CN105893999A - Method and device for extracting a region of interest - Google Patents
Method and device for extracting a region of interest Download PDFInfo
- Publication number
- CN105893999A CN105893999A CN201610196180.4A CN201610196180A CN105893999A CN 105893999 A CN105893999 A CN 105893999A CN 201610196180 A CN201610196180 A CN 201610196180A CN 105893999 A CN105893999 A CN 105893999A
- Authority
- CN
- China
- Prior art keywords
- pixel
- value
- undulating value
- described target
- target 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
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
- G06V10/255—Detecting or recognising potential candidate objects based on visual cues, e.g. shapes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
- G06V10/443—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components by matching or filtering
- G06V10/449—Biologically inspired filters, e.g. difference of Gaussians [DoG] or Gabor filters
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Multimedia (AREA)
- General Physics & Mathematics (AREA)
- Biodiversity & Conservation Biology (AREA)
- Molecular Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Compression Or Coding Systems Of Tv Signals (AREA)
- Image Analysis (AREA)
Abstract
The embodiment of the invention discloses a method and device for extracting a region of interest (ROI) and relates to the technical field of image processing. The method comprises steps of: acquiring an intensity sensitive parameter of a target image, wherein the intensity sensitive parameter is determined by the textural feature and the eye visual characteristic of the target image; selecting a pixel classification coefficient of the target image according to the intensity sensitive parameter; determining the classification of each pixel in the target image according to the pixel classification coefficient; acquiring the ROI in the target image according to the classification of each pixel in the target image and extracting the determined ROI. The method and the device may fast extract the ROI in the target image.
Description
Technical field
The application relates to technical field of image processing, particularly to a kind of area-of-interest exacting method and device.
Background technology
Along with the fast development of Internet technology and popularizing of individual's intelligent terminal, such as, individual's intelligence hand
Machine, video camera, camera etc., video and image generate and occur in that explosive growth, and user can will regard simultaneously
Frequency or image are shared with other users easily.
During it addition, process video image, the process of video image is mainly manifested on two levels,
One is that image, semantic understands level, and one is signal pixels level.At present, based on video image basis language
Justice and a lot of application of human-eye visual characteristic, e.g., image characteristics extraction, Object Segmentation etc., these application one
As carry out based on area-of-interest, it is therefore desirable to be able to the area-of-interest in rapid extraction image.
Summary of the invention
The embodiment of the present application discloses a kind of area-of-interest exacting method and device, with can rapid extraction figure
Area-of-interest in Xiang.
For reaching above-mentioned purpose, the embodiment of the present application discloses a kind of area-of-interest exacting method, described side
Method includes:
Obtaining the strength sensitive parameter of target image, wherein, described strength sensitive parameter, by described target figure
Textural characteristics and the human-eye visual characteristic of picture determine;
The pixel classification factor of described target image is selected according to described strength sensitive parameter;
The classification of each pixel in described target image is determined according to described pixel classification factor;
Classification according to pixel each in described target image, it is thus achieved that the region of interest in described target image
Territory, and area-of-interest determined by extraction.
In a kind of specific implementation of the application, described determine according to described pixel classification factor described
The classification of each pixel in target image, including:
Determine the classification of each pixel in described target image in such a way:
Calculate described target pixel points undulating value KH in the horizontal direction and undulating value KV vertically, its
In, described target pixel points is any pixel point in described target image, and described undulating value is used for representing
Pixel and about along the change between the pixel value of specific direction pixel;
Judge whether described undulating value KH and described undulating value KV is satisfied by the sentencing for flat site preset
Disconnected criterion, wherein, the described judgment criterion for flat site is: with pin in described pixel classification factor
The criterion relevant to the classification factor of flat site;
If it has, then judge that described target pixel points is as the vegetarian refreshments belonging to flat site picture.
In a kind of specific implementation of the application, judge that described target pixel points is as belonging to smooth described
After the pixel in region, also include:
According in described undulating value KH, described undulating value KV and described pixel classification factor for flat region
The classification factor in territory, determines the subclassification of described target pixel points.
In a kind of specific implementation of the application, learn described undulating value KH and described undulating value in judgement
In the case of KV can not be satisfied by the described default judgment criterion for flat site, also include:
Calculate the described target pixel points undulating value along preset direction, be designated as direction undulating value, wherein, 00< institute
State the angle < 90 between preset direction and horizontal direction0;
The first of value minimum is selected from described undulating value KH, described undulating value KV and described direction undulating value
Minimal ripple value;
Judge whether described first minimal ripple value meets the judgment criterion for direction borderline region preset,
Wherein, the described judgment criterion for direction borderline region is: with in described pixel classification factor for side
To the criterion that the classification factor of borderline region is relevant;
If it has, then judge that described target pixel points is as the pixel belonging to direction borderline region.
In a kind of specific implementation of the application, judge that described target pixel points is as belonging to direction described
After the pixel of borderline region, also include:
Determine that the subclassification of described target pixel points is direction corresponding to described first minimal ripple value.
In a kind of specific implementation of the application, learn that described first minimal ripple value is unsatisfactory in judgement
In the case of the described default judgment criterion for direction borderline region, also include:
Select rule according to default angle point, select the angle point of described target pixel points;
Calculate the undulating value of described angle point, be designated as angle point undulating value;
The second minimal ripple value that value is minimum is selected from described angle point undulating value;
Judge that whether described second minimal ripple value meets the judgment criterion for angle point region preset, wherein,
The described judgment criterion for angle point region is: with dividing for angle point region in described pixel classification factor
The criterion that class coefficient is relevant;
If it has, then judge that described target pixel points is as the pixel belonging to angle point region.
In a kind of specific implementation of the application, judge that described target pixel points is as belonging to angle point described
After the pixel in region, also include:
Determine that the subclassification of described target pixel points is direction corresponding to described second minimal ripple value.
In a kind of specific implementation of the application, learn that described second minimal ripple value is unsatisfactory in judgement
In the case of the judgment criterion for angle point region preset, also include:
Judge that described target pixel points is as pole type pixel.
In a kind of specific implementation of the application, judging that described target pixel points is as pole type pixel
Afterwards, also include:
Calculate the pixel value average of pixel in the default neighborhood region of described target pixel points;
According to described pixel value average and the pixel value of described target pixel points, determine described target pixel points
Subclassification.
In a kind of specific implementation of the application, the strength sensitive parameter of described acquisition target image, bag
Include:
According to default parameter value, it is thus achieved that the strength sensitive parameter of target image;Or
Compression ratio according to target image, it is thus achieved that the strength sensitive parameter of described target image;Or
In the case of many wheel codings, complete to take turns the coding result of coding according to target image more, it is thus achieved that
The strength sensitive parameter of described target image;Or
Coding result according to coding moment with target image immediate predetermined number image, it is thus achieved that described
The strength sensitive parameter of target image.
For reaching above-mentioned purpose, the embodiment of the present application discloses a kind of region of interesting extraction device, described dress
Put and include:
Sensitive parameter obtains module, for obtaining the strength sensitive parameter of target image, wherein, described intensity
Sensitive parameter, textural characteristics and human-eye visual characteristic by described target image determine;
Classification factor selects module, for selecting the pixel of described target image according to described strength sensitive parameter
Point classification factor;
Pixel classification determines module, for determining in described target image according to described pixel classification factor
The classification of each pixel;
Region extraction module, for the classification according to pixel each in described target image, it is thus achieved that described mesh
Area-of-interest in logo image, and area-of-interest determined by extraction.
In a kind of specific implementation of the application, the classification of described pixel determines module, specifically for really
The classification of each pixel in fixed described target image;
The classification of described pixel determines module, including:
First undulating value calculating sub module, for calculating described target pixel points undulating value KH in the horizontal direction
Undulating value KV vertically, wherein, described target pixel points is the arbitrary picture in described target image
Vegetarian refreshments, described undulating value, for representing pixel and about along between the pixel value of specific direction pixel
Change;
First undulating value judges submodule, is used for judging that described undulating value KH and described undulating value KV is the fullest
The judgment criterion for flat site that foot is preset, wherein, the described judgment criterion for flat site is:
To in described pixel classification factor for the criterion that the classification factor of flat site is relevant;
First pixel decision sub-module, for judging that at described first undulating value the judged result of submodule is
In the case of being, it is determined that described target pixel points is to belong to the vegetarian refreshments of flat site picture.
In a kind of specific implementation of the application, described region of interesting extraction device also includes:
First subclassification determines submodule, for judging that described target pixel points is as the picture belonging to flat site
After vegetarian refreshments, according in described undulating value KH, described undulating value KV and described pixel classification factor for
The classification factor of flat site, determines the subclassification of described target pixel points.
In a kind of specific implementation of the application, described region of interesting extraction device also includes:
Second undulating value calculating sub module, for judging that at described first undulating value the judged result of submodule is
In the case of no, calculate the described target pixel points undulating value along preset direction, be designated as direction undulating value, its
In, 00Angle < 90 between preset direction and horizontal direction described in <0;
First undulating value selects submodule, for from described undulating value KH, described undulating value KV and described direction
Undulating value selects the first minimal ripple value that value is minimum;
Second undulating value judges submodule, for judging whether described first minimal ripple value meets the pin preset
Judgment criterion to direction borderline region, wherein, the described judgment criterion for direction borderline region is: with
For the criterion that the classification factor of direction borderline region is relevant in described pixel classification factor;
Second pixel decision sub-module, for judging that at described second undulating value the judged result of submodule is
In the case of being, it is determined that described target pixel points is to belong to the pixel of direction borderline region.
In a kind of specific implementation of the application, described region of interesting extraction device also includes:
Second subclassification determines submodule, for judging that described target pixel points is as belonging to direction borderline region
Pixel after, determine that the subclassification of described target pixel points is side corresponding to described first minimal ripple value
To.
In a kind of specific implementation of the application, described region of interesting extraction device also includes:
Angle point selects submodule, for judging, at described second undulating value, the feelings that the judged result of submodule is no
Under condition, select rule according to default angle point, select the angle point of described target pixel points;
3rd undulating value calculating sub module, for calculating the undulating value of described angle point, is designated as angle point undulating value;
Second undulating value selects submodule, for select from described angle point undulating value value minimum second
Minor swing value;
3rd undulating value judges submodule, for judging whether described second minimal ripple value meets the pin preset
Judgment criterion to angle point region, wherein, the described judgment criterion for angle point region is: with described pixel
For the criterion that the classification factor in angle point region is relevant in some classification factor;
3rd pixel decision sub-module, for the judged result judging submodule at described 3rd undulating value be
In the case of being, it is determined that described target pixel points is to belong to the pixel in angle point region.
In a kind of specific implementation of the application, described region of interesting extraction device also includes:
3rd subclassification determines submodule, for judging that described target pixel points is as the picture belonging to angle point region
After vegetarian refreshments, determine that the subclassification of described target pixel points is direction corresponding to described second minimal ripple value.
In a kind of specific implementation of the application, described region of interesting extraction device also includes:
4th pixel decision sub-module, for the judged result judging submodule at described 3rd undulating value be
In the case of no, it is determined that described target pixel points is pole type pixel.
In a kind of specific implementation of the application, described region of interesting extraction device also includes:
Mean value computation submodule, for after judging that described target pixel points is as pole type pixel, calculates
The pixel value average of pixel in the default neighborhood region of described target pixel points;
4th subclassification determines submodule, for according to described pixel value average and the picture of described target pixel points
Element value, determines the subclassification of described target pixel points.
In a kind of specific implementation of the application, described sensitive parameter obtains module,
Specifically for according to the parameter value preset, it is thus achieved that the strength sensitive parameter of target image;Or
Specifically for the compression ratio according to target image, it is thus achieved that the strength sensitive parameter of described target image;Or
Specifically in the case of many wheel codings, tie according to the coding completing to take turns coding of target image more
Really, it is thus achieved that the strength sensitive parameter of described target image;Or
Specifically for the coding result according to coding moment with target image immediate predetermined number image,
Obtain the strength sensitive parameter of described target image.
As seen from the above, in the scheme that the embodiment of the present application provides, the strength sensitive of target image is first obtained
Parameter, then selects the pixel classification factor of target image according to strength sensitive parameter, and according to pixel
Classification factor determines the classification of each pixel in target image, afterwards further according to pixel each in target image
The classification of point, it is thus achieved that the area-of-interest in target image, and area-of-interest determined by extraction.This Shen
The scheme that theres is provided of embodiment please extracted before area-of-interest, owing to having determined that in target image
The classification of each pixel, it is possible to the area-of-interest in rapid extraction image.
Accompanying drawing explanation
In order to be illustrated more clearly that the embodiment of the present application or technical scheme of the prior art, below will be to enforcement
In example or description of the prior art, the required accompanying drawing used is briefly described, it should be apparent that, describe below
In accompanying drawing be only some embodiments of the application, for those of ordinary skill in the art, do not paying
On the premise of going out creative work, it is also possible to obtain other accompanying drawing according to these accompanying drawings.
The schematic flow sheet of a kind of area-of-interest exacting method that Fig. 1 provides for the embodiment of the present application;
Fig. 2 provide for the embodiment of the present application the first determine the schematic flow sheet of method that pixel classifies;
A kind of image slices vegetarian refreshments classifying and dividing schematic diagram that Fig. 3 provides for the embodiment of the present application;
Fig. 4 determines the schematic flow sheet of method that pixel classifies for the second that the embodiment of the present application provides;
Fig. 5 provide for the embodiment of the present application the third determine the schematic flow sheet of method that pixel classifies;
There is provided for the embodiment of the present application the 4th kind of Fig. 6 determines the schematic flow sheet of the method that pixel classifies;
The structural representation of a kind of region of interesting extraction device that Fig. 7 provides for the embodiment of the present application;
A kind of structural representation determining device that pixel classifies that Fig. 8 provides for the embodiment of the present application.
Detailed description of the invention
Below in conjunction with the accompanying drawing in the embodiment of the present application, the technical scheme in the embodiment of the present application is carried out clearly
Chu, be fully described by, it is clear that described embodiment be only some embodiments of the present application rather than
Whole embodiments.Based on the embodiment in the application, those of ordinary skill in the art are not making creation
Property work premise under the every other embodiment that obtained, broadly fall into the scope of the application protection.
When processing image, the process of image is mainly manifested on two levels, and one is image, semantic
Understanding level, one is signal pixels level.It is currently based on image basis semantic with human-eye visual characteristic very
Many application, e.g., image characteristics extraction, Object Segmentation etc., these application are substantially based on area-of-interest
Carry out, accordingly, it would be desirable to be able to the area-of-interest of rapid extraction video image.
Such as, when image is carried out image enhancement processing, need quickly to navigate to image boundary profile, front
Scape and background.On signal transacting level, complete video processing chain generally comprises: the collection of video image,
Pre-process, encode, transmit, post-process and render and the Pixel-level such as display processes, whole process, video
Image is it is possible that various distortion, e.g., and object boundary and the breaking strain of profile, object boundary and angle point
Fuzzy distortion, false contouring and ringing effect, blocking effect, region blur, color distortion etc..Various psychology
Experiment shows: the distortion of major part human eye opposite side circle is most sensitive.From the point of view of physiological angle, human eye is felt most
Interest is the border of object and profile and angle point, its corresponding human brain visual centre V1 district;And for structure
Changing flat site, user is then not intended to the inside and introduces lofty noise and false contouring.Therefore, quickly image
These regions with different human eye interest level such as border, angle point, flat site, limit are extracted respectively
Out, image can be carried out various process and guidance and help is provided.Such as, if can rapid extraction object
Border, angle point and flat site, then can improve video image pretreatment and post processing (as filtering
Strengthen, sharpen, segmentation etc.) and compression coding strategy, can adaptive holding image boundary and structure special
Levy, the noise within structured region is suppressed simultaneously, thus improve the final mass etc. of video image
Deng.
Based on above-mentioned situation, the embodiment of the present application provides a kind of area-of-interest exacting method and device, under
Face is explained by specific embodiment.
The schematic flow sheet of a kind of area-of-interest exacting method that Fig. 1 provides for the embodiment of the present application, the method
Including:
S101: obtain the strength sensitive parameter of target image.
Above-mentioned target image can be the two field picture in video, it is also possible to be an independent figure in non-video
Picture, this is not defined by the application.
Above-mentioned strength sensitive parameter is the textural characteristics by target image and human-eye visual characteristic determines.
Such as, the value of strength sensitive parameter can be { of 0,1,2,3,4}, table in this case
Showing five grades of existence, the application simply illustrates as example, strength sensitive parameter in actual application
Value is not limited to that.
In actual application, series of parameters can be pre-set according to human-eye visual characteristic etc., extract target
During the area-of-interest of image, the gain of parameter strength sensitive parameter that can pre-set according to these.Specifically
, can be according to default parameter value, it is thus achieved that the strength sensitive parameter of target image.
It should be understood that in the case of to coding standard or form, the video image of same resolution ratio,
File after its compression is the biggest, i.e. compression ratio is the least, illustrates that the detailed information comprised in this video image is got over
Many, in this case, bigger strength sensitive parameter need to be used, so that the discrimination of pixel classifications is more preferable.
Given this in the optional implementation of one of the application, can be according to the compression ratio of target image, it is thus achieved that mesh
The strength sensitive parameter of logo image, wherein, compression ratio is it is to be understood that the file of compression rear video image is big
Ratio between little and video image resolution ratio.
It addition, when carrying out Video coding, it may be desirable to carry out taking turns coding to same video image more, with
Reach preferably compression ratio, wherein, each take turns coding after all can obtain peak value to-noise ratio PSNR, SSIM etc.
Coding result, wherein, SSIM is a kind of New Set weighing two width image similarities, and its value is the biggest, explanation
The similarity of two width images is the highest, and the maximum of SSIM is 1.
It addition, as a example by PSNR, PSNR is the lowest, illustrate that the detailed information comprised in video image is the most,
In this case bigger strength sensitive parameter can be selected.
In consideration of it, in the optional implementation of the one of the application, in the case of many wheel codings, can basis
Target image complete to take turns the coding result of coding more, it is thus achieved that the strength sensitive parameter of target image, wherein,
Coding result can include PSNR, SSIM etc..
Concrete, can be the mean value of the PSNR obtained according to many wheel codings, it is thus achieved that the intensity of target image
Sensitive parameter.
Such as, representing the strength sensitive parameter of target image with f1, the value of f1 can be provided that
During PSNR >=42db, f1=0;
During 39db≤PSNR < 42db, f1=1;
During 36db≤PSNR < 39db, f1=2;
During 33db≤PSNR < 36db, f1=3;
During PSNR >=31db, f1=4.
It should be noted that when obtaining the strength sensitive parameter of target image, above-mentioned many except considering
Outside the PSNR of wheel coding situation, it is also contemplated that the image resolution ratio of target image and average quantisation parameter etc.
Deng, wherein, resolution ratio hour, can suitably tune up the value of strength sensitive parameter.
Those skilled in the art are it is understood that have relativity of time domain between adjacent image in video
And spatial correlation, often the coding result between adjacent image is similar, so, the one in the application can
Select in implementation, it is also possible to according to the volume in coding moment with target image immediate predetermined number image
Code result, it is thus achieved that the strength sensitive parameter of target image.
Wherein, predetermined number can determine in actual application, for example, it is possible to be set to 1,2,3 etc..
S102: select the pixel classification factor of target image according to strength sensitive parameter.
Owing to different strength sensitive parameters illustrates image, there is different textural characteristics, so, Ke Yiwei
Different strength sensitive parameters arranges different pixel classification factor.
Concrete, pixel classification factor corresponding to strength sensitive parameter can be set in advance.
Concrete, actual application can be stored in organizing pixel classification factor in a table more.
S103: determine the classification of each pixel in target image according to pixel classification factor.
Specifically determine that the method that pixel is classified is referred to the scheme that FIG. 2 below-embodiment illustrated in fig. 6 provides,
Here wouldn't describe in detail.
S104: according to the classification of pixel each in target image, it is thus achieved that the area-of-interest in target image,
And area-of-interest determined by extracting.
When determining the classification of each pixel in target image according to pixel classification factor, it is thus necessary to determine that target
In image, the classification of each pixel, concrete, sees Fig. 2, Fig. 2 and provides the first and determine that pixel divides
The schematic flow sheet of the method for class, the method includes:
S201: calculate target pixel points undulating value KH in the horizontal direction and undulating value KV vertically.
Wherein, any pixel point during target pixel points is target image.Above-mentioned undulating value, is used for representing picture
Vegetarian refreshments and about along the change between the pixel value of specific direction pixel.
Above-mentioned specific direction can be horizontally oriented, vertical direction and direction that horizontal direction angle is 45 degree
Etc., this is not defined by the application.
The surrounding pixel point of one pixel can be understood as the pixel adjacent with this pixel, it is also possible to
It it is the pixel in preset range of the distance between this pixel.
Concrete, above-mentioned undulating value can according to target pixel points with its cycle along specific direction pixel
Quadratic sum (MSE) between pixel value, all sides of difference and, variance etc. calculated, the application is not
This is defined.
When calculating undulating value, the quantity of selected surrounding pixel point can determine as the case may be,
For example, it is possible to select 2,4 or more etc. can be selected.
In a kind of specific implementation of the application, see Fig. 3, in Fig. 3 as a example by pixel C, its fluctuation
Value KH and undulating value KV can be calculated by following formula.
KH=(pC-pC_left)2+(pC-pC_right)2
KV=(pC-pC_up)2+(pC-pC_down)2
Wherein, pCRepresent the pixel value of pixel C, pC_left、pC_right、pC_up、pC_downRespectively it is positioned at picture
On the left of vegetarian refreshments C, right side, upside, downside and the pixel value of adjacent pixel, specifically, can be found in Fig. 3
Four gray pixels points around pixel C.
It addition, two gray pixels points adjacent with pixel D in Fig. 3 may be used for calculating pixel D level
The undulating value in direction, two the gray pixels points adjacent with pixel E may be used for calculating pixel E Vertical Square
To undulating value.
During it should be noted that calculate the undulating value of the undulating value of pixel horizontal direction and vertical direction, institute
The surrounding pixel point selected is not limited in 2, it is also possible to be 4,6 etc., but institute under normal circumstances
These pixels selected are symmetrically distributions on the basis of target pixel points.
S202: judge whether undulating value KH and undulating value KV is satisfied by the judgement for flat site preset
Criterion, if it is, perform S203.
Owing to there may be polytype region in image, such as, flat site, borderline region, angle point
Region etc., these regions are respectively provided with different features, so, permissible in above-mentioned pixel classification factor
Exist for the classification factor of above-mentioned zones of different.
Wherein, the above-mentioned judgment criterion for flat site is: with in pixel classification factor for flat region
The criterion that the classification factor in territory is relevant.
In the optional implementation of one of the application, the above-mentioned judgment criterion for flat site may is that
Judge whether undulating value KH is less than the first classification factor, and whether undulating value KV is less than the second classification system
Number, if undulating value KH is less than the second classification factor less than the first classification factor and undulating value KV, then judges fluctuation
Value KH and undulating value KV are satisfied by the judgment criterion for flat site preset.
Wherein, above-mentioned first classification factor and the second classification factor are for flat region in pixel classification factor
The classification factor in territory, both values can be equal, it is also possible to is unequal.Concrete, above-mentioned
First classification factor and the second classification factor can be 120,150 etc., and its concrete value can be by research and development people
Member determines according to concrete applicable cases.
S203: judge that target pixel points is as the vegetarian refreshments belonging to flat site picture.
It should be understood that in the case of being the most all flat site, some area flatness is higher, can claim
And strong flat site, and some area flatness is relatively low, can be referred to as weak flat site.In consideration of it,
In a kind of specific implementation of the application, judge target pixel points as belong to flat site pixel it
After, it is also possible to according to dividing for flat site in undulating value KH, undulating value KV and pixel classification factor
Class coefficient, determines the subclassification of target pixel points.
Such as, calculated undulating value KH and undulating value KV is respectively less than 120, it is believed that target pixel points
For belonging to the pixel of flat site, and its subclassification is the pixel belonging to strong flat site;
Calculated undulating value KH and undulating value KV is respectively less than 150, it is believed that target pixel points is for belonging to
The pixel of flat site, and its subclassification is the pixel belonging to weak flat site.
In a kind of specific implementation of the application, see Fig. 4, it is provided that the second determines that pixel is classified
The schematic flow sheet of method, compared with previous embodiment, in the present embodiment, judge to learn fluctuation at S202
In the case of value KH and undulating value KV can not be satisfied by the judgment criterion for flat site preset, also wrap
Include:
S204: calculating target pixel points, along the undulating value of preset direction, is designated as direction undulating value.
Wherein, 00Angle < 90 between < preset direction and horizontal direction0。
Due to the particularity of pixel arrangement, target pixel points and pixel about are formed under normal circumstances
Direction is fixing, including: horizontal direction, vertical direction, angle becomes 450Direction, so, the present embodiment
In, above-mentioned preset direction may be generally understood to: is 45 with the angle of horizontal direction0Direction.
Concrete, can be found in pixel F and the direction that two gray pixels points are formed about, pixel in Fig. 3
Point G and the direction that two gray pixels points are formed about.
Certainly, when calculating undulating value, selected surrounding pixel point is not limited in above-mentioned pixel F and pixel
Gray pixels point around some G, it is also possible to extend to bilateral symmetry along the direction in Fig. 3 and select more pixels
Point.
The mode of calculated direction undulating value can be similar to the mode of aforementioned calculating undulating value KH and undulating value KV,
Here repeat no more.
S205: select value minimum from undulating value KH, undulating value KV and direction undulating value first is minimum
Undulating value.
S206: judge whether the first minimal ripple value meets the judgment criterion for direction borderline region preset,
If it is, perform S207.
Wherein, the judgment criterion for direction borderline region is: with in pixel classification factor for limit, direction
The criterion that the classification factor of boundary region is relevant.
Concrete, the judgment criterion for direction borderline region can be to judge that whether undulating value is less than the 3rd point
Class coefficient, if being less than, then judges to meet the decision criteria for direction borderline region.
Wherein, above-mentioned 3rd classification factor is the classification system in pixel classification factor for direction borderline region
Number, its value can be 150 etc..
S207: judge that target pixel points is as the pixel belonging to direction borderline region.
It should be understood that the border belonging to its correspondence of pixel of direction borderline region there may be multiple direction,
In consideration of it, for ensureing the classification identifying pixel definitely, in a kind of relatively good implementation of the application,
After judging that target pixel points is as the pixel belonging to direction borderline region, it is also possible to determine target pixel points
Subclassification be direction corresponding to the first minimal ripple value.
In a kind of specific implementation of the application, see Fig. 5, it is provided that the third determines that pixel is classified
The schematic flow sheet of method, compared with previous embodiment, in the present embodiment, judge to learn first at S206
In the case of minimal ripple value is unsatisfactory for the judgment criterion for direction borderline region preset, also include:
S208: select rule according to default angle point, selects the angle point of target pixel points.
Four pixels of surrounding that can select a pixel are its angle point, concrete, see the pixel in Fig. 3
Point H, I, J, K can be as the angle points of target pixel points, and wherein, target pixel points is and these four pixels
Equidistant pixel between point.
S209: calculate the undulating value of above-mentioned angle point, is designated as angle point undulating value.
The mode calculating angle point undulating value can be similar to the mode of aforementioned calculating undulating value KH and undulating value KV,
Here repeat no more.
S210: select the second minimal ripple value that value is minimum from angle point undulating value.
S211: judge whether the second minimal ripple value meets the judgment criterion for angle point region preset, if
It is yes, performs S212.
Wherein, the judgment criterion for angle point region is: with in pixel classification factor for angle point region
The criterion that classification factor is relevant.
Concrete, the judgment criterion for angle point region can be to judge that whether undulating value is less than the 4th classification system
Number, if being less than, then judges to meet the decision criteria for angle point region.
Wherein, above-mentioned 4th classification factor is the classification factor in pixel classification factor for angle point region,
Its value can be 200 etc..
S212: judge that target pixel points is as the pixel belonging to angle point region.
In a kind of relatively good implementation of the application, judging that target pixel points is as the picture belonging to angle point region
After vegetarian refreshments, it is also possible to determine that the subclassification of target pixel points is direction corresponding to the second minimal ripple value.
In a kind of specific implementation of the application, see Fig. 6, it is provided that the 4th kind determines that pixel is classified
The schematic flow sheet of method, compared with previous embodiment, in the present embodiment, judge to learn second at S211
In the case of minimal ripple value is unsatisfactory for the judgment criterion for angle point region preset, also include:
S213: judge that target pixel points is as pole type pixel.
In a kind of relatively good implementation of the application, after judging that target pixel points is as pole type pixel,
Can also calculate the pixel value average of pixel in the default neighborhood region of target pixel points further, and according to
Pixel value average and the pixel value of target pixel points, determine the subclassification of target pixel points.
Concrete, above-mentioned default areas can be the neighborhood region of 3x3, it is also possible to be the neighborhood district of 5x5
Pixel C and the region of four gray pixels point compositions about etc. in territory, Fig. 3, the application is not
This is defined.
Between above-mentioned pixel value average and the pixel value of target pixel points when differing greatly, be greater than 30
Deng, then explanation target pixel points is stronger limit, can be referred to as strong limit, such as, white noise point,
Particular point etc. easily causes the point of vision attention;
When difference between above-mentioned pixel value average and the pixel value of target pixel points is less, e.g., less than etc.
In 30 etc., then explanation target pixel points is more weak limit, can be referred to as weak limit.
It addition, for ensureing that being pointed to the borderline pixel of image surrounding smoothly classifies, the one of the application
Plant in optional implementation, it is also possible to determining each pixel in target image according to pixel classification factor
Classification before, according to default extension rule, the external boundary of target image is carried out image spreading, obtains
Target image after extension, target image the most after expansion carries out aforesaid operations.
Concrete, above-mentioned default extension rule can be to extend a pixel column, example on the top circle of image
As, the pixel column at pixel A place in Fig. 3, the value of each pixel target front with extension in this pixel column
In image, in the first pixel column, the value of each pixel is equal;
The leftmost border of image extends a pixel column, such as, the pixel column at pixel B place in figure, should
The taking of each pixel in first pixel column in target image before the value of each pixel and extension in pixel column
It is worth equal.
Certainly, the application simply illustrates as a example by above-mentioned, and the extension rule preset in actual application is not
It is only limitted to this.
As seen from the above, in the scheme that each embodiment above-mentioned provides, the intensity first obtaining target image is quick
Sense parameter, then selects the pixel classification factor of target image according to strength sensitive parameter, and according to pixel
Point classification factor determines the classification of each pixel in target image, afterwards further according to picture each in target image
The classification of vegetarian refreshments, it is thus achieved that the area-of-interest in target image, and area-of-interest determined by extraction.On
State in the scheme that each embodiment provides, before extracting area-of-interest, owing to having determined that target figure
The classification of each pixel in Xiang, it is possible to the area-of-interest in rapid extraction image.
Corresponding with above-mentioned area-of-interest exacting method, the embodiment of the present application additionally provides a kind of region of interest
Territory extraction element.
The structural representation of a kind of region of interesting extraction device that Fig. 7 provides for the embodiment of the present application, this device
Including:
Sensitive parameter obtains module 701, for obtaining the strength sensitive parameter of target image, wherein, described by force
Degree sensitive parameter, textural characteristics and human-eye visual characteristic by described target image determine;
Classification factor selects module 702, for selecting the picture of described target image according to described strength sensitive parameter
Vegetarian refreshments classification factor;
Pixel classification determines module 703, for determining described target image according to described pixel classification factor
In the classification of each pixel;
Region extraction module 704, for the classification according to pixel each in described target image, it is thus achieved that described
Area-of-interest in target image, and area-of-interest determined by extraction.
Concrete, the classification of described pixel determines module, specifically for determining each picture in described target image
The classification of vegetarian refreshments.See a kind of device determining that pixel is classified that Fig. 8, Fig. 8 provide for the embodiment of the present application
Structural representation, in the present embodiment, described pixel classification determine module 703, including:
First undulating value calculating sub module 703A, for calculating the fluctuation in the horizontal direction of described target pixel points
Value KH and undulating value KV vertically, wherein, described target pixel points is appointing in described target image
One pixel, described undulating value, for representing pixel and about along the pixel value of specific direction pixel
Between change;
First undulating value judges submodule 703B, is used for judging that described undulating value KH and described undulating value KV is
The no judgment criterion for flat site being satisfied by presetting, wherein, the described judgement for flat site is accurate
Be then: in described pixel classification factor for the criterion that the classification factor of flat site is relevant;
First pixel decision sub-module 703C, for judging sentencing of submodule 703B at described first undulating value
In the case of disconnected result is for being, it is determined that described target pixel points is to belong to the vegetarian refreshments of flat site picture.
It is also preferred that the left the described device determining that pixel is classified can also include:
First subclassification determines submodule, for judging that described target pixel points is as the picture belonging to flat site
After vegetarian refreshments, according in described undulating value KH, described undulating value KV and described pixel classification factor for
The classification factor of flat site, determines the subclassification of described target pixel points.
Concrete, the described device determining that pixel is classified can also include:
Second undulating value calculating sub module, for judging the judgement knot of submodule 703B at described first undulating value
Fruit be no in the case of, calculate the described target pixel points undulating value along preset direction, be designated as direction undulating value,
Wherein, 00Angle < 90 between preset direction and horizontal direction described in <0;
First undulating value selects submodule, for from described undulating value KH, described undulating value KV and described direction
Undulating value selects the first minimal ripple value that value is minimum;
Second undulating value judges submodule, for judging whether described first minimal ripple value meets the pin preset
Judgment criterion to direction borderline region, wherein, the described judgment criterion for direction borderline region is: with
For the criterion that the classification factor of direction borderline region is relevant in described pixel classification factor;
Second pixel decision sub-module, for judging that at described second undulating value the judged result of submodule is
In the case of being, it is determined that described target pixel points is to belong to the pixel of direction borderline region.
It is also preferred that the left the described device determining that pixel is classified can also include:
Second subclassification determines submodule, for judging that described target pixel points is as belonging to direction borderline region
Pixel after, determine the direction that the first minimal ripple value described in the subclassification of described target pixel points is corresponding.
Concrete, the described device determining that pixel is classified can also include:
Angle point selects submodule, for judging, at described second undulating value, the feelings that the judged result of submodule is no
Under condition, select rule according to default angle point, select the angle point of described target pixel points;
3rd undulating value calculating sub module, for calculating the undulating value of described angle point, is designated as angle point undulating value;
Second undulating value selects submodule, for select from described angle point undulating value value minimum second
Minor swing value;
3rd undulating value judges submodule, for judging whether described second minimal ripple value meets the pin preset
Judgment criterion to angle point region, wherein, the described judgment criterion for angle point region is: with described pixel
For the criterion that the classification factor in angle point region is relevant in some classification factor;
3rd pixel decision sub-module, for the judged result judging submodule at described 3rd undulating value be
In the case of being, it is determined that described target pixel points is to belong to the pixel in angle point region.
It is also preferred that the left the described device determining that pixel is classified can also include:
3rd subclassification determines submodule, for judging that described target pixel points is as the picture belonging to angle point region
After vegetarian refreshments, determine the direction that the second minimal ripple value described in the subclassification of described target pixel points is corresponding.
Concrete, the described device determining that pixel is classified can also include:
4th pixel decision sub-module, for the judged result judging submodule at described 3rd undulating value be
In the case of no, it is determined that described target pixel points is pole type pixel.
It is also preferred that the left the described device determining that pixel is classified can also include:
Mean value computation submodule, for after judging that described target pixel points is as pole type pixel, calculates
The pixel value average of pixel in the default neighborhood region of described target pixel points;
4th subclassification determines submodule, for according to described pixel value all with the pixel of described target pixel points
Value, determines the subclassification of described target pixel points.
Concrete, described sensitive parameter obtains module 701,
Specifically for according to the parameter value preset, it is thus achieved that the strength sensitive parameter of target image;Or
Specifically for the compression ratio according to target image, it is thus achieved that the strength sensitive parameter of described target image;Or
Specifically in the case of many wheel codings, tie according to the coding completing to take turns coding of target image more
Really, it is thus achieved that the strength sensitive parameter of described target image;Or
Specifically for the coding result according to coding moment with target image immediate predetermined number image,
Obtain the strength sensitive parameter of described target image.
As seen from the above, in the scheme that each embodiment above-mentioned provides, the intensity first obtaining target image is quick
Sense parameter, then selects the pixel classification factor of target image according to strength sensitive parameter, and according to pixel
Point classification factor determines the classification of each pixel in target image, afterwards further according to picture each in target image
The classification of vegetarian refreshments, it is thus achieved that the area-of-interest in target image, and area-of-interest determined by extraction.On
State in the scheme that each embodiment provides, before extracting area-of-interest, owing to having determined that target figure
The classification of each pixel in Xiang, it is possible to the area-of-interest in rapid extraction image.
For device embodiment, owing to it is substantially similar to embodiment of the method, so the comparison described is simple
Single, relevant part sees the part of embodiment of the method and illustrates.
It should be noted that in this article, the relational terms of such as first and second or the like be used merely to by
One entity or operation separate with another entity or operating space, and not necessarily require or imply these
Relation or the order of any this reality is there is between entity or operation.And, term " includes ", " comprising "
Or its any other variant is intended to comprising of nonexcludability, so that include the mistake of a series of key element
Journey, method, article or equipment not only include those key elements, but also other including being not expressly set out
Key element, or also include the key element intrinsic for this process, method, article or equipment.Do not having
In the case of more restrictions, statement " including ... " key element limited, it is not excluded that including described wanting
Process, method, article or the equipment of element there is also other identical element.
One of ordinary skill in the art will appreciate that all or part of step realizing in said method embodiment
The program that can be by completes to instruct relevant hardware, and described program can be stored in computer-readable
Take in storage medium, the storage medium obtained designated herein, such as: ROM/RAM, magnetic disc, CD etc..
The foregoing is only the preferred embodiment of the application, be not intended to limit the protection domain of the application.
All any modification, equivalent substitution and improvement etc. made within spirit herein and principle, are all contained in
In the protection domain of the application.
Claims (20)
1. an area-of-interest exacting method, it is characterised in that described method includes:
Obtaining the strength sensitive parameter of target image, wherein, described strength sensitive parameter, by described target figure
Textural characteristics and the human-eye visual characteristic of picture determine;
The pixel classification factor of described target image is selected according to described strength sensitive parameter;
The classification of each pixel in described target image is determined according to described pixel classification factor;
Classification according to pixel each in described target image, it is thus achieved that the region of interest in described target image
Territory, and area-of-interest determined by extraction.
Method the most according to claim 1, it is characterised in that described according to described pixel classification system
Number determines the classification of each pixel in described target image, including:
Determine the classification of each pixel in described target image in such a way:
Calculate described target pixel points undulating value KH in the horizontal direction and undulating value KV vertically, its
In, described target pixel points is any pixel point in described target image, and described undulating value is used for representing
Pixel and about along the change between the pixel value of specific direction pixel;
Judge whether described undulating value KH and described undulating value KV is satisfied by the sentencing for flat site preset
Disconnected criterion, wherein, the described judgment criterion for flat site is: with pin in described pixel classification factor
The criterion relevant to the classification factor of flat site;
If it has, then judge that described target pixel points is as the vegetarian refreshments belonging to flat site picture.
Method the most according to claim 2, it is characterised in that in the described target pixel points of described judgement
For belong to flat site pixel after, also include:
According in described undulating value KH, described undulating value KV and described pixel classification factor for flat region
The classification factor in territory, determines the subclassification of described target pixel points.
Method the most according to claim 2, it is characterised in that judge learn described undulating value KH and
In the case of described undulating value KV can not be satisfied by the described default judgment criterion for flat site, also wrap
Include:
Calculate the described target pixel points undulating value along preset direction, be designated as direction undulating value, wherein, 0 ° of < institute
State the angle < 90 ° between preset direction and horizontal direction;
The first of value minimum is selected from described undulating value KH, described undulating value KV and described direction undulating value
Minimal ripple value;
Judge whether described first minimal ripple value meets the judgment criterion for direction borderline region preset,
Wherein, the described judgment criterion for direction borderline region is: with in described pixel classification factor for side
To the criterion that the classification factor of borderline region is relevant;
If it has, then judge that described target pixel points is as the pixel belonging to direction borderline region.
Method the most according to claim 4, it is characterised in that in the described target pixel points of described judgement
For belong to direction borderline region pixel after, also include:
Determine that the subclassification of described target pixel points is direction corresponding to described first minimal ripple value.
Method the most according to claim 4, it is characterised in that learn the described first small echo judging
In the case of dynamic value is unsatisfactory for the described default judgment criterion for direction borderline region, also include:
Select rule according to default angle point, select the angle point of described target pixel points;
Calculate the undulating value of described angle point, be designated as angle point undulating value;
The second minimal ripple value that value is minimum is selected from described angle point undulating value;
Judge that whether described second minimal ripple value meets the judgment criterion for angle point region preset, wherein,
The described judgment criterion for angle point region is: with dividing for angle point region in described pixel classification factor
The criterion that class coefficient is relevant;
If it has, then judge that described target pixel points is as the pixel belonging to angle point region.
Method the most according to claim 6, it is characterised in that in the described target pixel points of described judgement
For belong to angle point region pixel after, also include:
Determine that the subclassification of described target pixel points is direction corresponding to described second minimal ripple value.
Method the most according to claim 6, it is characterised in that learn the described second small echo judging
In the case of dynamic value is unsatisfactory for the judgment criterion for angle point region preset, also include:
Judge that described target pixel points is as pole type pixel.
Method the most according to claim 8, it is characterised in that judging that described target pixel points is as pole
After point-type pixel, also include:
Calculate the pixel value average of pixel in the default neighborhood region of described target pixel points;
According to described pixel value average and the pixel value of described target pixel points, determine described target pixel points
Subclassification.
10. according to the method according to any one of claim 1-9, it is characterised in that described acquisition target figure
The strength sensitive parameter of picture, including:
According to default parameter value, it is thus achieved that the strength sensitive parameter of target image;Or
Compression ratio according to target image, it is thus achieved that the strength sensitive parameter of described target image;Or
In the case of many wheel codings, complete to take turns the coding result of coding according to target image more, it is thus achieved that
The strength sensitive parameter of described target image;Or
Coding result according to coding moment with target image immediate predetermined number image, it is thus achieved that described
The strength sensitive parameter of target image.
11. 1 kinds of region of interesting extraction devices, it is characterised in that described device includes:
Sensitive parameter obtains module, for obtaining the strength sensitive parameter of target image, wherein, described intensity
Sensitive parameter, textural characteristics and human-eye visual characteristic by described target image determine;
Classification factor selects module, for selecting the pixel of described target image according to described strength sensitive parameter
Point classification factor;
Pixel classification determines module, for determining in described target image according to described pixel classification factor
The classification of each pixel;
Region extraction module, for the classification according to pixel each in described target image, it is thus achieved that described mesh
Area-of-interest in logo image, and area-of-interest determined by extraction.
12. devices according to claim 11, it is characterised in that the classification of described pixel determines module,
Specifically for determining the classification of each pixel in described target image;
The classification of described pixel determines module, including:
First undulating value calculating sub module, for calculating described target pixel points undulating value KH in the horizontal direction
Undulating value KV vertically, wherein, described target pixel points is the arbitrary picture in described target image
Vegetarian refreshments, described undulating value, for representing pixel and about along between the pixel value of specific direction pixel
Change;
First undulating value judges submodule, is used for judging that described undulating value KH and described undulating value KV is the fullest
The judgment criterion for flat site that foot is preset, wherein, the described judgment criterion for flat site is:
To in described pixel classification factor for the criterion that the classification factor of flat site is relevant;
First pixel decision sub-module, for judging that at described first undulating value the judged result of submodule is
In the case of being, it is determined that described target pixel points is to belong to the vegetarian refreshments of flat site picture.
13. devices according to claim 12, it is characterised in that described device also includes:
First subclassification determines submodule, for judging that described target pixel points is as the picture belonging to flat site
After vegetarian refreshments, according in described undulating value KH, described undulating value KV and described pixel classification factor for
The classification factor of flat site, determines the subclassification of described target pixel points.
14. devices according to claim 12, it is characterised in that described device also includes:
Second undulating value calculating sub module, for judging that at described first undulating value the judged result of submodule is
In the case of no, calculate the described target pixel points undulating value along preset direction, be designated as direction undulating value, its
In, angle < 90 ° between preset direction and horizontal direction described in 0 ° of <;
First undulating value selects submodule, for from described undulating value KH, described undulating value KV and described direction
Undulating value selects the first minimal ripple value that value is minimum;
Second undulating value judges submodule, for judging whether described first minimal ripple value meets the pin preset
Judgment criterion to direction borderline region, wherein, the described judgment criterion for direction borderline region is: with
For the criterion that the classification factor of direction borderline region is relevant in described pixel classification factor;
Second pixel decision sub-module, for judging that at described second undulating value the judged result of submodule is
In the case of being, it is determined that described target pixel points is to belong to the pixel of direction borderline region.
15. devices according to claim 14, it is characterised in that described device also includes:
Second subclassification determines submodule, for judging that described target pixel points is as belonging to direction borderline region
Pixel after, determine that the subclassification of described target pixel points is side corresponding to described first minimal ripple value
To.
16. devices according to claim 14, it is characterised in that described device also includes:
Angle point selects submodule, for judging, at described second undulating value, the feelings that the judged result of submodule is no
Under condition, select rule according to default angle point, select the angle point of described target pixel points;
3rd undulating value calculating sub module, for calculating the undulating value of described angle point, is designated as angle point undulating value;
Second undulating value selects submodule, for select from described angle point undulating value value minimum second
Minor swing value;
3rd undulating value judges submodule, for judging whether described second minimal ripple value meets the pin preset
Judgment criterion to angle point region, wherein, the described judgment criterion for angle point region is: with described pixel
For the criterion that the classification factor in angle point region is relevant in some classification factor;
3rd pixel decision sub-module, for the judged result judging submodule at described 3rd undulating value be
In the case of being, it is determined that described target pixel points is to belong to the pixel in angle point region.
17. devices according to claim 16, it is characterised in that described device also includes:
3rd subclassification determines submodule, for judging that described target pixel points is as the picture belonging to angle point region
After vegetarian refreshments, determine that the subclassification of described target pixel points is direction corresponding to described second minimal ripple value.
18. devices according to claim 16, it is characterised in that described device also includes:
4th pixel decision sub-module, for the judged result judging submodule at described 3rd undulating value be
In the case of no, it is determined that described target pixel points is pole type pixel.
19. devices according to claim 18, it is characterised in that described device also includes:
Mean value computation submodule, for after judging that described target pixel points is as pole type pixel, calculates
The pixel value average of pixel in the default neighborhood region of described target pixel points;
4th subclassification determines submodule, for according to described pixel value average and the picture of described target pixel points
Element value, determines the subclassification of described target pixel points.
20. according to the device according to any one of claim 11-19, it is characterised in that described sensitive parameter
Obtain module,
Specifically for according to the parameter value preset, it is thus achieved that the strength sensitive parameter of target image;Or
Specifically for the compression ratio according to target image, it is thus achieved that the strength sensitive parameter of described target image;Or
Specifically in the case of many wheel codings, tie according to the coding completing to take turns coding of target image more
Really, it is thus achieved that the strength sensitive parameter of described target image;Or
Specifically for the coding result according to coding moment with target image immediate predetermined number image,
Obtain the strength sensitive parameter of described target image.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610196180.4A CN105893999B (en) | 2016-03-31 | 2016-03-31 | region-of-interest extraction method and device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610196180.4A CN105893999B (en) | 2016-03-31 | 2016-03-31 | region-of-interest extraction method and device |
Publications (2)
Publication Number | Publication Date |
---|---|
CN105893999A true CN105893999A (en) | 2016-08-24 |
CN105893999B CN105893999B (en) | 2019-12-13 |
Family
ID=57011713
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610196180.4A Active CN105893999B (en) | 2016-03-31 | 2016-03-31 | region-of-interest extraction method and device |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN105893999B (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112561830A (en) * | 2020-12-23 | 2021-03-26 | 安徽大学 | Endoscope image highlight repair method and device |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101383008A (en) * | 2008-10-23 | 2009-03-11 | 上海交通大学 | Image classification method based on visual attention model |
CN101651772A (en) * | 2009-09-11 | 2010-02-17 | 宁波大学 | Method for extracting video interested region based on visual attention |
CN102332161A (en) * | 2011-09-13 | 2012-01-25 | 中国科学院深圳先进技术研究院 | Image-based intima-media thickness automatic extraction method and system |
CN102567731A (en) * | 2011-12-06 | 2012-07-11 | 北京航空航天大学 | Extraction method for region of interest |
JP2013243486A (en) * | 2012-05-18 | 2013-12-05 | Sharp Corp | Image data inspection device and image data inspection method |
CN103971343A (en) * | 2014-05-21 | 2014-08-06 | 浙江宇视科技有限公司 | Image denoising method based on similar pixel detection |
US20140348386A1 (en) * | 2013-05-22 | 2014-11-27 | Osram Gmbh | Method and a system for occupancy location |
CN104182756A (en) * | 2014-09-05 | 2014-12-03 | 大连理工大学 | Method for detecting barriers in front of vehicles on basis of monocular vision |
-
2016
- 2016-03-31 CN CN201610196180.4A patent/CN105893999B/en active Active
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101383008A (en) * | 2008-10-23 | 2009-03-11 | 上海交通大学 | Image classification method based on visual attention model |
CN101651772A (en) * | 2009-09-11 | 2010-02-17 | 宁波大学 | Method for extracting video interested region based on visual attention |
CN102332161A (en) * | 2011-09-13 | 2012-01-25 | 中国科学院深圳先进技术研究院 | Image-based intima-media thickness automatic extraction method and system |
CN102567731A (en) * | 2011-12-06 | 2012-07-11 | 北京航空航天大学 | Extraction method for region of interest |
JP2013243486A (en) * | 2012-05-18 | 2013-12-05 | Sharp Corp | Image data inspection device and image data inspection method |
US20140348386A1 (en) * | 2013-05-22 | 2014-11-27 | Osram Gmbh | Method and a system for occupancy location |
CN103971343A (en) * | 2014-05-21 | 2014-08-06 | 浙江宇视科技有限公司 | Image denoising method based on similar pixel detection |
CN104182756A (en) * | 2014-09-05 | 2014-12-03 | 大连理工大学 | Method for detecting barriers in front of vehicles on basis of monocular vision |
Non-Patent Citations (1)
Title |
---|
刘紫燕 等: "层次聚类算法的实时图像边缘检测及FPGA实现", 《红外技术》 * |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112561830A (en) * | 2020-12-23 | 2021-03-26 | 安徽大学 | Endoscope image highlight repair method and device |
CN112561830B (en) * | 2020-12-23 | 2022-11-18 | 安徽大学 | Endoscope image highlight repair method and device |
Also Published As
Publication number | Publication date |
---|---|
CN105893999B (en) | 2019-12-13 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US20170365046A1 (en) | Algorithm and device for image processing | |
JP5045817B2 (en) | Image identifier extraction device | |
Saha et al. | Mutual spectral residual approach for multifocus image fusion | |
Ye et al. | Co-saliency detection via co-salient object discovery and recovery | |
CN103020992B (en) | A kind of video image conspicuousness detection method based on motion color-associations | |
CN110706235B (en) | Far infrared pedestrian detection method based on two-stage cascade segmentation | |
CN110929593A (en) | Real-time significance pedestrian detection method based on detail distinguishing and distinguishing | |
US20220239908A1 (en) | Method for processing images, method for processing videos, computer device and medium | |
CN103714181A (en) | Stratification specific figure search method | |
CN106650615A (en) | Image processing method and terminal | |
CN104966305A (en) | Foreground detection method based on motion vector division | |
CN104933679A (en) | A method for enlarging an image and a system corresponding to the method | |
CN111183630B (en) | Photo processing method and processing device of intelligent terminal | |
US8311269B2 (en) | Blocker image identification apparatus and method | |
Chen et al. | Visual depth guided image rain streaks removal via sparse coding | |
CN112926667B (en) | Method and device for detecting saliency target of depth fusion edge and high-level feature | |
Manoharan et al. | Image processing-based framework for continuous lane recognition in mountainous roads for driver assistance system | |
CN105893999A (en) | Method and device for extracting a region of interest | |
KR20080079443A (en) | Method and apparatus for extracting object from image | |
Chen et al. | Head-shoulder detection using joint HOG features for people counting and video surveillance in library | |
CN116665156A (en) | Multi-scale attention-fused traffic helmet small target detection system and method | |
CN111950411A (en) | Model determination method and related device | |
Li et al. | A saliency model based on wavelet transform and visual attention | |
CN109800758A (en) | A kind of natural scene character detecting method of maximum region detection | |
CN108521575A (en) | The noise-reduction method and device of picture noise |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |