CN101540050B - Method and device for obtaining scene boundary - Google Patents
Method and device for obtaining scene boundary Download PDFInfo
- Publication number
- CN101540050B CN101540050B CN200910083420XA CN200910083420A CN101540050B CN 101540050 B CN101540050 B CN 101540050B CN 200910083420X A CN200910083420X A CN 200910083420XA CN 200910083420 A CN200910083420 A CN 200910083420A CN 101540050 B CN101540050 B CN 101540050B
- Authority
- CN
- China
- Prior art keywords
- boundary
- key points
- search
- border
- threshold value
- 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.)
- Expired - Fee Related
Links
Images
Abstract
The invention discloses a method and a device for obtaining scene boundary, and belongs to the field of computer multimedia. The method comprises the steps of obtaining divergent boundary points of each piece of single-channel gray information of multiple pieces of single-channel gray information of an image; obtaining multiple key points at the scene boundary and arranging the multiple key points according to sequence; searching to obtain the effective boundary between two adjacent key points using a shortest path search algorithm according to the divergent boundary point of each piece of single-channel gray information; and connecting the effective boundary between two adjacent key points to obtain the boundary of the scene. The device comprises a first obtaining module, a second obtaining module, a searching module and a connecting module. The invention avoids the wrong segmentation phenomenon caused by complex scene texture by comprehensively utilising multiple pieces of single-channel scene boundary information and setting the key points, improves accuracy for obtaining the scene boundary, improves the efficiency for obtaining the scene boundary and can save lots of time. The human intervention in the process of obtaining the scene boundary is low.
Description
Technical field
The present invention relates to field of computer multimedia, particularly a kind of method and apparatus that obtains scene border.
Background technology
Along with the development of computer media, convert planar video to three-dimensional video-frequency and be subjected to paying close attention to widely and paying attention to.The key that planar video converts three-dimensional video-frequency to is the extraction of depth map, and because the scenery in the image has the different degree of depth with background, when generating depth map, often need to obtain earlier scene border, scenery is separated from background, and then respectively scenery and background are carried out depth value and give, generate depth map.
Obtain scene border in the prior art and mainly contain following two kinds of methods:
1) utilizes mapping software such as Photoshop, manual scenery on the image is extracted, obtain scene border.
2) by image segmentation algorithm image is cut apart, scenery and background separation are opened, obtain scene border.
In realizing process of the present invention, the inventor finds that there is following problem at least in prior art:
1) the artificial participation height of manual extraction method, treatment effeciency are low, need expend the plenty of time; And rely on artificial experience and technology, portable poor.
2) image segmentation algorithm tends to it is divided into different scenery, poor accuracy for the abundant same scenery of superficial makings; And, manually be difficult to intervene, interactive poor.
Summary of the invention
In order to improve efficient and the accuracy of obtaining scene border, the embodiment of the invention provides a kind of method and apparatus that obtains scene border.Described technical scheme is as follows:
On the one hand, the embodiment of the invention provides a kind of method of obtaining scene border, and this method comprises:
Obtain the divergent boundary point of each the single channel half-tone information in a plurality of single channel half-tone informations of image;
Obtain a plurality of key points on the scene border, and described a plurality of key points are arranged in order;
According to the divergent boundary point of described each single channel half-tone information, calculate the boundary intensity of each pixel of described image;
Determine the boundary search zone between per two adjacent key points;
Boundary intensity according to each pixel of described image utilizes Shortest Path Searching Algorithm, in the boundary search zone between described per two adjacent key points of determining, searches for and obtain efficiency frontier between described per two adjacent key points;
Connect the efficiency frontier between per two adjacent key points, obtain the border of scenery.
On the other hand, the embodiment of the invention provides a kind of device that obtains scene border, and this device comprises:
First acquisition module is used for obtaining the divergent boundary point of each single channel half-tone information of a plurality of single channel half-tone informations of image;
Second acquisition module is used to obtain a plurality of key points on the scene border, and described a plurality of key points are arranged in order;
Search module is used for the divergent boundary point according to described each single channel half-tone information, calculates the boundary intensity of each pixel of described image; Determine the boundary search zone between per two adjacent key points; Boundary intensity according to each pixel of described image utilizes Shortest Path Searching Algorithm, in the boundary search zone between described per two adjacent key points of determining, searches for and obtain efficiency frontier between described per two adjacent key points;
Link block is used for connecting the efficiency frontier between described per two adjacent key points after described search module obtains efficiency frontier between per two adjacent key points, obtains the border of scenery.
The beneficial effect of the technical scheme that the embodiment of the invention provides is:
By fully utilizing the scene border information under a plurality of single channels and the setting of key point, avoided having improved the accuracy of obtaining scene border because the mistake that scenery texture complexity causes is cut apart phenomenon; It is low to obtain in the scene border process artificial participation, has improved the efficient of obtaining scene border, can save the plenty of time.
Description of drawings
Fig. 1 is a kind of method flow diagram that obtains scene border that the embodiment of the invention 1 provides;
Fig. 2 is a kind of method flow diagram that obtains scene border that the embodiment of the invention 2 provides;
Fig. 3 is a kind of apparatus structure synoptic diagram that obtains scene border that the embodiment of the invention 3 provides.
Embodiment
For making the purpose, technical solutions and advantages of the present invention clearer, embodiment of the present invention is described further in detail below in conjunction with accompanying drawing.
Embodiment 1
Referring to Fig. 1, the embodiment of the invention provides a kind of method of obtaining scene border, and this method comprises:
101: the divergent boundary point that obtains each the single channel half-tone information in a plurality of single channel half-tone informations of image;
102: obtain a plurality of key points on the scene border, and a plurality of key points are arranged in order;
103: the divergent boundary point according to each single channel half-tone information, utilize Shortest Path Searching Algorithm, search for and obtain efficiency frontier between per two adjacent key points;
104: connect the efficiency frontier between per two adjacent key points, obtain the border of scenery.
Particularly, the divergent boundary point of each single channel half-tone information that obtains a plurality of single channel half-tone informations of image can comprise:
Image is carried out color notation conversion space, obtain a plurality of single channel half-tone informations;
Utilize the boundary search algorithm, each the single channel half-tone information in a plurality of single channel half-tone informations is carried out boundary search respectively, obtain the divergent boundary point of each single channel half-tone information.
Particularly, the divergent boundary point according to each single channel half-tone information utilizes Shortest Path Searching Algorithm, searches for and obtain efficiency frontier between per two adjacent key points, can comprise:
According to the divergent boundary point of each single channel half-tone information, calculate the boundary intensity of each pixel of image;
Determine the boundary search zone between per two adjacent key points;
Boundary intensity according to each pixel of image utilizes Shortest Path Searching Algorithm, in the boundary search zone between per two adjacent key points of determining, searches for and obtain efficiency frontier between per two adjacent key points.
Particularly, the boundary intensity according to each pixel of image utilizes Shortest Path Searching Algorithm, in the boundary search zone between per two adjacent key points of determining, searches for and obtain efficiency frontier between per two adjacent key points, can comprise:
The boundary intensity of each pixel of image and default passage threshold value are compared, obtain the efficiency frontier point of image;
Carry out the cum rights shortest path through search by the efficiency frontier point, in the boundary search zone between per two adjacent key points of determining, search for and obtain current border between per two adjacent key points;
Judge that whether ratio between the pixel total in the pixel of filling in the current border and the current border is greater than default proportion threshold value;
When judged result is during smaller or equal to default proportion threshold value, with current border as and two adjacent key points of should the front bound pair answering between efficiency frontier.
Further, when judged result is during greater than default proportion threshold value, can also comprise:
Judge that whether default passage threshold value is greater than 1;
When judged result is the passage threshold value preset greater than 1 the time, default passage threshold value is subtracted 1, carries out above-mentioned boundary intensity and compare with the passage threshold value of presetting with each pixel in the image, obtain the step of the efficiency frontier point of image.
Further, when judged result is the passage threshold value preset smaller or equal to 1 the time, can also comprise:
With current border as and deserve efficiency frontier between two adjacent key points that the front bound pair answers.
The described method of present embodiment by fully utilizing the scene border information under a plurality of single channels and the setting of key point, has been avoided having improved the accuracy of obtaining scene border because the mistake that scenery texture complexity causes is cut apart phenomenon; It is low to obtain in the scene border process artificial participation, has improved the efficient of obtaining scene border, can save the plenty of time, and not need to rely on artificial experience and technology, portable high; And can manually choose key point, manually can intervene, strengthen interactive; Further, the passage threshold value screening frontier point by default when obscure boundary is clear, can reduces the passage threshold value and obtain abundanter boundary information, makes sharpness of border; When boundary information too enriches, can improve the passage threshold value and reduce interference; In addition,, reduced the working strength of manually a plurality of single channel lower boundary points being screened, improved work efficiency, can save the plenty of time by default passage threshold adaptive ground screening frontier point is set.
Embodiment 2
Referring to Fig. 2, the embodiment of the invention provides a kind of method of obtaining scene border, and this method comprises:
201: image is carried out color notation conversion space, obtain a plurality of single channel half-tone informations.
Wherein, image is the image that the needs that receive therefrom extract scene border, can be various images such as coloured image, gray level image.And image generally is the information combination of a plurality of passages, and for example: image is the information combination of R (Red, red), G (Green, green) and three passages of B (Blue, indigo plant).Can be by image be carried out color notation conversion space, obtain R, G, B, X, Y, Z, H (Hue, tone), S (saturation, saturation degree) and V (value, lightness) waits half-tone information under each single channel, for convenience of description the half-tone information under the single channel is called the single channel half-tone information, promptly the single channel half-tone information only comprises the information under the passage.And can be provided with according to actual needs and specifically obtain what single channel half-tone informations.
Particularly, rgb space is as follows to the formula of the mutual conversion of other color spaces:
Rgb space is to the conversion formula of XYZ space as the formula (1):
XYZ space is to the conversion formula of rgb space as the formula (2):
Rgb space is to the conversion formula in HSV space as the formula (3):
Rgb space is to the conversion formula in YCrCb space as the formula (4):
The YCrCb space is to the conversion formula of rgb space as the formula (5):
Can be in the practical application according to concrete needs, utilize above-mentioned the separate equations to carry out color notation conversion space and obtain corresponding single channel half-tone information, and can select to obtain what single channel half-tone informations as required, as selecting to obtain 1 single channel half-tone information under the R passage; Selection obtains 2 single channel half-tone informations (being respectively R passage, G passage or R passage, B passage etc.), but need to prove that the port number of selection is many more, and it is accurate more to obtain scene border.
202: utilize the boundary search algorithm, each the single channel half-tone information in a plurality of single channel half-tone informations is carried out frontier point respectively extract, obtain the divergent boundary point of each single channel half-tone information.???
Wherein, the boundary search algorithm can be Canny Operator Method, filtering 2 value algorithms etc.; The preferred Canny Operator Method that adopts in the present embodiment.
203: obtain a plurality of key points on the scene border, and a plurality of key points are arranged in order.
Wherein, the point of key point for choosing on scene border can obtain by manual demarcation or other the whole bag of tricks; A plurality of key points are arranged in order, can be for a plurality of key points be sorted by the direction that travels through scene border clockwise; Maybe can be for a plurality of key points are sorted by the direction that travels through scene border counterclockwise.
Adopt manual method of demarcating in the present embodiment, on scene border, mark key point; And a plurality of key points are sorted by the direction that travels through the scenery profile clockwise.
204: the divergent boundary point of each the single channel half-tone information that obtains according to step 202, utilize Shortest Path Searching Algorithm, search for and obtain border between per two adjacent key points.
Detailed process is: 2041: according to the divergent boundary point of each single channel half-tone information, calculate the boundary intensity of each pixel of image.
Wherein, for each pixel of image, its boundary intensity is for being identified as the number of active lanes of divergent boundary point under a plurality of passages.For example: for pixel A, it is identified as the divergent boundary point at R passage, G passage and H passage, and then the boundary intensity of pixel A is 3; For pixel B, it is identified as the divergent boundary point at R passage, G passage, H passage and V passage, and then the boundary intensity of pixel B is 4.
2042: determine the boundary search zone between per two adjacent key points.
Wherein, the boundary search zone between two adjacent key points can be to be cornerwise rectangular area with this line of 2, as the boundary search zone between these 2; Also can be to be the center with the mid point with this line of 2, a bigger square area that value is the length of side in this 2 Diff E and the Diff N be as the boundary search zone between these 2; Can also be to be the center of circle with the mid point with this line of 2, a bigger border circular areas that value is a diameter in this 2 Diff E and the Diff N be as the boundary search zone between these 2.Boundary search zone between two adjacent key points is not limited to above-mentioned several, in the practical application two boundary search zones between the adjacent key point can be set flexibly as required.
2043: the boundary intensity according to each pixel of image, utilize Shortest Path Searching Algorithm, in the boundary search zone between per two adjacent key points of determining, search for and obtain efficiency frontier between per two adjacent key points.
Particularly, 2043a: the boundary intensity and the passage threshold value of presetting of each pixel in the image are compared, obtain the efficiency frontier point of image.
Wherein, boundary intensity is regarded as the efficiency frontier point greater than the pixel of default passage threshold value.And the initial value of default passage threshold value can be the value smaller or equal to the number of single channel half-tone information, and for example: if obtain 7 single channel half-tone informations in step 201, then Yu She passage threshold value can be 7,6 or 5 etc.
2043b: carry out the cum rights shortest path through search by the efficiency frontier point, in the boundary search zone between per two adjacent key points of determining, search for and obtain current border between per two adjacent key points.
Wherein, the establishing method of cum rights shortest path weight in search has a lot, establishing in the embodiment of the invention needs the quantity of the pixel of filling to be directly proportional between weight and two the efficiency frontier points, and the angle between the ray formed of the ray of forming with these two efficiency frontier points and this two key points also is directly proportional.The direction of the ray of two efficiency frontier point compositions is that initial frontier point is to the terminal point frontier point, the direction of the ray of two key point compositions is that initial key point is to the terminal point key point, for example: current search be key point A to the border between the key point B, then key point A is to key point B from key point A to the directions of rays between the key point B.
2043c: whether judge ratio between the pixel total in the pixel of filling in the current border and the current border greater than default proportion threshold value,, carry out 2043d when judged result is during greater than default proportion threshold value; Otherwise, carry out 2043f.
Wherein, the pixel of filling in the current border is meant the pixel that does not belong to efficiency frontier point in the current border; Ratio in pixel of filling in the current border and the current border between the total pixel is meant: the quantity of the pixel of filling in the current border shared ratio in the total pixel quantity in current border; Default proportion threshold value can be provided with according to actual needs flexibly, and for example: default proportion threshold value is 40%, 30% etc.Ratio in pixel of filling in the current border and the current border between the total pixel thinks that the efficiency frontier point in this moment of the current border is very few during greater than default proportion threshold value, need search for the border again.
2043d: whether judge default passage threshold value greater than 1,, carry out 2043e when judged result is greater than 1 the time; Otherwise, carry out 2043f.
2043e: default passage threshold value is subtracted 1, then execution in step 2043a.
2043f: will deserve preceding border as and deserve efficiency frontier between two adjacent key points that the front bound pair answers.
205: connect the efficiency frontier between two whenever adjacent key points, obtain the integral edge of scenery.
Need to prove, may comprise a plurality of scenery in the image, as comprising a plurality of personages in the sub-picture, when comprising a plurality of scenery in the sub-picture, only need repeated execution of steps 203-205, the border that obtains each scenery successively gets final product.
The described method of present embodiment by fully utilizing the scene border information under a plurality of single channels and the setting of key point, has been avoided having improved the accuracy of obtaining scene border because the mistake that scenery texture complexity causes is cut apart phenomenon; It is low to obtain in the scene border process artificial participation, has improved the efficient of obtaining scene border, can save the plenty of time, and not need to rely on artificial experience and technology, portable high; And can manually choose key point, manually can intervene, strengthen interactive; Further, the passage threshold value screening frontier point by default when obscure boundary is clear, can reduces the passage threshold value and obtain abundanter boundary information, makes sharpness of border; When boundary information too enriches, can improve the passage threshold value and reduce interference; In addition,, reduced the working strength of manually a plurality of single channel lower boundary points being screened, improved work efficiency, can save the plenty of time by default passage threshold adaptive ground screening frontier point is set.
Embodiment 3
Referring to Fig. 3, the embodiment of the invention provides a kind of device that obtains scene border, and this device comprises:
Particularly, first acquisition module 301 can comprise:
Converter unit is used for image is carried out color notation conversion space, obtains a plurality of single channel half-tone informations;
First search unit, be used for after converter unit obtains a plurality of single channel half-tone informations, utilize the boundary search algorithm, each the single channel half-tone information in a plurality of single channel half-tone informations is carried out boundary search respectively, obtain the divergent boundary point of each single channel half-tone information.
Particularly, search module can comprise:
Computing unit is used for the divergent boundary point of each single channel half-tone information of obtaining according to first acquisition module 301, calculates the boundary intensity of each pixel of image;
Determining unit is used for the boundary search zone between definite per two adjacent key points;
Second search unit, the boundary intensity that is used for each pixel of the image that obtains according to computing unit, utilize Shortest Path Searching Algorithm, in the boundary search zone between per two adjacent key points that determining unit is determined, search for and obtain efficiency frontier between per two adjacent key points.
Particularly, second search unit can comprise:
Subelement relatively is used for the boundary intensity of each pixel of image and default passage threshold value are compared, and obtains the efficiency frontier point of image;
The search subelement, be used for after relatively subelement obtains the efficiency frontier point of image, carry out the cum rights shortest path through search by the efficiency frontier point, in the boundary search zone between per two adjacent key points that determining unit is determined, search for and obtain current border between per two adjacent key points;
First judgment sub-unit is used for after the search subelement obtains current border between per two adjacent key points, judges that whether ratio between the pixel total in the pixel of filling in the current border and the current border is greater than the proportion threshold value of presetting;
First handles subelement, when being used for judged result when first judgment sub-unit and being smaller or equal to default proportion threshold value, with current border as and two adjacent key points of should the front bound pair answering between efficiency frontier.
Further, second search unit can also comprise:
Second judgment sub-unit when being used for judged result when first judgment sub-unit and being greater than default proportion threshold value, judges that whether default passage threshold value is greater than 1;
Second handles subelement, be used for judged result when second judgment sub-unit and be the passage threshold value preset greater than 1 the time, default passage threshold value is subtracted 1, notice relatively subelement execution compares the boundary intensity of each pixel in the image and default passage threshold value, obtains the efficiency frontier point of image.
Further, second search unit also comprises:
The 3rd handles subelement, is used for judged result when second judgment sub-unit and is the passage threshold value preset smaller or equal to 1 the time, with current border as and two adjacent key points of should the front bound pair answering between efficiency frontier.
The described device of present embodiment by fully utilizing the scene border information under a plurality of single channels and the setting of key point, has been avoided having improved the accuracy of obtaining scene border because the mistake that scenery texture complexity causes is cut apart phenomenon; It is low to obtain in the scene border process artificial participation, has improved the efficient of obtaining scene border, can save the plenty of time, and not need to rely on artificial experience and technology, portable high; And can manually choose key point, manually can intervene, strengthen interactive; Further, the passage threshold value screening frontier point by default when obscure boundary is clear, can reduces the passage threshold value and obtain abundanter boundary information, makes sharpness of border; When boundary information too enriches, can improve the passage threshold value and reduce interference; In addition,, reduced the working strength of manually a plurality of single channel lower boundary points being screened, improved work efficiency, can save the plenty of time by default passage threshold adaptive ground screening frontier point is set.
All or part of content in the technical scheme that above embodiment provides can realize that its software program is stored in the storage medium that can read by software programming, storage medium for example: the hard disk in the computing machine, CD or floppy disk.
The above only is preferred embodiment of the present invention, and is in order to restriction the present invention, within the spirit and principles in the present invention not all, any modification of being done, is equal to replacement, improvement etc., all should be included within protection scope of the present invention.
Claims (10)
1. a method of obtaining scene border is characterized in that, comprising:
Obtain the divergent boundary point of each the single channel half-tone information in a plurality of single channel half-tone informations of image;
Obtain a plurality of key points on the scene border, and described a plurality of key points are arranged in order;
According to the divergent boundary point of described each single channel half-tone information, calculate the boundary intensity of each pixel of described image;
Determine the boundary search zone between per two adjacent key points;
Boundary intensity according to each pixel of described image utilizes Shortest Path Searching Algorithm, in the boundary search zone between described per two adjacent key points of determining, searches for and obtain efficiency frontier between described per two adjacent key points;
Connect the efficiency frontier between described per two adjacent key points, obtain the border of scenery.
2. the method for obtaining scene border according to claim 1 is characterized in that, the divergent boundary point of each the single channel half-tone information in the described a plurality of single channel half-tone informations that obtain image comprises:
Described image is carried out color notation conversion space, obtain a plurality of single channel half-tone informations;
Utilize the boundary search algorithm, each the single channel half-tone information in described a plurality of single channel half-tone informations is carried out boundary search respectively, obtain the divergent boundary point of described each single channel half-tone information.
3. the method for obtaining scene border according to claim 1, it is characterized in that, the boundary intensity of described each pixel according to described image, utilize Shortest Path Searching Algorithm, in the boundary search zone between described per two adjacent key points of determining, search also obtains efficiency frontier between described per two adjacent key points, specifically comprises:
The boundary intensity of each pixel of described image and default passage threshold value are compared, obtain the efficiency frontier point of described image;
Carry out the cum rights shortest path through search by described efficiency frontier point, in the boundary search zone between described per two adjacent key points of determining, search for and obtain current border between described per two adjacent key points;
Judge that whether ratio between the pixel total in the pixel of filling in the described current border and the described current border is greater than default proportion threshold value;
When judged result is during smaller or equal to described default proportion threshold value, with described current border as the efficiency frontier between two adjacent key points corresponding with described current border.
4. the method for obtaining scene border according to claim 3 is characterized in that, when judged result is during greater than described default proportion threshold value, also comprises:
Judge that whether described default passage threshold value is greater than 1;
When judged result is described default passage threshold value greater than 1 the time, described default passage threshold value is subtracted 1, carry out described boundary intensity and compare, obtain the step of the efficiency frontier point of described image with the passage threshold value of presetting with each pixel in the described image.
5. the method for obtaining scene border according to claim 4 is characterized in that, when judged result is described default passage threshold value smaller or equal to 1 the time, also comprises:
With described current border as the efficiency frontier between two adjacent key points corresponding with described current border.
6. a device that obtains scene border is characterized in that, comprising:
First acquisition module is used for obtaining the divergent boundary point of each single channel half-tone information of a plurality of single channel half-tone informations of image;
Second acquisition module is used to obtain a plurality of key points on the scene border, and described a plurality of key points are arranged in order;
Search module comprises computing unit, determining unit and second search unit; Described computing unit is used for the divergent boundary point of each single channel half-tone information of obtaining according to described first acquisition module, calculates the boundary intensity of each pixel of described image; Described determining unit is used for the boundary search zone between definite per two adjacent key points; Described second search unit is used for the boundary intensity of each pixel of the described image that obtains according to described computing unit, utilize Shortest Path Searching Algorithm, in the boundary search zone between described per two adjacent key points that described determining unit is determined, search for and obtain efficiency frontier between described per two adjacent key points;
Link block is used for connecting the efficiency frontier between described per two adjacent key points after described search module obtains efficiency frontier between per two adjacent key points, obtains the border of scenery.
7. the device that obtains scene border according to claim 6 is characterized in that, described first acquisition module comprises:
Converter unit is used for described image is carried out color notation conversion space, obtains a plurality of single channel half-tone informations;
First search unit, be used for after described converter unit obtains a plurality of single channel half-tone informations, utilize the boundary search algorithm, each the single channel half-tone information in described a plurality of single channel half-tone informations is carried out boundary search respectively, obtain the divergent boundary point of described each single channel half-tone information.
8. the device that obtains scene border according to claim 6 is characterized in that, described second search unit comprises:
Subelement relatively is used for the boundary intensity of each pixel of described image and default passage threshold value are compared, and obtains the efficiency frontier point of described image;
The search subelement, be used for after described comparison subelement obtains the efficiency frontier point of described image, carry out the cum rights shortest path through search by described efficiency frontier point, in the boundary search zone between described per two adjacent key points that described determining unit is determined, search for and obtain current border between described per two adjacent key points;
First judgment sub-unit, be used for after described search subelement obtains current border between described per two adjacent key points, judge that whether ratio between the pixel total in the pixel of filling in the described current border and the described current border is greater than default proportion threshold value;
First handles subelement, when being used for judged result when described first judgment sub-unit and being smaller or equal to described default proportion threshold value, with described current border as the efficiency frontier between two adjacent key points corresponding with described current border.
9. the device that obtains scene border according to claim 8 is characterized in that, described second search unit also comprises:
Second judgment sub-unit when being used for judged result when described first judgment sub-unit and being greater than described default proportion threshold value, judges that whether described default passage threshold value is greater than 1;
Second handles subelement, be used for judged result when described second judgment sub-unit and be described default passage threshold value greater than 1 the time, described default passage threshold value is subtracted 1, the boundary intensity of notifying described relatively subelement to carry out each pixel in the described image compares with the passage threshold value of presetting, and obtains the efficiency frontier point of image.
10. the device that obtains scene border according to claim 9 is characterized in that, described second search unit also comprises:
The 3rd handles subelement, is used for judged result when described second judgment sub-unit and is described default passage threshold value smaller or equal to 1 the time, with described current border as the efficiency frontier between two adjacent key points corresponding with described current border.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN200910083420XA CN101540050B (en) | 2009-04-30 | 2009-04-30 | Method and device for obtaining scene boundary |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN200910083420XA CN101540050B (en) | 2009-04-30 | 2009-04-30 | Method and device for obtaining scene boundary |
Publications (2)
Publication Number | Publication Date |
---|---|
CN101540050A CN101540050A (en) | 2009-09-23 |
CN101540050B true CN101540050B (en) | 2011-10-26 |
Family
ID=41123226
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN200910083420XA Expired - Fee Related CN101540050B (en) | 2009-04-30 | 2009-04-30 | Method and device for obtaining scene boundary |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN101540050B (en) |
Families Citing this family (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103353991B (en) * | 2013-06-24 | 2016-02-24 | 西安交通大学 | A kind of iterative resolution generation method of profile fragment |
CN105243380A (en) * | 2015-11-18 | 2016-01-13 | 哈尔滨工业大学 | Single facial image recognition method based on combination of selective median filtering and PCA |
CN113159112B (en) * | 2021-03-09 | 2023-04-07 | 西南石油大学 | Automatic labeling method for ancient biological fossil image |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5202928A (en) * | 1988-09-09 | 1993-04-13 | Agency Of Industrial Science And Technology | Surface generation method from boundaries of stereo images |
CN1380543A (en) * | 2001-04-12 | 2002-11-20 | 清华大学 | Image segmentation and identification method in industrial radiation imaging system |
US20060120590A1 (en) * | 2004-12-07 | 2006-06-08 | Lockheed Martin Corporation | Automatic scene correlation and identification |
CN101377850A (en) * | 2008-09-27 | 2009-03-04 | 北京航空航天大学 | Method of multi-formwork image segmentation based on ant colony clustering |
-
2009
- 2009-04-30 CN CN200910083420XA patent/CN101540050B/en not_active Expired - Fee Related
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5202928A (en) * | 1988-09-09 | 1993-04-13 | Agency Of Industrial Science And Technology | Surface generation method from boundaries of stereo images |
CN1380543A (en) * | 2001-04-12 | 2002-11-20 | 清华大学 | Image segmentation and identification method in industrial radiation imaging system |
US20060120590A1 (en) * | 2004-12-07 | 2006-06-08 | Lockheed Martin Corporation | Automatic scene correlation and identification |
CN101377850A (en) * | 2008-09-27 | 2009-03-04 | 北京航空航天大学 | Method of multi-formwork image segmentation based on ant colony clustering |
Non-Patent Citations (1)
Title |
---|
于勇 郭雷.《噪声图像中提取边缘的蚁群搜索算法》.《噪声图像中提取边缘的蚁群搜索算法》.2008,第30卷(第6期), * |
Also Published As
Publication number | Publication date |
---|---|
CN101540050A (en) | 2009-09-23 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US8345998B2 (en) | Compression scheme selection based on image data type and user selections | |
EP1507232B1 (en) | Method for classifying a digital image | |
US8666156B2 (en) | Image-based backgrounds for images | |
CN101122999B (en) | Method for automatically extracting stamp image from Chinese painting and calligraphy | |
CN105761202B (en) | A kind of color image color moving method | |
KR102351725B1 (en) | Color gamut mapping method and color gamut mapping device | |
CN107633252A (en) | Skin color detection method, device and storage medium | |
CN101540050B (en) | Method and device for obtaining scene boundary | |
CN112270653A (en) | Data enhancement method for unbalance of image sample | |
JPH0816789A (en) | Judging method for landscape by means of color | |
CN106415596B (en) | image conversion based on segmentation | |
CN103325101B (en) | The extracting method of color property and device | |
CN101561316A (en) | On-line test visual data processing system based on region of interest (ROI) | |
CN104392443B (en) | Method for detecting main characteristic colors of two-dimensional 24-bit color image | |
Nakajima et al. | A novel color image processing scheme in HSI color space with negative image processing | |
CN103679658A (en) | Image processing method for achieving intelligent decoloring according to dominant hue | |
CN108205812B (en) | Method for matching pigment color mixing proportion | |
CN1220960C (en) | Colour signal processing device and method for efficient storage colour gamut | |
CN103366343A (en) | Bitmap scaling method and system | |
CN106851168A (en) | Video format recognition methods, device and player | |
KR101216578B1 (en) | Method and apparatus for correcting color | |
CN108876800B (en) | Information processing method and equipment | |
JP5933987B2 (en) | Image coding apparatus and control method thereof | |
Wang et al. | Specific color-pair edge detection using quaternion convolution | |
CN103514611B (en) | The extracting method of a kind of area of skin color and device |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20111026 Termination date: 20190430 |
|
CF01 | Termination of patent right due to non-payment of annual fee |