CN111192280A - Method for detecting optic disc edge based on local feature - Google Patents
Method for detecting optic disc edge based on local feature Download PDFInfo
- Publication number
- CN111192280A CN111192280A CN201911348067.3A CN201911348067A CN111192280A CN 111192280 A CN111192280 A CN 111192280A CN 201911348067 A CN201911348067 A CN 201911348067A CN 111192280 A CN111192280 A CN 111192280A
- Authority
- CN
- China
- Prior art keywords
- point
- value
- boundary
- points
- pixel
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- 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
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30041—Eye; Retina; Ophthalmic
Abstract
The invention belongs to the technical field of computer image processing, and discloses a method for detecting an optic disc edge based on local characteristics, which searches for adjacent boundary points by using the morphological characteristics of an optic disc in a screened local area to be processed, and comprises the following steps: (1) preprocessing and synthesizing the fundus images, and determining the characteristic value of each pixel point; (2) determining horizontal initial value row by average maximum characteristic value of each row, determining vertical initial value by difference derivation to obtain local region to be processedS(ii) a (3) By usingSobelThe operator calculates the gradient of each point in turnG(ii) a (4) Finding a gradient extreme value in the horizontal initial value row, and determining two initial boundary points; (5) starting from two initial boundary points, searching for adjacent boundary points. The invention decomposes the identification problem of the optic disk in the whole eye fundus image into the adjacent boundary point searching problem, thereby effectively reducing the number of the optic disk in the whole eye fundus imageThe algorithm complexity is low, and the identification precision is improved.
Description
Technical Field
The invention relates to a method for detecting the edge of a video disc based on local characteristics, belonging to the technical field of computer image processing.
Background
The optic disc is the most important component in the fundus image, and the accurate extraction of the optic disc is the basis for effectively diagnosing the fundus diseases and extracting the characteristics of the fundus. The optic disc appears as a light red or light yellow area of an approximate circle in the fundus image, but the detection effect of the commonly used edge detection algorithm is not good due to factors such as the shooting environment, personal differences, fundus diseases, and the like. Therefore, it is necessary to propose an appropriate optic disc edge detection method for the particularity of the fundus image.
For the problem, some corresponding solutions are proposed in the current research, but the problems of high image quality requirement and low lesion image accuracy rate generally exist. Most methods perform disc edge detection by converting a color image into a grayscale image or a G-channel image. Existing detection techniques fall into two broad categories: a method for positioning based on the relationship between the optic disc and the retinal vessel and a method for analyzing based on the morphological characteristics of the optic disc. The data dimensionality is reduced in the image transformation process, so that the problem of reduction of the edge detection accuracy is caused, and meanwhile, the problem that algorithm complexity and robustness cannot be considered at the same time exists. In addition, in the prior art, the edge of the optic disc is determined by detecting the whole image, the processing data volume is large, and the process is complex.
Disclosure of Invention
The invention overcomes the defects of the prior art, and solves the technical problems that: a method for detecting the edge of a video disc based on local features is provided to improve the accuracy of extracting the edge of the video disc and reduce the amount of calculation.
In order to solve the technical problems, the invention adopts the technical scheme that: 1. a method for detecting the edge of an optic disc based on local features is characterized by comprising the following steps:
s1, smoothing the color fundus image to be processed by adopting a Gaussian filter, and recording as an image Grap;
s2, performing the following processing on the image gray: carrying out gray level transformation by a weighted average method to obtain the gray level value of each pixel point; obtaining the brightness value of each pixel point through HSV space transformation; taking the mean value of the gray value and the brightness value as a characteristic value e to obtain a new image Grap';
s3, establishing a rectangular coordinate system x-y taking pixels as a unit by taking the upper left corner of the Grap' as an origin;
s4, sequentially calculating the average value avgy of the characteristic values e of each pixel point of each line in the image Grap', and calculating the value y corresponding to the maximum value of avgy0I.e. the horizontal initial value row of the optic disc;
s5, finding y0First derivative, y, of the characteristic value e of each pixel point in a line0The calculation formula of the first derivative dx of the pixel point with the horizontal coordinate i is as follows: dx ═ e (i, y)0)-e(i-1，y0) Wherein e (i, y)0) Denotes y0Characteristic value e (i-1, y) of pixel point with horizontal coordinate i0) Denotes y0Characteristic values of pixel points with horizontal coordinates of i-1 are obtained; finding out the abscissa (xmax and xmin) corresponding to the maximum value and the minimum value of the first derivative, and recording the midpoint coordinate of the two as x0With O1(x0，y0) Determining a local area S to be processed as a center;
s6, removing blood vessels in the region S by using an image restoration method, and sequentially calculating the gradient G of each point in the region S to be processed by using a Sobel operator; at y0Finding out two peak points with the maximum gradient value change, and recording the peak points as peak and peak; at the same time, pixel point M (peak, y)0) And N (peak, y)0) As disc image in y0Two boundary points of a line, wherein peak<peakr；
S7, using M point as starting point, according to clockwise or anticlockwise rotation direction, searching boundary points M 'adjacent to M one by one, after finding out, setting M as M', repeating searching until coinciding with N point, connecting all boundary points to be located at y0An upper or lower optic disc edge; similarly, with the point N as the starting point, the boundary points N ' adjacent to the point N are searched one by one according to the same rotating direction, after the boundary points N ' are found, the boundary points N ' are set to be N ', the search is repeated until the boundary points N ' coincide with the point M, and all the boundary points are connected to be positioned at the point y0A lower or upper optic disc edge;
the specific method for searching the adjacent boundary points comprises the following steps: setting a gradient threshold value T and a characteristic value threshold value, classifying pixel points with characteristic values larger than the characteristic value threshold value as the interior of a video disc, classifying pixel points with characteristic values smaller than the characteristic value threshold value as the exterior of the video disc, and searching whether adjacent points thereof meet the following conditions from a starting point one by one: and (3) determining 8 neighborhood pixel points of the boundary point, wherein the gradient value of a certain point is greater than the gradient threshold value, at least one of the neighborhood pixel points of the point is positioned in the video disc, and if the gradient value of the certain point is met, the certain point is used as the next boundary point.
In step S7, if the boundary point found finally does not coincide with the point N when the point M is used as the starting point, the adjacent boundary points are searched in the opposite direction using the point N as the starting point until the boundary point is found to coincide with the boundary point found from the point M or y is reached0Taking the central pixel point of the boundary points found twice as the boundary point finally found; similarly, when the point N is used as the starting point, if the boundary point found finally is not coincident with the point M, the adjacent boundary points are searched in the opposite direction by using the point M as the starting point until the found boundary point is coincident with the boundary point found from the point M or reaches y0And taking the central pixel point of the boundary point found twice as the boundary point finally found.
In step S7, when the boundary points M 'adjacent to M are searched one by one in the clockwise direction, the boundary points M' are searched in the order of priority from the top right, the right side, and the bottom right, and when the boundary points N 'adjacent to N are searched one by one in the clockwise direction, the boundary points M' are searched in the order of priority from the bottom left, the left side, and the top left.
In the step S6, the value range of the gradient threshold T is 60-80% of the mean value of the peak value peak and peak of the gradient, and the eigenvalue threshold is the eigenvalue of the last boundary point.
In step S6, the calculation formula for calculating the gradient G of each point in the region S to be processed is:
wherein the content of the first and second substances,
Gx＝(ei-1，j+1+2ei，j+1+ei+1，j+1)-(ei-1，j-1+2ei，j-1+ei+1，j-1)；
Gy＝(ei-1，j-1+2ei-1，j+ei-1，j+1)-(ei+1，j-1+2ei+1，j+ei+1，j+1)；
ei，jand (d) representing the characteristic value of the pixel point with the coordinate (i, j).
In step S5, the height and width of the local region to be processed S are 1/3 of the image height and width.
Compared with the prior art, the invention has the following beneficial effects:
(1) the invention combines the gray value and the brightness value, makes up the defect that the brightness information can not be embodied in the gray image, and the brightness information of the video disc is obvious, so the edge extraction of the next step can be more accurate by adopting the method of superposing the brightness information on the gray image and then taking the median.
(2) The invention simplifies the edge extraction process of the whole fundus image, determines the local to-be-processed range by a primary positioning method and reduces the processing range. In addition, two edge points are determined by finding a gradient extreme value in the maximum average characteristic value row, and then edge extraction is sequentially carried out by using the morphological characteristics of the optic disc through a characteristic value analysis method of 8 neighborhood nodes, so that the computation amount is effectively reduced.
Drawings
Fig. 1 is an original image Grap and a preprocessed image;
fig. 2 is a schematic diagram of a comparison of a grayscale image, a luminance image and a combined image gray';
fig. 3 is a schematic view of the coordinate system x-y and the local region to be processed S of the image Grap';
fig. 4 is a schematic view of determination of a local region to be processed S;
FIG. 5 is a horizontal initial row y0A schematic diagram of the positions of the lines in the image and their characteristic value curves and gradient value curves;
FIG. 6 is a schematic view of a disc;
FIG. 7 is a schematic diagram of the characteristic analysis of 8 neighborhood nodes of a pixel M;
FIG. 8 is a schematic diagram of 10 edge points M' found starting from point M;
fig. 9 is a diagram showing the result of rapid detection of the edge of the optic disc.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below, and it is obvious that the described embodiments are some embodiments of the present invention, but not all embodiments; all other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The embodiment of the invention provides a method for detecting the edge of an optic disc based on local features, which comprises the following 7 steps:
and S1, smoothing by adopting a Gaussian filter to obtain a color fundus image Grap to be processed.
In this embodiment, the size of the color fundus image to be processed gray is 3048 × 2432, and the size of the image is adjusted to 600 × 800 for convenience of processing because the image is too large. And the image is smoothed by adopting a Gaussian filter, so that the noise influence is reduced. The image to be processed Grap and its smoothed image are shown in fig. 1.
S2, performing the following processing on the image gray: carrying out gray level transformation by a weighted average method to obtain the gray level value of each pixel point; obtaining the brightness value of each pixel point through HSV space transformation; and averaging the gray value and the brightness value to obtain a new image Grap'.
In the present embodiment, the pair of the grayscale image, the luminance image, and the combined image gray' is shown in fig. 2. By combining the gray value and the brightness value, the defect that the brightness information cannot be embodied in the gray image is made up, and the brightness information of the video disc is highlighted, so that the edge extraction in the next step can be more accurate by adopting a method of superposing the brightness information on the gray image and then taking the median.
And S3, establishing a rectangular coordinate system x-y by taking the upper left corner of the image Grap' as an origin and taking the pixel as a unit. As shown in fig. 3.
S4, sequentially calculating the average value avgy of the characteristic values e of each pixel point of each line in the image Grap',and find the value y corresponding to the maximum value of avgy0I.e. the horizontal initial value row of the optic disc;
s5, using j-1 as initial value and j-600 as maximum value, finding y0First derivative, y, of the characteristic value e of each pixel point in a line0The calculation formula of the first derivative dx of the pixel point with the horizontal coordinate i is as follows: dx ═ e (i, y)0)-e(i-1，y0) Wherein e (i, y)0) Denotes y0Characteristic value e (i-1, y) of pixel point with horizontal coordinate i0) Denotes y0Characteristic values of pixel points with horizontal coordinates of i-1 are obtained; finding out the abscissa (xmax and xmin) corresponding to the maximum value and the minimum value of the first derivative, and recording the midpoint coordinate of the two as x0With O1(x0，y0) As the center, a local to-be-processed region S is determined as shown in fig. 3.
Specifically, in the present embodiment, the height of the local region to be processed S
Width of
The area S is marked by a rectangle in fig. 4.
S6, removing blood vessels in the region S by using an image restoration method, and sequentially calculating the gradient G of each point in the region S to be processed by using a Sobel operator; at y0Finding out two peak points with the maximum gradient value change, and recording the peak points as peak and peak; at the same time, pixel point M (peak, y)0) And N (peak, y)0) As disc image in y0Two boundary points of a line, wherein peak<peak r. The positions of the M point and the N point in the local region to be processed S are shown as (1) in fig. 5. y is0The characteristic value curve and the gradient value curve of the line are shown in (2) and (3) of FIG. 5, and it can be seen from the figure that y0The gradient value curve of a row has two very prominent peak points, which correspond to the optic disc edge.
Among them, in the known Sobel operator,
let the characteristic value of the pixel point (i, j) be ei，jAnd the pixel values of 8 neighborhoods are as follows:
then: gx＝(ei-1，j+1+2ei，j+1+ei+1，j+1)-(ei-1，j-1+2ei，j-1+ei+1，j-1)；(3)
Gy＝(ei-1，j-1+2ei-1，j+ei-1，j+1)-(ei+1，j-1+2ei+1，j+ei+1，j+1)；(4)
The gradient value calculation formula of the pixel point (i, j) is as follows:
therefore, the gradient G of each point in the region S to be processed can be calculated by the equation (5).
S7, as shown in fig. 6, with point M as the starting point, looking up boundary points M ' adjacent to M one by one in the clockwise direction, then setting M ' to M ', repeating the looking up until the M is coincident with the N, and connecting all boundary points to be located at y0The upper optic disc edge; similarly, with the point N as the starting point, the boundary points N ' adjacent to the point N are searched one by one according to the same rotating direction, after the boundary points N ' are found, the boundary points N ' are set to be N ', the search is repeated until the boundary points N ' coincide with the point M, and all the boundary points are connected to be positioned at the point y0The lower optic disc edge.
The specific method for searching the adjacent boundary points comprises the following steps: firstly, setting a characteristic value threshold E and a gradient threshold T, and dividing points on two sides of a boundary point into two types: on one side of the boundary point, the characteristic value of the pixel point is greater than the characteristic value threshold value and is classified as the inside of the video disc, on the other side of the boundary point, the characteristic value of the pixel point is less than the characteristic value threshold value of the boundary point and is classified as the outside of the video disc, and then whether the adjacent points thereof meet the following conditions is searched one by one from the starting point: and in 8 neighborhoods of the determined boundary points, the gradient value of a certain point is greater than the gradient threshold value, at least one pixel point in the neighborhood of the point is positioned in the video disc, and if the gradient value is met, the pixel point is taken as the next boundary point. The gradient threshold value T can be 60-80% of the mean value of the peak point peak and peak of the gradient, and the characteristic value threshold value E can be the characteristic value of the last boundary point.
Specifically, when M point is used as a starting point, if the boundary point found finally does not coincide with N point, N is used as a starting point, and the adjacent boundary points are searched in the opposite direction until the boundary point is found to coincide with the boundary point found from M point or y is reached0Taking the central pixel point of the boundary points found twice as the boundary point finally found; similarly, when the point N is used as the starting point, if the boundary point found finally is not coincident with the point M, the adjacent boundary points are searched in the opposite direction by using the point M as the starting point until the found boundary point is coincident with the boundary point found from the point M or reaches y0And (4) taking the median of the positions of the boundary points found twice in the y direction, namely taking the central pixel point of the pixel points found twice on the same abscissa as the finally found boundary point.
Specifically, as shown in fig. 7, a characteristic analysis diagram of 8 neighborhood nodes of the pixel M is shown, in fig. 7, the pixel located at the edge of the video disk and inside the video disk is marked with gray, and the pixel located outside the video disk is marked with white. According to the characteristics of the optic disc, if a certain pixel point is an optic disc boundary point, the gradient value of the certain pixel point is obviously increased, and 8 neighborhood pixel points of the certain pixel point must simultaneously meet the following characteristics: at least one pixel point is positioned in the video disc.
And in the right direction of the M point, searching pixel points M 'meeting the boundary point condition in the neighborhood pixel points one by one according to the clockwise direction (namely according to the priority sequence of the upper right, the right side and the lower right), and recording the position of the M'. Setting M to M', repeating the searching process until the M is coincident with N. Fig. 8 is a schematic diagram showing 10 edge points found from the point M. If the distance is not coincident with N, the N is used as a starting point, pixel points meeting all the characteristics are searched in the opposite direction, the pixel point with the closest or same characteristic value is found, and the median value of all the pixel point positions between the pixel point and the N point is taken in the y direction. This is achieved byWhen y is0The upper edge is found.
In a similar way, y can be found starting from N0The lower edge. The video disc has a circular or oval characteristic, so that when boundary points M 'adjacent to M are searched one by one in the clockwise rotation direction, the boundary points M' are searched in the priority order of upper right, right side and lower right, and when boundary points N 'adjacent to N are searched one by one in the clockwise rotation direction, the boundary points M' are searched in the priority order of lower left, left side and upper left.
Further, in this embodiment, the boundary points may be searched clockwise from the M points and the N points, respectively, and when searching from the M points, the boundary points are searched in the preferred order of upper right, right side, and lower right, and when searching from the N points, the boundary points are searched in the preferred order of lower left, left side, and upper left.
Fig. 9 shows the result of executing the embodiment of the present invention, in which an edge curve drawn by all the found boundary points is displayed. It can be seen that the present invention accurately detects the edge of the disk. Therefore, the invention can be proved to be accurate and effective. Therefore, the invention simplifies the edge extraction process of the whole fundus image, determines the local range to be processed by a primary positioning method and reduces the processing range. In addition, two edge points are determined by finding a gradient extreme value in the maximum average characteristic value row, and then edge extraction is sequentially carried out by using the morphological characteristics of the optic disc through a characteristic value analysis method of 8 neighborhood nodes, so that the computation amount is effectively reduced.
Finally, it should be noted that: the above embodiments are only used to illustrate the technical solution of the present invention, and not to limit the same; while the invention has been described in detail and with reference to the foregoing embodiments, it will be understood by those skilled in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some or all of the technical features may be equivalently replaced; and the modifications or the substitutions do not make the essence of the corresponding technical solutions depart from the scope of the technical solutions of the embodiments of the present invention.
Claims (6)
1. A method for detecting the edge of an optic disc based on local features is characterized by comprising the following steps:
s1, smoothing the color fundus image to be processed by adopting a Gaussian filter, and recording as an image Grap;
s2, performing the following processing on the image gray: carrying out gray level transformation by a weighted average method to obtain the gray level value of each pixel point; obtaining the brightness value of each pixel point through HSV space transformation; taking the mean value of the gray value and the brightness value as a characteristic value e to obtain a new image Grap';
s3, establishing a rectangular coordinate system x-y taking pixels as a unit by taking the upper left corner of the Grap' as an origin;
s4, sequentially calculating the average value avgy of the characteristic values e of each pixel point of each line in the image Grap', and calculating the value y corresponding to the maximum value of avgy0I.e. the horizontal initial value row of the optic disc;
s5, finding y0First derivative, y, of the characteristic value e of each pixel point in a line0The calculation formula of the first derivative dx of the pixel point with the horizontal coordinate i is as follows: dx ═ e (i, y)0)-e(i-1，y0) Wherein e (i, y)0) Denotes y0Characteristic value e (i-1, y) of pixel point with horizontal coordinate i0) Denotes y0Characteristic values of pixel points with horizontal coordinates of i-1 are obtained; finding out the abscissa (xmax and xmin) corresponding to the maximum value and the minimum value of the first derivative, and recording the midpoint coordinate of the two as x0With O1(x0，y0) Determining a local area S to be processed as a center;
s6, removing blood vessels in the region S by using an image restoration method, and sequentially calculating the gradient G of each point in the region S to be processed by using a Sobel operator; at y0Finding out two peak points with the maximum gradient value change, and recording the peak points as peak and peak; at the same time, pixel point M (peak, y)0) And N (peak, y)0) As disc image in y0Two boundary points of a line, wherein peak<peakr；
S7, using M point as starting point, according to clockwise or anticlockwise rotation direction, searching boundary points M ' adjacent to M one by one, after finding out M-M ', repeating searching until the M-M ' is coincident with N point, connecting all boundary pointsThen becomes located at y0An upper or lower optic disc edge; similarly, with the point N as the starting point, the boundary points N ' adjacent to the point N are searched one by one according to the same rotating direction, after the boundary points N ' are found, the boundary points N ' are set to be N ', the search is repeated until the boundary points N ' coincide with the point M, and all the boundary points are connected to be positioned at the point y0A lower or upper optic disc edge;
the specific method for searching the adjacent boundary points comprises the following steps: setting a gradient threshold value T and a characteristic value threshold value, classifying pixel points with characteristic values larger than the characteristic value threshold value as the interior of a video disc, classifying pixel points with characteristic values smaller than the characteristic value threshold value as the exterior of the video disc, and searching whether adjacent points thereof meet the following conditions from a starting point one by one: and in 8 neighborhood pixel points of the determined boundary point, the gradient value of a certain point is greater than the gradient threshold value, at least one of the neighborhood pixel points of the certain point is positioned in the video disc, and if the gradient value of the certain point is met, the certain point is used as the next boundary point.
2. The method according to claim 1, wherein in step S7, if the boundary point found finally does not coincide with the point N when the point M is used as the starting point, then the method searches for the adjacent boundary points in the opposite direction using the point N as the starting point until the boundary point is found to coincide with the boundary point found from the point M or reach y0Taking the central pixel point of the boundary points found twice as the boundary point finally found; similarly, when the point N is used as the starting point, if the boundary point found finally is not coincident with the point M, the adjacent boundary points are searched in the opposite direction by using the point M as the starting point until the found boundary point is coincident with the boundary point found from the point M or reaches y0And taking the central pixel point of the boundary point found twice as the boundary point finally found.
3. The method according to claim 1, wherein in step S7, when searching for M 'S boundary points adjacent to M one by one in a clockwise rotation direction, the boundary points M' are searched in an upper-right, right-side, and lower-right priority order, and when searching for N 'S boundary points adjacent to N one by one in a clockwise rotation direction, the boundary points M' are searched in a lower-left, left-side, and upper-left priority order.
4. The method for detecting the edge of the optic disc based on the local features of claim 1, wherein in the step S6, the gradient threshold T is in a range of 60 to 80% of a mean value of peak values peak and peak of the gradient, and the feature threshold is a feature value of a previous boundary point.
5. The method according to claim 1, wherein in step S6, the formula for calculating the gradient G of each point in the region S to be processed is:
wherein the content of the first and second substances,
Gx＝(ei-1，j+1+2ei，j+1+ei+1，j+1)-(ei-1，j-1+2ei，j-1+ei+1，j-1)；
Gy＝(ei-1，j-1+2ei-1，j+ei-1，j+1)-(ei+1，j-1+2ei+1，j+ei+1，j+1)；
ei，jand (d) representing the characteristic value of the pixel point with the coordinate (i, j).
6. The method according to claim 1, wherein in step S5, the height and width of the local region S to be processed is 1/3 of the image height and width.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911348067.3A CN111192280B (en) | 2019-12-24 | 2019-12-24 | Method for detecting optic disc edge based on local feature |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911348067.3A CN111192280B (en) | 2019-12-24 | 2019-12-24 | Method for detecting optic disc edge based on local feature |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111192280A true CN111192280A (en) | 2020-05-22 |
| CN111192280B CN111192280B (en) | 2022-10-18 |
Family
ID=70707571
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911348067.3A Active CN111192280B (en) | 2019-12-24 | 2019-12-24 | Method for detecting optic disc edge based on local feature |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111192280B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111667402A (en) * | 2020-06-19 | 2020-09-15 | 洛阳师范学院 | Binary gray level changing method in high-precision laser engraving rubber plate making |
| CN113642588A (en) * | 2021-08-11 | 2021-11-12 | 深圳福鸽科技有限公司 | Method for analyzing crawler on outer edge of planar image object |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102842136A (en) * | 2012-07-19 | 2012-12-26 | 湘潭大学 | Optic disc projection location method synthesizing vascular distribution with video disc appearance characteristics |
| CN104794721A (en) * | 2015-04-30 | 2015-07-22 | 湘潭大学 | Quick optic disc positioning method based on multi-scale macula detection |
| CN107292868A (en) * | 2017-05-31 | 2017-10-24 | 瑞达昇科技(大连)有限公司 | A kind of optic disk localization method and device |
| WO2017186635A1 (en) * | 2016-04-26 | 2017-11-02 | Optos Plc | Retinal image processing |
| CN108846827A (en) * | 2018-04-16 | 2018-11-20 | 江南大学 | A method of based on more circle Fast Segmentation eyeground optic disks |
-
2019
- 2019-12-24 CN CN201911348067.3A patent/CN111192280B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102842136A (en) * | 2012-07-19 | 2012-12-26 | 湘潭大学 | Optic disc projection location method synthesizing vascular distribution with video disc appearance characteristics |
| CN104794721A (en) * | 2015-04-30 | 2015-07-22 | 湘潭大学 | Quick optic disc positioning method based on multi-scale macula detection |
| WO2017186635A1 (en) * | 2016-04-26 | 2017-11-02 | Optos Plc | Retinal image processing |
| CN107292868A (en) * | 2017-05-31 | 2017-10-24 | 瑞达昇科技(大连)有限公司 | A kind of optic disk localization method and device |
| CN108846827A (en) * | 2018-04-16 | 2018-11-20 | 江南大学 | A method of based on more circle Fast Segmentation eyeground optic disks |
Non-Patent Citations (4)
| Title |
|---|
| ARTURO AQUINO等: "Detecting the Optic Disc Boundary in Digital Fundus Images Using Morphological, Edge Detection, and Feature Extraction Techniques", 《IEEE TRANSACTIONS ON MEDICAL IMAGING》 * |
| YUAN GAO等: "Automatic rapid optic disc detection from infrared radiation retinal fundus images", 《2012 5TH INTERNATIONAL CONFERENCE ON BIOMEDICAL ENGINEERING AND INFORMATICS》 * |
| 姜平: "眼底图像分割方法研究", 《中国博士学位论文全文数据库 (信息科技辑)》 * |
| 张亚男: "结合改进型GBVS模型和眼底血管结构特征的视盘检测方法研究", 《中国优秀硕士学位论文全文数据库 (医药卫生科技辑)》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111667402A (en) * | 2020-06-19 | 2020-09-15 | 洛阳师范学院 | Binary gray level changing method in high-precision laser engraving rubber plate making |
| CN113642588A (en) * | 2021-08-11 | 2021-11-12 | 深圳福鸽科技有限公司 | Method for analyzing crawler on outer edge of planar image object |
| CN113642588B (en) * | 2021-08-11 | 2023-10-13 | 深圳福鸽科技有限公司 | Analysis method for outer edge crawler of planar image object |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111192280B (en) | 2022-10-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108596849B (en) | Single image defogging method based on sky region segmentation | |
| CN106934803B (en) | method and device for detecting surface defects of electronic device | |
| US9082165B2 (en) | Inspection of region of interest | |
| CN102426649B (en) | Simple steel seal digital automatic identification method with high accuracy rate | |
| Xiong et al. | An approach to locate optic disc in retinal images with pathological changes | |
| CN105844655B (en) | A kind of laser rays stripe edge extracting method | |
| CN111915704A (en) | Apple hierarchical identification method based on deep learning | |
| WO2013080868A1 (en) | Image processing device, image processing method, and image processing program | |
| CN110782477A (en) | Moving target rapid detection method based on sequence image and computer vision system | |
| WO2021109697A1 (en) | Character segmentation method and apparatus, and computer-readable storage medium | |
| CN107292835B (en) | Method and device for automatically vectorizing retinal blood vessels of fundus image | |
| CN106096491B (en) | Automatic identification method for microaneurysms in fundus color photographic image | |
| CN114118144A (en) | Anti-interference accurate aerial remote sensing image shadow detection method | |
| CN110647836B (en) | Robust single-target tracking method based on deep learning | |
| CN111192280B (en) | Method for detecting optic disc edge based on local feature | |
| CN112862744B (en) | Intelligent detection method for internal defects of capacitor based on ultrasonic image | |
| CN113205494B (en) | Infrared small target detection method and system based on adaptive scale image block weighting difference measurement | |
| CN112288780B (en) | Multi-feature dynamically weighted target tracking algorithm | |
| CN113781523A (en) | Football detection tracking method and device, electronic equipment and storage medium | |
| CN113379649A (en) | Image enhancement method, device, equipment and storage medium | |
| CN106372593B (en) | Optic disk area positioning method based on vascular convergence | |
| CN117197064A (en) | Automatic non-contact eye red degree analysis method | |
| CN116309780A (en) | Water gauge water level identification method based on target detection | |
| CN116012594A (en) | Fundus image feature extraction method, fundus image feature extraction device and diagnosis system | |
| CN110751064B (en) | Blink frequency analysis method and system based on image processing |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |