CN106033608B - The objective contour detection method of bionical object smooth pursuit eye movement information processing mechanism - Google Patents
The objective contour detection method of bionical object smooth pursuit eye movement information processing mechanism Download PDFInfo
- Publication number
- CN106033608B CN106033608B CN201510442062.2A CN201510442062A CN106033608B CN 106033608 B CN106033608 B CN 106033608B CN 201510442062 A CN201510442062 A CN 201510442062A CN 106033608 B CN106033608 B CN 106033608B
- Authority
- CN
- China
- Prior art keywords
- pixel
- directions
- gabor
- maximum value
- 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.)
- Active
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 20
- 230000004424 eye movement Effects 0.000 title claims abstract description 18
- 230000010365 information processing Effects 0.000 title claims abstract description 12
- 238000010586 diagram Methods 0.000 claims abstract description 42
- 230000009466 transformation Effects 0.000 claims abstract description 37
- 230000005764 inhibitory process Effects 0.000 claims abstract description 28
- 230000002401 inhibitory effect Effects 0.000 claims abstract description 7
- 238000001914 filtration Methods 0.000 claims abstract description 6
- 238000012360 testing method Methods 0.000 claims description 10
- 230000001629 suppression Effects 0.000 claims description 6
- 230000009977 dual effect Effects 0.000 claims description 5
- 238000013178 mathematical model Methods 0.000 claims description 5
- 235000013399 edible fruits Nutrition 0.000 claims 1
- 238000012545 processing Methods 0.000 abstract description 4
- 230000007547 defect Effects 0.000 abstract description 3
- 238000000034 method Methods 0.000 description 7
- 230000000694 effects Effects 0.000 description 3
- 241000208340 Araliaceae Species 0.000 description 2
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 2
- 235000003140 Panax quinquefolius Nutrition 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 235000008434 ginseng Nutrition 0.000 description 2
- 208000003443 Unconsciousness Diseases 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 244000309464 bull Species 0.000 description 1
- 230000000052 comparative effect Effects 0.000 description 1
- 238000003708 edge detection Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000012913 prioritisation Methods 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 230000002207 retinal effect Effects 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
Abstract
The present invention is intended to provide a kind of objective contour detection method of bionical object smooth pursuit eye movement information processing mechanism, includes the following steps：Input image to be detected through gray proces；The Gabor filtering for carrying out all directions, obtains Gabor energy diagrams；Initial DoG templates are established, initial DoG templates are converted, each transformation DoG templates is used in combination to be filtered respectively to Gabor energy diagrams；It is generated according to above-mentioned filter result figure and inhibits maximum value figure；To inhibiting maximum value figure to be modified, obtain correcting inhibition maximum value figure；The Gabor energy values of each pixel in all directions Gabor energy diagrams are subtracted into the filter result value that the same amendment in direction inhibits corresponding pixel in maximum value figure, as the initial profile of the pixel, and makees subsequent processing, obtains final profile figure.Detection method overcomes the defect that prior art fidelity is low, outline identification rate is low, has the characteristics that fidelity is high, outline identification rate is high.
Description
Technical field
The present invention relates to image processing fields, and in particular to a kind of mesh of bionical object smooth pursuit eye movement information processing mechanism
Mark profile testing method.
Background technology
Contour detecting is a critically important component part in image procossing and computer vision.It is correct from complicated background
Ground detection object profile is an extremely important and difficult job.In numerous traditional image processing methods, it is applied to wheel
Exterior feature detection more successfully has Canny operators, movable contour model etc..The letter of the difference in brightness in image is mainly utilized in these methods
Breath is detected, and objective contour and other mixed and disorderly boundaries cannot be distinguished.So in image contrast variation it is bigger,
When the more situation of background interference, these methods hardly result in satisfied result；
The further prioritization scheme of the prior art is：For input picture, existed using the Gabor filter group of multiple directions
Image is filtered successively respectively under two different scale (frequency) parameters；Obtain each pixel maximum gradation value and
Optimal direction is led to filter result obtained by low-frequency parameter as amount of suppression based on filter result obtained by high-frequency parameter respectively
It crosses after dimensional Gaussian difference function (DoG) template is filtered and subtracts each other, obtain final filter profile；The eye movement of human eye refers to human eye
The unconscious small movements under fixation status, refer mainly to flash, when people watches a special object or background attentively, eye
It is dynamic also to have corresponding meaning and effect for the sensitivity and precision of vision；And smooth pursuit eye movement refers to eyes follows regard at leisure
Target movement is felt, to avoid the defocus of target is caused；And the above method uses preset fixed DoG templates, it is assumed that
Human eye is to maintain fixed, ignores the smooth pursuit eye movement effect of human eye, has that fidelity is low, outline identification rate is low
Defect.
Invention content
The present invention is intended to provide a kind of objective contour detection method of bionical object smooth pursuit eye movement information processing mechanism, it should
Detection method overcomes the defect that prior art fidelity is low, outline identification rate is low, the spy high with fidelity, outline identification rate is high
Point.
Technical scheme is as follows, the objective contour detection side of bionical object smooth pursuit eye movement information processing mechanism
Method includes the following steps：
A, image to be detected through gray proces is inputted；
B, the Gabor filter group for presetting multiple directions parameter, to each pixel in image to be detected respectively according to each
A directioin parameter carries out Gabor energy balanes, obtains the Gabor energy values of all directions of each pixel, obtains all directions
Gabor energy diagrams, the direction of the corresponding Gabor filter in direction of the Gabor energy diagrams is consistent；
C, initial DoG templates are established using dimensional Gaussian difference function, the initial DoG templates are circle, and include circle
The center of shape, each pixel point value is zero in the center；
D, it is radiated around using center as basic point in initial DoG templates, presets and filter directioin parameter phase with each Gabor
Corresponding direction divides, and default one of each direction is circular to wait for that zeroed extents and the center of circle are located at the direction and divide, and respectively waits setting
Null range respectively waits for distance phase of the zeroed extents apart from the center of initial DoG templates other than the center of initial DoG templates
Together, successively to respectively waiting for that whole pixel point values in zeroed extents carry out zero setting, each zero setting generates a transformation DoG template, and
Each zero setting is based on initial DoG templates；
E, each Gabor energy diagrams are filtered with each transformation DoG templates, obtain the transformation filter result of all directions
Figure, the direction of the corresponding Gabor energy diagrams in direction are identical；
F, for unidirectional transformation filter result value figure, the maximum of the filter result value of wherein each pixel is chosen
Value, as the pixel in the inhibition maximum value of the direction, generates the inhibition maximum value figure of the direction, repetition of changing direction is above-mentioned
Operation, obtains the inhibition maximum value figure of all directions, is modified to the inhibition maximum value figure of all directions, obtains the amendment of all directions
Inhibit maximum value figure；
G, the Gabor energy values of each pixel in all directions Gabor energy diagrams the same amendment in direction is subtracted to press down
The filter result value of corresponding pixel, as the initial profile of the pixel direction, and then obtains in maximum value figure processed
The initial profile of each pixel all directions；
H, the maximum value in the initial profile of each pixel all directions is chosen, as the largest contours value of the pixel,
Largest contours value is handled using non-maxima suppression and dual threshold, obtains the final profile value of each pixel, and then obtain most
Whole profile diagram.
Preferably, in the step E, further include with converting DoG templates and being filtered to each Gabor energy diagrams：It uses
Each transformation DoG templates are filtered each Gabor energy diagrams successively, obtain the transformation filter result figure of all directions.
Preferably, the two-dimensional Gabor function expression of Gabor filter group is as follows in the step B：
Whereinγ is one and indicates oval receptive field axial ratio
The constant of example, parameter lambda are wavelength, and σ is the bandwidth in area of DoG template center, and 1/ λ is the spatial frequency of cosine function,It is phase angle
Parameter, θ are the directioin parameter of Gabor filtering；
Gabor energy diagram computation models are as follows：
Wherein
I is image to be detected, and * is convolution operator.
Preferably, the corresponding expression formula of initial DoG templates is as follows in the step C：
The corresponding distance weighting template expression formula of DoG templates is as follows：
Wherein||·||1For single order (L1)
Norm.
Preferably, wait for that the corresponding mathematical model of zeroed extents is as follows in the step D：
Wait for zeroed extents set expression be x, y | x2+(y-d)2≤R2} (6)；
X=d*cos α, y=d*sin α (12)；
Wherein d is with a distance from center, and R is the radius for waiting for zeroed extents, and α is deviation angle.
Preferably, in the step E, further include with converting DoG templates and being filtered to each Gabor energy diagrams：It uses
Each transformation DoG templates are filtered each Gabor energy diagrams successively, obtain the transformation filter result figure of all directions；
Corresponding formula is as follows：
Ii, θ (x, y)=Eλ,σ,θ(x,y)*ωd,i(x,y,σ) (7)；
Wherein Ii,θ(x, y) is i-th of filter result value for converting pixel (x, y) in filter result figure in the directions θ,
Eλ,σ,θ(x, y) is the Gabor energy values of pixel (x, y) in the Gabor energy diagrams in the directions θ, and * is convolution algorithm, ωd,i(x,y,
σ) refer to i-th of transformation DoG template.
Preferably, maximum value is inhibited to be obtained by the following formula in the step F：
WhereinFor the inhibition maximum value in the directions pixel (x, y) θ, Ii,θ(x, y) is i-th of transformation in the directions θ
The filter result value of pixel (x, y), M in filter result figureiIndicate the number of transformation DoG templates；
Amendment in the step F refers to inhibiting the value of each pixel in maximum value figure to be multiplied by preset parameter.
Preferably, the corresponding mathematical models of the step G are as follows：
Wherein bθ(x, y) is the initial profile of pixel (x, y), Eλ,σ,θ(x, y) is in the Gabor energy diagrams in the directions θ
The Gabor energy values of pixel (x, y), a are corrected parameter,For the inhibition maximum value in the directions pixel (x, y) θ；
Wherein H function is：
Preferably, the Gabor filter group of the multiple directions parameter in the step B, the filter of different directions
Number is 8-12, and radians distribution is waited in 360 degree.
Preferably, the amendment in the step F refers to inhibiting the value of each pixel in maximum value figure to be multiplied by preset ginseng
Number.
Preferably, wait for that the radius of zeroed extents is 4-6 in the step D.
Technical solution of the present invention is innovatively by the smooth pursuit eye movement phase of the contour extraction method of classical receptive field and human eye
In conjunction with carrying out analogue simulation to the smooth pursuit eye movement of human eye using the DoG templates of transformation, improve classical receptive field model
Fidelity；Smooth pursuit eye movement can accomplish quickly and accurately edge extracting to the small and weak information of concern, also can be to retinal periphery
Area information generates appropriate Alernting effect, can further increase the discrimination of the contour extraction method of classical receptive field.
Description of the drawings
Fig. 1 is the flow chart of the objective contour detection method of the bionical object smooth pursuit eye movement information processing mechanism of the present invention
Fig. 2 is 1DoG template transformation area schematics of the embodiment of the present invention
Fig. 3 is the image 1 that the embodiment of the present invention 1 carries out contour detecting
Fig. 4 is the nominal contour figure of image 1
Fig. 5 is the profile diagram that image 1 is obtained through 1 detection method of document
Fig. 6 is the profile diagram that image 1 is obtained through 1 detection method of embodiment
Fig. 7 is the image 2 that the embodiment of the present invention 1 carries out contour detecting
Fig. 8 is the nominal contour figure of image 2
Fig. 9 is the profile diagram that image 2 is obtained through 1 detection method of document
Figure 10 is the profile diagram that image 2 is obtained through 1 detection method of embodiment
Each section title and serial number are as follows in Fig. 2：1 is DoG templates, and 2 be area of DoG template center, and 3 be to wait for zeroed extents.
Specific implementation mode
The present invention is illustrated with reference to the accompanying drawings and examples.
Embodiment 1
As shown in Figure 1, the objective contour detection method packet of the bionical object smooth pursuit eye movement information processing mechanism of the present embodiment
Include following steps：
A, image to be detected through gray proces is inputted；
B, the Gabor filter group for presetting multiple directions parameter, to each pixel in image to be detected respectively according to each
A directioin parameter carries out Gabor energy balanes, obtains the Gabor energy values of all directions of each pixel, obtains all directions
Gabor energy diagrams, the direction of the corresponding Gabor filter in direction of the Gabor energy diagrams is consistent；
The two-dimensional Gabor function expression of Gabor filter group is as follows in the step B：
Whereinγ is one and indicates oval receptive field major and minor axis ratio
Constant, parameter lambda is wavelength, and σ is the bandwidth in area of DoG template center, and 1/ λ is the spatial frequency of cosine function,It is phase angle ginseng
Number, θ are the directioin parameter of Gabor filtering；
Gabor energy diagram computation models are as follows：
Wherein
I is image to be detected, and * is convolution operator；
C, initial DoG templates are established using dimensional Gaussian difference function, the initial DoG templates are circle, and include circle
The center of shape, each pixel point value is zero in the center；
The corresponding expression formula of initial DoG templates is as follows in the step C：
The corresponding distance weighting template expression formula of DoG templates is as follows：
Wherein||·||1For single order (L1)
Norm；
D, it is radiated around using center as basic point in initial DoG templates, presets and filter directioin parameter phase with each Gabor
Corresponding direction divides, and default one of each direction is circular to wait for that zeroed extents and the center of circle are located at the direction and divide, and respectively waits setting
Null range respectively waits for distance phase of the zeroed extents apart from the center of initial DoG templates other than the center of initial DoG templates
Together, successively to respectively waiting for that whole pixel point values in zeroed extents carry out zero setting, each zero setting generates a transformation DoG template, and
Each zero setting is based on initial DoG templates；
Wait for that the corresponding mathematical model of zeroed extents is as follows in the step D：
Wait for zeroed extents set expression be x, y | x2+(y-d)2≤R2} (7)；
X=d*cos α, y=d*sin α (8)；
Wherein d is with a distance from center, and R is the radius for waiting for zeroed extents, and α is deviation angle；
E, each Gabor energy diagrams are filtered with each transformation DoG templates, obtain the transformation filter result of all directions
Figure, the direction of the corresponding Gabor energy diagrams in direction are identical；
Corresponding formula is as follows：
Ii,θ(x, y)=Eλ,σ,θ(x,y)*ωd,i(x,y,σ) (9)；
Wherein Ii,θ(x, y) is i-th of filter result value for converting pixel (x, y) in filter result figure in the directions θ,
Eλ,σ,θ(x, y) is the Gabor energy values of pixel (x, y) in the Gabor energy diagrams in the directions θ, and * is convolution algorithm, ωd,i(x,y,
σ) refer to i-th of transformation DoG template；
F, for unidirectional transformation filter result value figure, the maximum of the filter result value of wherein each pixel is chosen
Value, as the pixel in the inhibition maximum value of the direction, generates the inhibition maximum value figure of the direction, repetition of changing direction is above-mentioned
Operation, obtains the inhibition maximum value figure of all directions, is modified to the inhibition maximum value figure of all directions, obtains the amendment of all directions
Inhibit maximum value figure；
Maximum value is inhibited to be obtained by the following formula in the step F：
WhereinFor the inhibition maximum value in the directions pixel (x, y) θ, Ii,θ(x, y) is i-th of transformation filter in the directions θ
The filter result value of pixel (x, y), M in wave result figureiIndicate the number of transformation DoG templates；
Amendment in the step F refers to inhibiting the value of each pixel in maximum value figure to be multiplied by preset parameter；
G, the Gabor energy values of each pixel in all directions Gabor energy diagrams the same amendment in direction is subtracted to press down
The filter result value of corresponding pixel, as the initial profile of the pixel direction, and then obtains in maximum value figure processed
The initial profile of each pixel all directions；
The corresponding mathematical model of the step G is as follows：
Wherein bθ(x, y) is the initial profile of pixel (x, y), Eλ,σ,θ(x, y) is in the Gabor energy diagrams in the directions θ
The Gabor energy values of pixel (x, y), a are corrected parameter,For the inhibition maximum value in the directions pixel (x, y) θ；
Wherein H function is：
H, the maximum value in the initial profile of each pixel all directions is chosen, as the largest contours value of the pixel,
Largest contours value is handled using non-maxima suppression and dual threshold, obtains the final profile value of each pixel, and then obtain most
Whole profile diagram.
Above-mentioned non-maxima suppression and dual threshold processing are using the method provided in following documents：
Canny,J.,A Computational Approach To Edge Detection,IEEE
Trans.Pattern Analysis and M achine Intelligence,8(6):679–698,1986.。
The Gabor filter group of multiple directions parameter in step B described in the present embodiment, the filter of different directions
Number is 8, and radians distribution is waited in 360 degree；Amendment in the step H refers to inhibiting each pixel in maximum value figure
Value is multiplied by preset parameter, and the parameter is 1.2；The radius for waiting for zeroed extents is 4, and excentric distance is 20.
I.e.：
Image to be detected through gray proces is inputted in step A；
The Gabor filter group that 8 directions are preset in step B, obtains the Gabor energy values in 8 directions of each pixel,
And then obtain the Gabor energy diagrams in 8 directions；
Initial DoG templates are established using dimensional Gaussian difference function in step C, the initial DoG templates are circle, and are wrapped
Containing circular center, each pixel point value is zero in the center；
It presets 8 directions in step D to divide, i.e. 8 directions wait for that zeroed extents, each direction wait for zeroed extents from
The distance of the heart is 20, radius 4, to obtain 8 transformation DoG templates；
The Gabor energy diagrams in 8 directions are filtered respectively with 8 transformation DoG templates in step E, obtain 8 directions
Transformation filter result figure, the transformation filter result figure in each direction has 8；
F, for unidirectional transformation filter result value figure, the maximum of the filter result value of wherein each pixel is chosen
Value, as the pixel in the inhibition maximum value of the direction, generates the inhibition maximum value figure of the direction, repetition of changing direction is above-mentioned
Operation, obtains the inhibition maximum value figure of all directions, is modified to the inhibition maximum value figure of all directions, obtains the amendment of all directions
Inhibit maximum value figure；
For unidirectional 8 transformation filter result value figures in step F, the filter result value of wherein each pixel is chosen
Maximum value generate the inhibition maximum value figure of the direction as the pixel in the inhibition maximum value of the direction, change direction weight
Multiple aforesaid operations, obtain the inhibition maximum value figure in 8 directions, are modified to the inhibition maximum value figure in 8 directions, obtain 8
The amendment in direction inhibits maximum value figure；
It is same that the Gabor energy values of each pixel in 8 direction Gabor energy diagrams are subtracted into direction in step G
The filter result value for inhibiting corresponding pixel in maximum value figure is corrected, as the profile value of the pixel direction, and then
To the initial profile in 8 directions of each pixel；
The maximum value in the initial profile in 8 directions of each pixel is chosen in step H, the most bull wheel as the pixel
Exterior feature value is handled largest contours value using non-maxima suppression and dual threshold, obtains the final profile value of each pixel, and then obtain
To final profile figure.
As shown in Fig. 2, the present embodiment DoG template transformation area schematics, wherein 1 is DoG templates, 2 be DoG template center
Area, 3 be to wait for zeroed extents；
As shown in figs. 3-10, the present embodiment image more classical to two images process field carries out contour detecting, and
Comparative result is carried out with contour detecting field classic algorithm document 1, document 1 is " Cosmin Grigorescu, Nicolai
Petkov,and Michel A.Westenberg.Contour Detection Based on Nonclassical
Receptive Field Inhibition[J].IEEE Transactions on image processing,vol.12,
2003 729-739 of no.7, july ", comparing result is referring to table 1；
1 document of table, 1 profile testing method and the testing result P of 1 method of embodiment are compared：
Above-mentioned testing result P uses following evaluation and test formula：
Evaluating standard P is between [0,1].Card (X) indicates the number of member in set X in formula；C, CFPAnd CFNTable respectively
Show the profile correctly detected, the profile of false profile and omission.If all true profiles have all correctly detected out,
And it is contour pixel, then P=1 by false retrieval without background edge；When false retrieval (missing inspection) is more, P is closer to 0.
Claims (9)
1. the objective contour detection method of bionical object smooth pursuit eye movement information processing mechanism, it is characterised in that including following step
Suddenly：
A, image to be detected through gray proces is inputted；
B, the Gabor filter group for presetting multiple directions parameter, to each pixel in image to be detected respectively according to each side
Gabor energy balanes are carried out to parameter, the Gabor energy values of all directions of each pixel is obtained, obtains all directions
The direction of Gabor energy diagrams, the corresponding Gabor filter in direction of the Gabor energy diagrams is consistent；
C, initial DoG templates are established using dimensional Gaussian difference function, the initial DoG templates are circle, and include circular
Center, each pixel point value is zero in the center；
D, it radiates, presets corresponding with each Gabor filtering directioin parameters around using center as basic point in initial DoG templates
Direction divide, default one of each direction is circular to wait for that zeroed extents and the center of circle are located in direction division, respectively waits for zero setting area
Domain respectively waits for that zeroed extents are identical apart from the distance at the center of initial DoG templates other than the center of initial DoG templates, according to
It is secondary to respectively waiting for that whole pixel point values in zeroed extents carry out zero setting, each zero setting generates a transformation DoG template, and sets every time
Zero based on initial DoG templates；
E, each Gabor energy diagrams are filtered with each transformation DoG templates, obtain the transformation filter result figure of all directions,
The direction of the corresponding Gabor energy diagrams in direction is identical；
F, for unidirectional transformation filter result value figure, the maximum value of the filter result value of wherein each pixel is chosen, is made
It is the pixel in the inhibition maximum value of the direction, generates the inhibition maximum value figure of the direction, repetition aforesaid operations of changing direction,
The inhibition maximum value figure of all directions is obtained, the inhibition maximum value figure of all directions is modified, the amendment for obtaining all directions inhibits
Maximum value figure；
G, the Gabor energy values of each pixel in all directions Gabor energy diagrams the same amendment in direction is subtracted to inhibit most
The filter result value of corresponding pixel, as the initial profile of the pixel direction, and then obtains each picture in big value figure
The initial profile of vegetarian refreshments all directions；
H, the maximum value in the initial profile of each pixel all directions is chosen, as the largest contours value of the pixel, to most
Big profile value is handled using non-maxima suppression and dual threshold, obtains the final profile value of each pixel, and then finally taken turns
Exterior feature figure.
2. the objective contour detection method of bionical object smooth pursuit eye movement information processing mechanism as described in claim 1, special
Sign is：The two-dimensional Gabor function expression of Gabor filter group is as follows in the step B：
Whereinγ is one and indicates the normal of oval receptive field major and minor axis ratio
Number, parameter lambda are wavelength, and σ is the bandwidth in area of DoG template center, and 1/ λ is the spatial frequency of cosine function,It is phase angular dimensions, θ
For the directioin parameter of Gabor filtering；
Gabor energy diagram computation models are as follows：
Wherein
I is image to be detected, and * is convolution operator.
3. the objective contour detection method of bionical object smooth pursuit eye movement information processing mechanism as described in claim 1, special
Sign is：The corresponding expression formula of initial DoG templates is as follows in the step C：
The corresponding distance weighting template expression formula of DoG templates is as follows：
Wherein||·||1For single order (L1) norm.
4. profile testing method as described in claim 1, it is characterised in that：The corresponding number of zeroed extents is waited in the step D
It is as follows to learn model：
Wait for zeroed extents set expression be x, y | x2+(y-d)2≤R2} (7)；
X=d*cos α, y=d*sin α (8)；
Wherein d is with a distance from center, and R is the radius for waiting for zeroed extents, and α is deviation angle.
5. the objective contour detection method of bionical object smooth pursuit eye movement information processing mechanism as described in claim 1, special
Sign is：
In the step E, further include with converting DoG templates and being filtered to each Gabor energy diagrams：Use each transformation DoG
Template is filtered each Gabor energy diagrams successively, obtains the transformation filter result figure of all directions；
Corresponding formula is as follows：
Ii,θ(x, y)=Eλ,σ,θ(x,y)*ωd,i(x,y,σ) (9)；
Wherein Ii,θ(x, y) is i-th of filter result value for converting pixel (x, y) in filter result figure in the directions θ, Eλ,σ,θ(x,
Y) it is the Gabor energy values of pixel (x, y) in the Gabor energy diagrams in the directions θ, * is convolution algorithm, ωd,i(x, y, σ) refers to
I-th of transformation DoG template.
6. profile testing method as described in claim 1, it is characterised in that：
Maximum value is inhibited to be obtained by the following formula in the step F：
WhereinFor the inhibition maximum value in the directions pixel (x, y) θ, Ii,θ(x, y) is i-th of transformation filtering knot in the directions θ
The filter result value of pixel (x, y), M in fruit figureiIndicate the number of transformation DoG templates；
Amendment in the step F refers to inhibiting the value of each pixel in maximum value figure to be multiplied by preset parameter.
7. profile testing method as described in claim 1, it is characterised in that：
The corresponding mathematical model of the step G is as follows：
Wherein bθ(x, y) is the initial profile of pixel (x, y), Eλ,σ,θ(x, y) is pixel in the Gabor energy diagrams in the directions θ
The Gabor energy values of point (x, y), a are corrected parameter,For the inhibition maximum value in the directions pixel (x, y) θ；
Wherein H function is：
8. profile testing method as described in claim 1, it is characterised in that：Multiple directions parameter in the step B
The number of filter of Gabor filter group, different directions is 8-12, and radians distribution is waited in 360 degree.
9. profile testing method as described in claim 1, it is characterised in that：
Wait for that the radius of zeroed extents is 4-6 in the step D.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510442062.2A CN106033608B (en) | 2015-07-24 | 2015-07-24 | The objective contour detection method of bionical object smooth pursuit eye movement information processing mechanism |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510442062.2A CN106033608B (en) | 2015-07-24 | 2015-07-24 | The objective contour detection method of bionical object smooth pursuit eye movement information processing mechanism |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106033608A CN106033608A (en) | 2016-10-19 |
| CN106033608B true CN106033608B (en) | 2018-10-19 |
Family
ID=57149422
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510442062.2A Active CN106033608B (en) | 2015-07-24 | 2015-07-24 | The objective contour detection method of bionical object smooth pursuit eye movement information processing mechanism |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106033608B (en) |
Families Citing this family (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107067408B (en) * | 2017-04-11 | 2020-01-31 | 广西科技大学 | Image contour detection method for simulating human eye micromotion |
| CN107909593B (en) * | 2017-11-07 | 2020-05-05 | 广西科技大学 | Non-directional selective contour detection method based on receptive field region |
| CN107766866B (en) * | 2017-11-07 | 2020-05-05 | 广西科技大学 | Direction selective contour detection method based on receptive field region |
| CN108053415B (en) * | 2017-12-14 | 2020-05-22 | 广西科技大学 | Bionic contour detection method based on improved non-classical receptive field |
| CN109087324B (en) * | 2018-08-03 | 2021-08-31 | 广西科技大学 | Contour detection method based on color antagonistic receptive field and black and white channel |
| CN109146902B (en) * | 2018-08-03 | 2021-09-21 | 广西科技大学 | Contour detection method based on color antagonistic receptive field and black and white channel |
| CN109146901B (en) * | 2018-08-03 | 2021-09-17 | 广西科技大学 | Contour detection method based on color antagonistic receptive field |
| CN115984952B (en) * | 2023-03-20 | 2023-11-24 | 杭州叶蓁科技有限公司 | Eye movement tracking system and method based on bulbar conjunctiva blood vessel image recognition |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2009098743A1 (en) * | 2008-02-04 | 2009-08-13 | Honda Motor Co., Ltd. | Vehicle periphery monitoring device, vehicle, vehicle periphery monitoring program, and vehicle periphery monitoring method |
| CN101763641B (en) * | 2009-12-29 | 2011-09-14 | 电子科技大学 | Method for detecting contour of image target object by simulated vision mechanism |
| CN102201120B (en) * | 2011-06-17 | 2012-08-29 | 电子科技大学 | Multifeature-based target object contour detection method |
| CN103903251A (en) * | 2012-12-30 | 2014-07-02 | 南京理工大学 | Night vision image salient contour extracting method based on non-classical receptive field composite modulation |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070230754A1 (en) * | 2006-03-30 | 2007-10-04 | Jain Anil K | Level 3 features for fingerprint matching |
-
2015
- 2015-07-24 CN CN201510442062.2A patent/CN106033608B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2009098743A1 (en) * | 2008-02-04 | 2009-08-13 | Honda Motor Co., Ltd. | Vehicle periphery monitoring device, vehicle, vehicle periphery monitoring program, and vehicle periphery monitoring method |
| CN101763641B (en) * | 2009-12-29 | 2011-09-14 | 电子科技大学 | Method for detecting contour of image target object by simulated vision mechanism |
| CN102201120B (en) * | 2011-06-17 | 2012-08-29 | 电子科技大学 | Multifeature-based target object contour detection method |
| CN103903251A (en) * | 2012-12-30 | 2014-07-02 | 南京理工大学 | Night vision image salient contour extracting method based on non-classical receptive field composite modulation |
Non-Patent Citations (3)
| Title |
|---|
| Contour detection based on a non-classical receptive field model with butterfly-shaped inhibition subregions;Zeng Chi 等;《Neurocomputing》;20110531;第74卷(第10期);第1527-1534页 * |
| 引入低对比度环境下视觉感知机制的轮廓检测模型;陈建军 等;《模式识别与人工智能》;20121015;第25卷(第5期);第845-850页 * |
| 视觉感知结合学习的自然图像轮廓检测测模型;唐奇伶 等;《中国科学:信息科学》;20130920;第43卷(第9期);第1124-1135页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106033608A (en) | 2016-10-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106033608B (en) | The objective contour detection method of bionical object smooth pursuit eye movement information processing mechanism | |
| CN107767387B (en) | Contour detection method based on variable receptive field scale global modulation | |
| CN106033609B (en) | The objective contour detection method of bionical object jump eye movement information processing mechanism | |
| Zilly et al. | Boosting convolutional filters with entropy sampling for optic cup and disc image segmentation from fundus images | |
| CN110232389B (en) | Stereoscopic vision navigation method based on invariance of green crop feature extraction | |
| CN106033610B (en) | Profile testing method based on the modulation of non-classical receptive field spatial summation | |
| EP1229489A3 (en) | Image processing method and apparatus | |
| CN109919960B (en) | Image continuous edge detection method based on multi-scale Gabor filter | |
| CN110232390B (en) | Method for extracting image features under changed illumination | |
| CN102096824A (en) | Multi-spectral image ship detection method based on selective visual attention mechanism | |
| US20150186753A1 (en) | Image Processing Apparatus, Image Processing Method, and Image Processing Program | |
| CN103425998A (en) | Method for identifying SAR target under shielding conditions | |
| CN107909593B (en) | Non-directional selective contour detection method based on receptive field region | |
| CN107038432B (en) | Fingerprint image direction field extraction method based on frequency information | |
| CN107886539B (en) | High-precision gear visual detection method in industrial scene | |
| CN106780465A (en) | Retinal images aneurysms automatic detection and recognition methods based on gradient vector analysis | |
| CN105678245A (en) | Target position identification method based on Haar features | |
| CN105893916A (en) | New method for detection of face pretreatment, feature extraction and dimensionality reduction description | |
| CN105809173A (en) | Bionic vision transformation-based image RSTN (rotation, scaling, translation and noise) invariant attributive feature extraction and recognition method | |
| CN111461654A (en) | Face recognition sign-in method and device based on deep learning algorithm | |
| CN107766866B (en) | Direction selective contour detection method based on receptive field region | |
| CN108133211B (en) | Power distribution cabinet detection method based on mobile terminal visual image | |
| CN106033606B (en) | The objective contour detection method of bionical object smooth pursuit eye movement information processing mechanism | |
| CN101533475A (en) | Method for extracting feature of shape-adaptive neighborhood based remote sensing image | |
| CN108010046A (en) | Based on the bionical profile testing method for improving classical receptive field |
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 | ||
| EE01 | Entry into force of recordation of patent licensing contract | ||
| EE01 | Entry into force of recordation of patent licensing contract |
Application publication date: 20161019 Assignee: HUALI FAMILY PRODUCTS CO.,LTD. Assignor: GUANGXI University OF SCIENCE AND TECHNOLOGY Contract record no.: X2023980054119 Denomination of invention: Target contour detection method based on biomimetic smooth tracking eye tracking information processing mechanism Granted publication date: 20181019 License type: Common License Record date: 20231226 |