kind of automatic image cutting out method, methods described includes:
The aesthetic feeling response diagram and gradient energy figure of cutting image is treated in extraction;
The intensive extraction candidate's cutting image of cutting image is treated to described;
Based on the aesthetic feeling response diagram, candidate's cutting image is screened;
Based on the aesthetic feeling response diagram and the gradient energy figure, the composition for estimating the candidate's cutting image for filtering out divides
Number, and candidate's cutting image of highest scoring is defined as into cutting image.
Further, it is described to extract the aesthetic feeling response diagram and gradient energy figure for treating cutting image, specifically include:
Using depth convolutional neural networks and classification response mapping method, and cutting figure is treated using described in equation below extraction
The aesthetic feeling response diagram of picture:
Wherein, the M (x, y) represents the aesthetic feeling response at locus (x, y) place;The K represents depth convolution god
The overall channel number of the characteristic pattern of last layer of convolutional layer of Jing networks;The k represents k-th passage;The fk(x, y) is represented
Characteristic value of k-th passage at locus (x, the y) place;The wkRepresent the characteristic pattern pond of k-th passage
Weights of the result after change to high aesthetic feeling classification;
Treat that cutting image is smoothed to described, and calculate the Grad of each pixel, so as to obtain the ladder
Degree energy diagram.
Further, the depth convolutional neural networks are trained obtain in the following manner:
In the bottom of the depth convolutional neural networks structure, convolutional layer is set;
By the method in global average pond after last convolutional layer of the depth convolutional neural networks structure,
Each characteristic pattern pond is turned into a point;
Connection and the full articulamentum of aesthetic qualities class categories number identical and loss function.
Further, it is described based on the aesthetic feeling response diagram, candidate's cutting image is screened, specifically include:
The aesthetic feeling retention score of candidate's cutting image is calculated by equation below:
Wherein, the Sa(C) the aesthetic feeling retention score of candidate's cutting image is represented;The C represents the time
Select cutting image;(i, j) represents the position of pixel;The I represents original image;The A(i,j)Represent in (i, j) position
The aesthetic feeling response at place;
All candidate's cutting images are ranked up from big to small according to the aesthetic feeling retention score;
Choose a part of candidate's cutting image of highest scoring.
Further, it is described based on the aesthetic feeling response diagram and the gradient energy figure, estimate the candidate's cutting for filtering out
The composition fraction of image, and candidate's cutting image of highest scoring is defined as into cutting image, specifically include:
Composition model is set up based on the aesthetic feeling response diagram and the gradient energy figure;
Using the composition fraction of the candidate's cutting image filtered out described in composition model estimation, and by the score most
High candidate's cutting image is defined as the cutting image.
Further, the composition model is obtained in the following manner:
Training image collection is set up based on the aesthetic feeling response diagram and the gradient energy figure;
The mark of aesthetic qualities classification is carried out to training image;
Using the training image training depth convolutional neural networks of mark;
For the training image for having marked, using the depth convolutional neural networks for training, extract the aesthetic feeling and ring
The spatial pyramid feature with the gradient energy figure should be schemed;
By the spatial pyramid merging features for extracting together;
It is trained using grader, automatically study composition rule, obtains composition model.
The embodiment of the present invention provides a kind of automatic image cutting out method.The method includes:The aesthetic feeling of cutting image is treated in extraction
Response diagram and gradient energy figure;Treat the intensive extraction candidate's cutting image of cutting image;Based on aesthetic feeling response diagram, screening candidate cut out
Cut image;Based on aesthetic feeling response diagram and gradient energy figure, the composition fraction of candidate's cutting image for filtering out is estimated, and by score
Highest candidate's cutting image is defined as cutting image.This programme goes to probe into the aesthetic feeling zone of influence of picture using aesthetic feeling response diagram
Domain, using aesthetic feeling response diagram aesthetic feeling member-retaining portion is determined, so as to more farthest remain the high aesthetic feeling matter of cutting image
Amount, while this programme also goes to analyze gradient distribution rule using gradient energy figure, and based on aesthetic feeling response diagram and gradient energy
Figure is assessing the composition fraction of cutting figure.The embodiment of the present invention compensate for the defect of image composition expression, solves and how to improve
The robustness of image automatic cutting and the technical problem of precision.The embodiment of the present invention can apply to be related to the crowd of image automatic cutting
It is multi-field, including the reorientation of picture editting, photography and image etc..
Description of the drawings
Fig. 1 is the schematic flow sheet of automatic image cutting out method according to embodiments of the present invention;
Fig. 2 is the structural representation of depth convolutional neural networks according to embodiments of the present invention;
Fig. 3 a are according to embodiments of the present invention to treat cutting image schematic diagram;
Fig. 3 b are the image schematic diagrames after cutting according to embodiments of the present invention.
Specific embodiment
Below in conjunction with the accompanying drawings and the specific embodiment technical problem that the embodiment of the present invention is solved, the technical side that adopted
Case and the technique effect of realization carry out clear, complete description.Obviously, described embodiment is only of the application
Divide embodiment, be not whole embodiments.Based on the embodiment in the application, those of ordinary skill in the art are not paying creation
Property work on the premise of, all other equivalent for being obtained or the embodiment of obvious modification are all fallen within protection scope of the present invention.
The embodiment of the present invention can embody according to the multitude of different ways being defined and covered by claim.
Deep learning has obtained quick development and well effect in every field.The embodiment of the present invention is considered using deep
The automatically study influence area important to image cropping is gone in degree study, with comprehensively learning rules automatically, so that in cutting
When retain high aesthetic feeling region as much as possible.
For this purpose, the embodiment of the present invention provides a kind of automated graphics method of cutting out.Fig. 1 schematically illustrates image automatic cutting
The flow process of shear method.As shown in figure 1, the method can include:
S100:The aesthetic feeling response diagram and gradient energy figure of cutting image is treated in extraction.
Specifically, this step can include:
S101:Using depth convolutional neural networks and classification response mapping method, and cutting is treated using equation below extraction
The aesthetic feeling response diagram of image:
Wherein, M (x, y) represents the aesthetic feeling response at locus (x, y) place;K represents the depth convolution god for training
The overall channel number of the characteristic pattern f of last layer of convolutional layer of Jing networks;K represents k-th passage;fk(x, y) represents k-th to lead to
Characteristic value of the road at locus (x, y) place;wkRepresent the result behind the characteristic pattern pond of k-th passage to high aesthetic feeling classification
Weights.
Above-mentioned steps can according to actual needs train depth convolutional neural networks when aesthetic feeling response diagram is extracted.Depth is rolled up
The training of product neutral net can be carried out in the following manner:
Step 1:In the bottom of depth convolutional neural networks structure, convolutional layer is set.
Step 2:By the side in global average pond after last convolutional layer of depth convolutional neural networks structure
Method, by each characteristic pattern pond a point is turned to.
Step 3:Connection one and the full articulamentum of aesthetic qualities class categories number identical and loss function.
Fig. 2 schematically illustrates a depth convolutional neural networks structure.
One depth convolutional neural networks model under aesthetic qualities classification task can be trained by step 1-3.So
Afterwards, depth convolutional neural networks and classification response mapping method that aesthetic qualities classification task is trained are utilized as;Again using upper
Formula is stated, the aesthetic feeling response diagram M that cutting image is treated under high aesthetic feeling classification can be calculated.
S102:Treat cutting image to be smoothed, and calculate the Grad of each pixel, so as to obtain gradient energy
Spirogram.
S110:Treat the intensive extraction candidate's cutting image of cutting image.
Here it is possible to using the sliding window of all sizes less than image size, treat intensive extraction of cutting image and wait
Crop window is selected, candidate's cutting image is extracted by candidate's crop window.
S120:Based on aesthetic feeling response diagram, candidate's cutting image is screened.
Specifically, this step can include:
S121:The aesthetic feeling retention score of candidate's cutting image is calculated by equation below:
Wherein, Sa(C) the aesthetic feeling retention score of candidate's cutting image is represented;C represents candidate's cutting image;(i, j) is represented
The position of pixel;I represents original image;A(i,j)Represent the aesthetic feeling response at (i, j) place.
Aesthetic feeling reserving model can be built by this step.Candidate's crop window is filtered out into U.S. through aesthetic feeling reserving model
The higher candidate window of sense retention score.
S122:All candidate's cutting images are ranked up from big to small according to aesthetic feeling retention score.
S123:Choose a part of candidate's cutting image of highest scoring.
For example:The candidate's cutting image retained in front 10000 candidate's crop windows can be set in practical application.
S130:The composition fraction of the candidate's cutting image filtered out based on aesthetic feeling response diagram and gradient energy figure, estimation, and
Candidate's cutting image of highest scoring is defined as into cutting image.
Specifically, this step can be realized by step S131 to step S133.
S131:Composition model is set up based on aesthetic feeling response diagram and gradient energy figure.
This step can train composition model when composition model is set up according to actual conditions.In the mistake of training composition model
Cheng Zhong, training data can adopt the preferable image of composition as positive sample, and using the image for having patterning defects as negative sample.
Composition model can in the following manner be trained:
Step a:Training image collection is set up based on aesthetic feeling response diagram and gradient energy figure.
Step b:The mark of aesthetic qualities classification is carried out to training image.
Step c:Using the training image training depth convolutional neural networks of mark.
The training process of this step may be referred to above-mentioned steps 1 to step 3, will not be described here.
Step d:For the training image for having marked, using the depth convolutional neural networks for training, aesthetic feeling response is extracted
The spatial pyramid feature of figure and gradient energy figure.
Step e:By the spatial pyramid merging features for extracting together.
Step f:It is trained using grader, automatically study composition rule, obtains composition model.
Wherein, grader can for example adopt support vector machine classifier.
S132:The composition fraction of candidate's cutting image for estimating to filter out using composition model, and by the time of highest scoring
Cutting image is selected to be defined as cutting image.
Fig. 3 a are schematically illustrated and are treated cutting image;Fig. 3 b schematically illustrate the image after cutting.
Again the present invention is better described with a preferred embodiment below.
Step A:The image data set for being labeled with aesthetic qualities classification is sent into depth convolutional neural networks carries out aesthetic feeling matter
Amount class models training.
Step B:The image data set for being labeled with composition classification is input into the depth convolutional neural networks for training, is extracted most
The characteristic pattern of later layer convolutional layer, and aesthetic feeling response diagram is calculated, while aesthetic feeling gradient map is calculated, then using SVMs point
Class device trains composition model.
Step C:Treat test image and extract aesthetic feeling response diagram and gradient energy figure.
The method that the extracting method of this step refers to the training stage.
Step D:The intensive candidate's crop window for gathering image to be tested.
For example, on 1000 × 1000 image to be tested, carried out using the sliding window at intervals of 30 pixels
Collection is extracted.
Step E:Candidate's crop window is screened using aesthetic feeling reserving model.
The aesthetic feeling retention score of candidate's crop window that this step is collected using aesthetic feeling reserving model computation-intensive, screening
Go out a part of candidate's crop window of aesthetic feeling classification highest, for example:Filter out 10000 candidate's crop windows.
Step F:The candidate's crop window filtered out using composition model evaluation.
The composition model that this step collection training stage trains removes the composition point for assessing the candidate's crop window for filtering out
Number, using highest scoring as last crop window, so as to obtain cutting image.
In sum, method provided in an embodiment of the present invention make use of well aesthetic feeling response diagram and gradient energy figure to come most
Big degree ground retains the composition rule of aesthetic qualities and image, obtains more robust, the automatic cutting of the higher image of precision
Can, aesthetic feeling response diagram and gradient energy figure have been further related to for the validity of image automatic cutting.
Although describing method provided in an embodiment of the present invention, ability according to above-mentioned precedence in above-described embodiment
Field technique personnel are appreciated that to realize the effect of the present embodiment, can be with parallel or reverse the right order etc. different
Performing, these simply change all within protection scope of the present invention order.
The above, the only specific embodiment in the present invention, but protection scope of the present invention is not limited thereto, and appoints
What be familiar with the people of the technology disclosed herein technical scope in, it will be appreciated that the conversion expected or replacement, all should cover
The present invention include within the scope of, therefore, protection scope of the present invention should be defined by the protection domain of claims.
Claims (6)
1. a kind of automatic image cutting out method, it is characterised in that methods described includes:
The aesthetic feeling response diagram and gradient energy figure of cutting image is treated in extraction;
The intensive extraction candidate's cutting image of cutting image is treated to described;
Based on the aesthetic feeling response diagram, candidate's cutting image is screened;
The composition fraction of the candidate's cutting image filtered out based on the aesthetic feeling response diagram and the gradient energy figure, estimation, and
Candidate's cutting image of highest scoring is defined as into cutting image.
2. method according to claim 1, it is characterised in that the aesthetic feeling response diagram and gradient of cutting image is treated in the extraction
Energy diagram, specifically includes:
Using depth convolutional neural networks and classification response mapping method, and cutting image is treated using described in equation below extraction
The aesthetic feeling response diagram:
Wherein, the M (x, y) represents the aesthetic feeling response at locus (x, y) place;The K represents depth convolutional Neural net
The overall channel number of the characteristic pattern of last layer of convolutional layer of network;The k represents k-th passage;The fk(x, y) represents described
Characteristic value of k-th passage at locus (x, the y) place;The wkAfter representing the characteristic pattern pond of k-th passage
Result to high aesthetic feeling classification weights;
Treat that cutting image is smoothed to described, and calculate the Grad of each pixel, so as to obtain the gradient energy
Spirogram.
3. method according to claim 2, it is characterised in that the depth convolutional neural networks are trained in the following manner
Obtain:
In the bottom of the depth convolutional neural networks structure, convolutional layer is set;
By the method in global average pond after last convolutional layer of the depth convolutional neural networks structure, will be every
One characteristic pattern pond turns to a point;
Connection and the full articulamentum of aesthetic qualities class categories number identical and loss function.
4. method according to claim 1, it is characterised in that described based on the aesthetic feeling response diagram, screens the candidate
Cutting image, specifically includes:
The aesthetic feeling retention score of candidate's cutting image is calculated by equation below:
Wherein, the Sa(C) the aesthetic feeling retention score of candidate's cutting image is represented;The C represents candidate's cutting
Image;(i, j) represents the position of pixel;The I represents original image;The A(i,j)Represent the U.S. at (i, j) position
Sense response;
All candidate's cutting images are ranked up from big to small according to the aesthetic feeling retention score;
Choose a part of candidate's cutting image of highest scoring.
5. method according to claim 1, it is characterised in that described based on the aesthetic feeling response diagram and the gradient energy
Figure, the composition fraction of candidate's cutting image that estimation is filtered out, and candidate's cutting image of highest scoring is defined as into cutting figure
Picture, specifically includes:
Composition model is set up based on the aesthetic feeling response diagram and the gradient energy figure;
Using the composition model estimate described in the composition fraction of candidate's cutting image that filters out, and by the highest scoring
Candidate's cutting image is defined as the cutting image.
6. method according to claim 5, it is characterised in that the composition model is obtained in the following manner:
Training image collection is set up based on the aesthetic feeling response diagram and the gradient energy figure;
The mark of aesthetic qualities classification is carried out to training image;
Using the training image training depth convolutional neural networks of mark;
For the training image for having marked, using the depth convolutional neural networks for training, the aesthetic feeling response diagram is extracted
With the spatial pyramid feature of the gradient energy figure;
By the spatial pyramid merging features for extracting together;
It is trained using grader, automatically study composition rule, obtains composition model.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611041091.9A CN106650737B (en) | 2016-11-21 | 2016-11-21 | Automatic image cutting method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611041091.9A CN106650737B (en) | 2016-11-21 | 2016-11-21 | Automatic image cutting method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106650737A true CN106650737A (en) | 2017-05-10 |
CN106650737B CN106650737B (en) | 2020-02-28 |
Family
ID=58811471
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201611041091.9A Active CN106650737B (en) | 2016-11-21 | 2016-11-21 | Automatic image cutting method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106650737B (en) |
Cited By (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107317962A (en) * | 2017-05-12 | 2017-11-03 | 广东网金控股股份有限公司 | A kind of intelligence, which is taken pictures, cuts patterning system and application method |
CN107392244A (en) * | 2017-07-18 | 2017-11-24 | 厦门大学 | The image aesthetic feeling Enhancement Method returned based on deep neural network with cascade |
CN107545576A (en) * | 2017-07-31 | 2018-01-05 | 华南农业大学 | Image edit method based on composition rule |
CN108154464A (en) * | 2017-12-06 | 2018-06-12 | 中国科学院自动化研究所 | The method and device of picture automatic cutting based on intensified learning |
CN108566512A (en) * | 2018-03-21 | 2018-09-21 | 珠海市魅族科技有限公司 | A kind of intelligence image pickup method, device, computer equipment and readable storage medium storing program for executing |
CN109518446A (en) * | 2018-12-21 | 2019-03-26 | 季华实验室 | Intelligent cutting method of cutting machine |
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 |
CN109886317A (en) * | 2019-01-29 | 2019-06-14 | 中国科学院自动化研究所 | General image aesthetics appraisal procedure, system and equipment based on attention mechanism |
CN110062173A (en) * | 2019-03-15 | 2019-07-26 | 北京旷视科技有限公司 | Image processor and image processing method, equipment, storage medium and intelligent terminal |
WO2020186385A1 (en) * | 2019-03-15 | 2020-09-24 | 深圳市大疆创新科技有限公司 | Image processing method, electronic device, and computer-readable storage medium |
WO2020232672A1 (en) * | 2019-05-22 | 2020-11-26 | 深圳市大疆创新科技有限公司 | Image cropping method and apparatus, and photographing 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 |
CN113436224A (en) * | 2021-06-11 | 2021-09-24 | 华中科技大学 | Intelligent image clipping method and device based on explicit composition rule modeling |
WO2023093851A1 (en) * | 2021-11-29 | 2023-06-01 | 维沃移动通信有限公司 | Image cropping method and apparatus, and electronic device |
Citations (5)
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 |
CN105488758A (en) * | 2015-11-30 | 2016-04-13 | 河北工业大学 | Image scaling method based on content awareness |
CN105528786A (en) * | 2015-12-04 | 2016-04-27 | 小米科技有限责任公司 | Image processing method and device |
CN105787966A (en) * | 2016-03-21 | 2016-07-20 | 复旦大学 | An aesthetic evaluation method for computer pictures |
CN105894025A (en) * | 2016-03-30 | 2016-08-24 | 中国科学院自动化研究所 | Natural image aesthetic feeling quality assessment method based on multitask deep learning |
-
2016
- 2016-11-21 CN CN201611041091.9A patent/CN106650737B/en active Active
Patent Citations (5)
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 |
CN105488758A (en) * | 2015-11-30 | 2016-04-13 | 河北工业大学 | Image scaling method based on content awareness |
CN105528786A (en) * | 2015-12-04 | 2016-04-27 | 小米科技有限责任公司 | Image processing method and device |
CN105787966A (en) * | 2016-03-21 | 2016-07-20 | 复旦大学 | An aesthetic evaluation method for computer pictures |
CN105894025A (en) * | 2016-03-30 | 2016-08-24 | 中国科学院自动化研究所 | Natural image aesthetic feeling quality assessment method based on multitask deep learning |
Non-Patent Citations (2)
Title |
---|
侯丹红: "相片中重要对象布局优化系统", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
王伟凝 等: "基于并行深度卷积神经网络的图像美感分类", 《自动化学报》 * |
Cited By (20)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107317962A (en) * | 2017-05-12 | 2017-11-03 | 广东网金控股股份有限公司 | A kind of intelligence, which is taken pictures, cuts patterning system and application method |
CN107317962B (en) * | 2017-05-12 | 2019-11-08 | 广东网金控股股份有限公司 | A kind of intelligence, which is taken pictures, cuts patterning system and application method |
CN107392244A (en) * | 2017-07-18 | 2017-11-24 | 厦门大学 | The image aesthetic feeling Enhancement Method returned based on deep neural network with cascade |
CN107392244B (en) * | 2017-07-18 | 2020-08-28 | 厦门大学 | Image aesthetic feeling enhancement method based on deep neural network and cascade regression |
CN107545576A (en) * | 2017-07-31 | 2018-01-05 | 华南农业大学 | Image edit method based on composition rule |
CN108154464A (en) * | 2017-12-06 | 2018-06-12 | 中国科学院自动化研究所 | The method and device of picture automatic cutting based on intensified learning |
CN108154464B (en) * | 2017-12-06 | 2020-09-22 | 中国科学院自动化研究所 | Method and device for automatically clipping picture based on reinforcement learning |
CN108566512A (en) * | 2018-03-21 | 2018-09-21 | 珠海市魅族科技有限公司 | A kind of intelligence image pickup method, device, computer equipment and readable storage medium storing program for executing |
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 |
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 |
CN109518446A (en) * | 2018-12-21 | 2019-03-26 | 季华实验室 | Intelligent cutting method of cutting machine |
CN109886317A (en) * | 2019-01-29 | 2019-06-14 | 中国科学院自动化研究所 | General image aesthetics appraisal procedure, system and equipment based on attention mechanism |
CN109886317B (en) * | 2019-01-29 | 2021-04-27 | 中国科学院自动化研究所 | General image aesthetic evaluation method, system and equipment based on attention mechanism |
CN110062173A (en) * | 2019-03-15 | 2019-07-26 | 北京旷视科技有限公司 | Image processor and image processing method, equipment, storage medium and intelligent terminal |
WO2020186385A1 (en) * | 2019-03-15 | 2020-09-24 | 深圳市大疆创新科技有限公司 | Image processing method, electronic device, and computer-readable storage medium |
WO2020232672A1 (en) * | 2019-05-22 | 2020-11-26 | 深圳市大疆创新科技有限公司 | Image cropping method and apparatus, and photographing 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 |
CN113436224A (en) * | 2021-06-11 | 2021-09-24 | 华中科技大学 | Intelligent image clipping method and device based on explicit composition rule modeling |
CN113436224B (en) * | 2021-06-11 | 2022-04-26 | 华中科技大学 | Intelligent image clipping method and device based on explicit composition rule modeling |
WO2023093851A1 (en) * | 2021-11-29 | 2023-06-01 | 维沃移动通信有限公司 | Image cropping method and apparatus, and electronic device |
Also Published As
Publication number | Publication date |
---|---|
CN106650737B (en) | 2020-02-28 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106650737A (en) | Image automatic cutting method | |
CN109800736B (en) | Road extraction method based on remote sensing image and deep learning | |
CN104616274B (en) | A kind of multi-focus image fusing method based on salient region extraction | |
EP3819859B1 (en) | Sky filter method for panoramic images and portable terminal | |
CN110376198B (en) | Cervical liquid-based cell slice quality detection system | |
CN107665492B (en) | Colorectal panoramic digital pathological image tissue segmentation method based on depth network | |
CN109241967B (en) | Thyroid ultrasound image automatic identification system based on deep neural network, computer equipment and storage medium | |
EP3343440A1 (en) | Identifying and excluding blurred areas of images of stained tissue to improve cancer scoring | |
WO2018090355A1 (en) | Method for auto-cropping of images | |
CN105989347A (en) | Intelligent marking method and system of objective questions | |
CN106651872A (en) | Prewitt operator-based pavement crack recognition method and system | |
CN107240096A (en) | A kind of infrared and visual image fusion quality evaluating method | |
CN104766097B (en) | Surface of aluminum plate defect classification method based on BP neural network and SVMs | |
CN103206208A (en) | Method for macroscopically quantizing microscopic remaining oil in different occurrence states | |
US20110002516A1 (en) | Method and device for dividing area of image of particle in urine | |
CN102750538A (en) | Go competition result analysis method based on image processing technique | |
CN107341790A (en) | A kind of image processing method of environment cleanliness detection | |
CN111008647B (en) | Sample extraction and image classification method based on void convolution and residual linkage | |
CN109978771A (en) | Cell image rapid fusion method based on content analysis | |
CN106780514A (en) | The computational methods of the heavy rain Ji Lao areas depth of accumulated water based on monitor video image | |
Bobbe et al. | A primer on mapping vegetation using remote sensing | |
CN110874824B (en) | Image restoration method and device | |
CN105760878A (en) | Method and device for selecting urinary sediment microscope image with optimal focusing performance | |
CN114419465B (en) | Method, device and equipment for detecting change of remote sensing image and storage medium | |
EP4261774A1 (en) | Object classification device, object classification system, and object classification program |
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 |