CN112446894A - Image segmentation method based on direction space - Google Patents
Image segmentation method based on direction space Download PDFInfo
- Publication number
- CN112446894A CN112446894A CN202011401152.4A CN202011401152A CN112446894A CN 112446894 A CN112446894 A CN 112446894A CN 202011401152 A CN202011401152 A CN 202011401152A CN 112446894 A CN112446894 A CN 112446894A
- Authority
- CN
- China
- Prior art keywords
- line segment
- gray
- convex
- concave
- maximum
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 44
- 238000003709 image segmentation Methods 0.000 title claims abstract description 23
- 230000007704 transition Effects 0.000 claims abstract description 32
- 238000012216 screening Methods 0.000 claims description 4
- 230000011218 segmentation Effects 0.000 claims description 4
- 238000003708 edge detection Methods 0.000 abstract description 7
- 238000012545 processing Methods 0.000 abstract description 3
- 230000000717 retained effect Effects 0.000 abstract description 2
- 238000004364 calculation method Methods 0.000 description 5
- 238000005286 illumination Methods 0.000 description 3
- 238000000605 extraction Methods 0.000 description 2
- 238000002372 labelling Methods 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000001174 ascending effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000014509 gene expression Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012067 mathematical method Methods 0.000 description 1
- 238000013106 supervised machine learning method Methods 0.000 description 1
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
- G06T7/12—Edge-based segmentation
-
- 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
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The invention belongs to the technical field of image processing, and particularly relates to an image segmentation method based on a direction space. The method and the device solve the problem of false mark of edge detection fundamentally, and are not influenced by image brightness, brightness distribution, resolution, target contrast, size and shape. Including defining a directional space; defining and searching a gray scale maximum point of a convex line segment and a gray scale minimum point of a concave line segment of the gray scale distribution curve in a direction space. Defining and calculating amplitudes of the convex line segment and the concave line segment and pixel points corresponding to the amplitudes; convex and concave line segments with large amplitudes are retained. Defining and searching the maximum value of the gradient between the convex line segment and the concave line segment and the pixel point corresponding to the maximum value. And combining the pixel points corresponding to the maximum values of the gradient of the transition line segments in the spaces in different directions. And combining the pixel points corresponding to the maximum gray values of the convex line segments and the pixel points corresponding to the minimum gray values of the concave line segments in the spaces in different directions.
Description
Technical Field
The invention belongs to the technical field of image processing, and particularly relates to an image segmentation method based on a direction space.
Background
The purpose of image segmentation is to extract the region or determine the location of the object of interest in the image. If the interested target in the image can not be extracted, the product with appearance defects on the production line is missed, and the focus in the medical picture is missed.
The image segmentation method based on edge detection is to calculate the gray level of pixels in one or more fixed windows by a certain mathematical method, and determine whether the pixels in the center of the window (or other positions of the window) belong to a target or a background according to the calculation result. The method is easily influenced by factors such as target size, contrast, boundary gradient, image illumination uniformity and the like, and is difficult to ensure that the target does not miss detection.
The image segmentation method based on the regional gray level threshold is to obtain a threshold through some mathematical operation globally or locally to determine whether the pixels in the range are the target or the background. This approach is effective for situations where there is no intersection of the target and background, but many images have difficulty ensuring that there is no grayscale intersection of the target and background.
The image segmentation method based on texture analysis is established on the basis of statistics, and the pixel belongs to a target and a background and is also determined by the gray level of the adjacent pixels through a certain operation result. This method is easily affected by the density, shape and contrast of the texture.
The image segmentation method based on template matching is to design a template similar to a target and search the target in the image by a similarity method. This method is susceptible to factors such as the gray level distribution within the target, the shape of the target, the contrast of the target, and the uniformity of the image illumination.
Based on the supervised machine learning method, one is that the feature extraction method is usually based on the above classical method, and therefore, the effectiveness of the selected feature extraction method is also involved; secondly, training of the classifier is based on the labeling, and errors of the labeling affect the training result of the classifier.
From the above discussion, each image segmentation method has certain limitations in application. Generally, the outline of a target is obtained by edge detection, and thus obtaining a target region is the mainstream of image segmentation.
In the field of image processing, the basic idea of the edge detection method is as follows: firstly, the gradient amplitude of each pixel point in the image is calculated to form a gradient image. In a gradient image, gradient magnitudes are classified into two categories, a large gradient magnitude and a small gradient magnitude. And marking the pixel points corresponding to the large gradient amplitude as the edges of the target. There is often no clear boundary between large and small gradient magnitudes, meaning that large gradient magnitudes do not necessarily correspond to edges and small gradient magnitudes do not necessarily correspond to edges. Therefore, edge marking errors are difficult to avoid, which is a common problem for edge detection. The marking error here includes two cases: the locations of real edges are not marked as edges and the locations of non-real edges are marked as edges.
The gradient amplitude is calculated in a window, and the size of the window and the calculation method of the gradient amplitude often influence the calculation result of the gradient amplitude. Therefore, various gradient amplitude calculation methods and multi-scale methods appear in succession, and although the error marking of the edge can be improved, the problem of error marking of edge detection is not fundamentally solved.
Disclosure of Invention
The invention provides an image segmentation method based on a direction space, which aims to fundamentally solve the problem of false mark of edge detection and is not influenced by image brightness, brightness distribution, resolution, target contrast, size and shape.
The invention is characterized in that: the gradient does not need to be calculated in a window, edges do not need to be determined by comparing the gradient sizes, the marked edges are unique, edge pixels do not need to be extracted in a two-dimensional space but in a one-dimensional space, and therefore the calculation amount is remarkably reduced.
In order to achieve the purpose, the invention adopts the following technical scheme that the method comprises the steps of defining a direction space;
defining and searching a gray scale maximum point of a convex line segment and a gray scale minimum point of a concave line segment of a gray scale distribution curve in a direction space;
defining and calculating amplitudes of the convex line segment and the concave line segment and pixel points corresponding to the amplitudes; keeping convex line segments and concave line segments with large amplitude;
(corresponding to a bright target and a dark target in the image, respectively; wherein convex line segments with large amplitudes correspond to bright targets in the image and concave line segments with large amplitudes correspond to dark targets in the image; and/or a dark target in the image)
Defining and searching a gradient maximum value between the convex line segment and the concave line segment and a pixel point corresponding to the maximum value;
combining pixel points corresponding to the maximum values of the gradients of the transition line segments in the spaces in different directions;
and combining the pixel points corresponding to the maximum gray values of the convex line segments and the pixel points corresponding to the minimum gray values of the concave line segments in the spaces in different directions.
Further, the defining direction space is a composition of decomposing the image space into an image direction space, that is, converting the target segmentation of the gray curved surface into the target segmentation of the gray curved surface.
Further, defining a convex line segment and a concave line segment of the gray distribution curve in the direction space, and calculating a pixel point corresponding to the gray maximum value of the convex line segment and a pixel point corresponding to the gray minimum value of the concave line segment.
Further, absolute amplitude, relative amplitude and average amplitude of convex line segments and concave line segments of the gray distribution curve are defined in the direction space.
Further, the amplitudes are divided into convex line segments and concave line segments with large amplitudes and convex line segments and concave line segments with small amplitudes through a mature method (for example, a triple standard deviation method, a maximum inter-class variance method and the like), and the convex line segments and the concave line segments with large amplitudes are reserved.
Furthermore, a transition line segment of the gray distribution curve is defined in the direction space, the gradient of the transition line segment is defined, and a pixel point corresponding to the maximum value of the gradient of the transition line segment is calculated.
Further, the pixel points corresponding to the maximum gradient values of the transition line segments in the spaces in different directions are merged together.
Further, the pixel points corresponding to the maximum gray values of the convex line segments and the pixel points corresponding to the minimum gray values of the concave line segments, which are reserved in the space in different directions through amplitude screening, are combined together.
Compared with the prior art, the invention has the beneficial effects.
The invention is not influenced by the size and shape of the target, the brightness and brightness distribution of the image and the change of the gray scale of the boundary of the target.
The invention converts the curved surface operation into the curve operation, and the calculated amount is obviously reduced.
Drawings
The invention is further described with reference to the following figures and detailed description. The scope of the invention is not limited to the following expressions.
FIG. 1 is a bright spot of the present hair.
Fig. 2 is a gray scale distribution curve of the present invention.
Detailed Description
The invention adopts the following technical scheme: an image segmentation method based on a direction space comprises the steps of defining the direction space; defining and searching a gray scale maximum point of a convex line segment and a gray scale minimum point of a concave line segment of a gray scale distribution curve in a direction space; defining and calculating amplitudes of the convex line segment and the concave line segment and pixel points corresponding to the amplitudes; keeping convex line segments and concave line segments with large amplitude; defining and searching a maximum value of the gradient of the transition line segment and a pixel point corresponding to the maximum value; combining pixel points corresponding to the maximum values of the gradients of the transition line segments in the spaces in different directions; and combining the pixel points corresponding to the maximum gray values of the convex line segments and the pixel points corresponding to the minimum gray values of the concave line segments in the spaces in different directions.
The digital image is a gray distribution curved surface in a Cartesian three-dimensional coordinate system space, wherein a plane formed by X-Y axes reflects the position coordinates of image pixels, and a numerical value of a Z axis reflects the gray value of the image. And the space in which the three-dimensional gray scale distribution curved surface is located is referred to as a gray scale space.
In an arbitrary plane perpendicular to the X-Y axis plane, a portion intersecting the curved surface of the gradation distribution is a curve called a gradation distribution curve. An angle θ between the vertical plane and the X axis is defined as a direction of the gray distribution curve, a set of the gray distribution curves in the θ direction forms a gray distribution curve in the θ direction, and a space in which the gray distribution curve in the θ direction is located is referred to as a gray space in the θ direction.
And the gray level distribution curve along the theta direction consists of a convex line segment and a concave line segment. The amplitudes of the convex line segments and the concave line segments can be respectively defined by the maximum gray value of the convex line segments and the minimum gray value of the concave line segments. The set of convex and concave line segment amplitudes is referred to as the theta-direction concave-convex line segment amplitude space.
Two gray difference values are formed between the gray maximum value of one convex line segment and the gray minimum values of two adjacent concave line segments, one difference value with a large difference value is defined as the relative amplitude of the convex line segment, one difference value with a small difference value is defined as the absolute amplitude of the convex line segment, and the average value of the two difference values is called as the average amplitude of the convex line segment. The convex segment magnitude is preceded by a "+" sign.
Two differences are formed between the gray minimum value of one concave line segment and the gray maximum values of two adjacent convex line segments, one difference with a large difference is defined as the relative amplitude of the concave line segment, one difference with a small difference is defined as the absolute amplitude of the concave line segment, and the average value of the two differences is called as the average amplitude of the convex line segment. The concave segment amplitude is preceded by a "-" sign.
And in the gray distribution curve of the theta direction space, if a convex line segment or a concave line segment with a significant amplitude exists, reserving the pixel point corresponding to the gray maximum value of the convex line segment and the pixel point corresponding to the gray minimum value of the concave line segment. The set of the pixels is called a theta-direction concave-convex line segment amplitude pixel space.
Along the theta direction gray level distribution curve, two transition line segments exist between the pixel point corresponding to the gray level maximum value of the convex line segment and the pixel point corresponding to the gray level minimum value of the adjacent concave line segment, wherein the two transition line segments are respectively a transition line segment which monotonically rises and a transition line segment which monotonically falls, and the transition line segments are collectively called transition line segments. The change in gray level is usually called a gradient, and at least one maximum in gradient exists in each transition line segment. The set of the pixel points corresponding to the maximum gradient value of the transition line segment is called a theta direction transition line segment gradient extreme point space.
And merging the gradient extreme points of the transition line segments in different directions to form a transition line segment gradient extreme point space.
The image target boundary is composed of transition line segment gradient extreme points.
Embodiments of the present invention are further described below in conjunction with FIG. 1:
in this embodiment, fig. 1 shows a screenshot of a bright spot in a fundus image, and fig. 2 shows a gray scale distribution curve passing through the bright spot in the horizontal direction.
In this embodiment, in the gray distribution curve of fig. 2, the 2 nd arrow from the left points to the pixel point corresponding to the gray maximum of the convex line segment. The 1 st and 3 rd arrows from the left point to the pixel points corresponding to the minimum gray value of the concave line segment respectively.
In this embodiment, in the gray-scale distribution curve of fig. 2, the first vertical line from the left corresponds to the pixel point corresponding to the maximum gradient value of the ascending transition line segment, and the second vertical line from the left corresponds to the pixel point corresponding to the maximum gradient value of the descending transition line segment.
In this embodiment, the following method may be adopted to find the pixel point corresponding to the maximum gray value of the convex line segment. Setting f (j) to represent the gray value of the gray distribution curve at the pixel point j; f (j-1) represents the gray value of the gray distribution curve at the pixel point (j-1); f (j +1) represents the gray value of the gray distribution curve at the pixel point (j + 1); if f (j) > f (j-1) and f (j) > f (j +1) exist, the pixel point j is the pixel point corresponding to the maximum gray value of the convex line segment. If f (j) > f (j-1), f (j) ═ f (j +1) and f (j +1) > f (j +2 exist, then the pixel points j and (j +1) are the pixel points corresponding to the maximum gray value of the convex line segment, at this time, the pixel point j is selected as the point of the maximum value, and the (j +1) is not marked. The searching method for more than two gray maximum points is based on the principle as above, and the maximum point close to the middle is selected as the mark point in principle.
In this embodiment, the following method is adopted to search for the pixel point corresponding to the minimum gray value of the concave line segment. Setting f (j) to express the gray value of the gray distribution curve pixel point j; f (j-1) represents the gray value of the gray distribution curve pixel point (j-1); f (j +1) represents the gray value of the gray distribution curve pixel point (j + 1); if f (j) < f (j-1) and f (j) < f (j +1), the pixel point j is the pixel point corresponding to the minimum gray value of the concave line segment. If f (j) < f (j-1), f (j) < f (j +1), f (j +1) < f (j +2), the pixel points j and (j +1) are both pixel points corresponding to the gray minimum value of the concave line segment, at this time, the pixel point j is selected as the pixel point corresponding to the gray minimum value, and the pixel point (j +1) is not marked. The searching method for more than two gray minimum value points is based on the principle that the minimum value point close to the middle is selected as the mark point in principle.
In this embodiment, the following method may be adopted for determining the amplitude of the convex line segment. Setting f (J) to represent the gray value of the maximum gray value point J of the convex line segment; f (J-1) represents the gray value of the gray minimum value point (J-1) of the previous concave line segment adjacent to the current convex line segment; f (J +1) represents the gray value of the next concave line segment gray minimum value point (J +1) adjacent to the current convex line segment; if f (J-1) < f (J +1), let fr(J) F (J) -f (J-1) represents the relative amplitude of the convex segment of the J point; f. ofa(J) F (J) -f (J +1) represents the absolute amplitude of the convex segment at point J. Otherwise, let fr(J) F (J) -f (J +1) represents the relative amplitude of the convex segment of the J point; f. ofa(J) F (J) -f (J-1) represents the absolute amplitude of the convex segment at point J.
In this embodiment, the following method may be adopted for determining the amplitude of the concave line segment. Setting f (J) to represent the gray value of the concave line segment gray minimum value point J; f (J-1) represents the gray value of the gray maximum point (J-1) of the previous convex line segment adjacent to the current concave line segment; f (J +1) represents the gray value of the gray maximum value point (J +1) of the next convex line segment adjacent to the current concave line segment; if f (J-1) > f (J +1), let fr(J) F (J-1) -f (J) represents the relative amplitude of the concave segment at point J; f. ofa(J) F (J +1) -f (J) represents the absolute amplitude of the concave segment at point J. Otherwise, let fr(J) F (J +1) -f (J) represents the relative amplitude of the concave segment at point J; f. ofa(J) F (J-1) -f (J) represents the absolute amplitude of the concave segment at point J.
In this embodiment, the screening of the convex line segment having a large amplitude may be performed as follows. Let f (J)n) The amplitude of the n-th convex line segment in the gray distribution curve is expressed if f (J)n) The convex line segment is retained if the following condition is satisfied.
Wherein
N is the number of convex line segments.
In the present embodiment, the screening method of the concave line segment having a large amplitude is similar to that of the previous embodiment.
In the present embodiment, the gradient of the transition line segment between the adjacent convex and concave line segments having large amplitudes is defined as the gray scale difference value that is one pixel apart, i.e., Δ f (j) ═ f (j +1) -f (j-1). When Δ f (j) is max Δ f (j), the pixel point j is the pixel point corresponding to the maximum gradient value.
And combining the pixel points corresponding to the maximum gradient values of the transition line segments of all the gray distribution curves in the horizontal direction and the vertical direction to obtain the boundary pixels of the target.
The invention discloses an image segmentation method, which can synthesize the maximum value points of the gradient of a transition line segment extracted from gray level distribution curves in different directions together to obtain the target boundary of an image. The convex line segment gray scale maximum value points and the concave line segment minimum value points extracted from the gray scale distribution curves in different directions are combined together, and the frameworks of the bright linear target and the dark linear target of the image can be obtained respectively. The method specifically comprises the following parts: extracting a convex line section gray scale maximum value point and a concave line section gray scale minimum value point along a direction gray scale distribution curve; respectively calculating the absolute amplitude, the relative amplitude and the average amplitude of the convex line segment and the concave line segment; selecting and reserving a convex line segment and a concave line segment with large amplitude; extracting a transition line segment gradient maximum value point along a direction gray distribution curve; combining the convex line segment gray scale maximum value points and the concave line segment gray scale minimum value points in different directions; and merging the gradient maximum points of the transition line segments in different directions.
The invention discloses an image segmentation method, which converts the gray curved surface operation of a three-dimensional space into the gray curve operation of a two-dimensional space, and is characterized in that: the method is simple in operation and independent of the number, brightness, size, shape, contrast, boundary gradient and illumination uniformity of the target in the image, and particularly gets rid of the constraint of a method for determining whether the image pixel is the target or the background through a certain operation in a fixed window.
It should be understood that the detailed description of the present invention is only for illustrating the present invention and is not limited by the technical solutions described in the embodiments of the present invention, and those skilled in the art should understand that the present invention can be modified or substituted equally to achieve the same technical effects; as long as the use requirements are met, the method is within the protection scope of the invention.
Claims (8)
1. The image segmentation method based on the direction space is characterized by comprising the following steps of:
defining a direction space;
defining and searching a gray scale maximum point of a convex line segment and a gray scale minimum point of a concave line segment of a gray scale distribution curve in a direction space;
defining and calculating amplitudes of the convex line segment and the concave line segment and pixel points corresponding to the amplitudes; keeping convex line segments and concave line segments with large amplitude; respectively corresponding to a bright target and a dark target in the image;
defining and searching a gradient maximum value between the convex line segment and the concave line segment and a pixel point corresponding to the maximum value;
combining pixel points corresponding to the maximum values of the gradients of the transition line segments in the spaces in different directions;
and combining the pixel points corresponding to the maximum gray values of the convex line segments and the pixel points corresponding to the minimum gray values of the concave line segments in the spaces in different directions.
2. The method of image segmentation based on directional space according to claim 1, characterized in that: the definition direction space is a composition of decomposing the image space into the image direction space, that is, converting the target segmentation of the gray curved surface into the target segmentation of the gray curve.
3. The method of image segmentation based on directional space according to claim 1, characterized in that: and defining a convex line segment and a concave line segment of the gray distribution curve in the direction space, and calculating a pixel point corresponding to the gray maximum value of the convex line segment and a pixel point corresponding to the gray minimum value of the concave line segment.
4. The method of image segmentation based on directional space according to claim 1, characterized in that: and defining the absolute amplitude, the relative amplitude and the average amplitude of the convex line segment and the concave line segment of the gray distribution curve in the direction space.
5. The method of image segmentation based on directional space according to claim 1, characterized in that: dividing the amplitude into a convex line segment and a concave line segment with large amplitude, a convex line segment and a concave line segment with small amplitude, and reserving the convex line segment and the concave line segment with large amplitude.
6. The method of image segmentation based on directional space according to claim 1, characterized in that: and defining a transition line segment of the gray distribution curve in the direction space, defining the gradient of the transition line segment, and calculating a pixel point corresponding to the maximum value of the gradient of the transition line segment.
7. The method of image segmentation based on directional space according to claim 1, characterized in that: and combining the pixel points corresponding to the maximum gradient values of the transition line segments in the spaces in different directions.
8. The method of image segmentation based on directional space according to claim 1, characterized in that: and combining the pixel points corresponding to the maximum gray values of the convex line segments and the pixel points corresponding to the minimum gray values of the concave line segments which are reserved in the space with different directions through amplitude screening.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011401152.4A CN112446894B (en) | 2020-12-04 | 2020-12-04 | Image segmentation method based on direction space |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011401152.4A CN112446894B (en) | 2020-12-04 | 2020-12-04 | Image segmentation method based on direction space |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112446894A true CN112446894A (en) | 2021-03-05 |
| CN112446894B CN112446894B (en) | 2024-03-26 |
Family
ID=74740056
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011401152.4A Active CN112446894B (en) | 2020-12-04 | 2020-12-04 | Image segmentation method based on direction space |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112446894B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080069421A1 (en) * | 2006-09-14 | 2008-03-20 | Siemens Medical Solutions Usa Inc. | Efficient Border Extraction Of Image Feature |
| CN102819841A (en) * | 2012-07-30 | 2012-12-12 | 中国科学院自动化研究所 | Global threshold partitioning method for partitioning target image |
| CN104077773A (en) * | 2014-06-23 | 2014-10-01 | 北京京东方视讯科技有限公司 | Image edge detection method, and image target identification method and device |
| CN105678750A (en) * | 2015-12-31 | 2016-06-15 | 上海联影医疗科技有限公司 | Gray scale mapping curve generation method and apparatus for medical images |
| CN107705314A (en) * | 2017-11-01 | 2018-02-16 | 齐鲁工业大学 | A kind of more subject image dividing methods based on intensity profile |
| CN111652825A (en) * | 2020-08-06 | 2020-09-11 | 上海海栎创微电子有限公司 | Edge tracking straight line segment rapid detection device and method based on gradient direction constraint |
-
2020
- 2020-12-04 CN CN202011401152.4A patent/CN112446894B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080069421A1 (en) * | 2006-09-14 | 2008-03-20 | Siemens Medical Solutions Usa Inc. | Efficient Border Extraction Of Image Feature |
| CN102819841A (en) * | 2012-07-30 | 2012-12-12 | 中国科学院自动化研究所 | Global threshold partitioning method for partitioning target image |
| CN104077773A (en) * | 2014-06-23 | 2014-10-01 | 北京京东方视讯科技有限公司 | Image edge detection method, and image target identification method and device |
| CN105678750A (en) * | 2015-12-31 | 2016-06-15 | 上海联影医疗科技有限公司 | Gray scale mapping curve generation method and apparatus for medical images |
| CN107705314A (en) * | 2017-11-01 | 2018-02-16 | 齐鲁工业大学 | A kind of more subject image dividing methods based on intensity profile |
| CN111652825A (en) * | 2020-08-06 | 2020-09-11 | 上海海栎创微电子有限公司 | Edge tracking straight line segment rapid detection device and method based on gradient direction constraint |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112446894B (en) | 2024-03-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109978839B (en) | Method for detecting wafer low-texture defects | |
| CN104715238B (en) | A kind of pedestrian detection method based on multi-feature fusion | |
| CN102629322B (en) | Character feature extraction method based on stroke shape of boundary point and application thereof | |
| CN110197153B (en) | Automatic wall identification method in house type graph | |
| CN109242870A (en) | A kind of sea horizon detection method divided based on image with textural characteristics | |
| CN105374015A (en) | Binary method for low-quality document image based on local contract and estimation of stroke width | |
| CN115170669B (en) | Identification and positioning method and system based on edge feature point set registration and storage medium | |
| JP4946878B2 (en) | Image identification apparatus and program | |
| CN107154044B (en) | Chinese food image segmentation method | |
| CN107895151A (en) | Method for detecting lane lines based on machine vision under a kind of high light conditions | |
| CN111754538B (en) | Threshold segmentation method for USB surface defect detection | |
| CN110378924A (en) | Level set image segmentation method based on local entropy | |
| CN104123554A (en) | SIFT image characteristic extraction method based on MMTD | |
| CN109961065B (en) | Sea surface ship target detection method | |
| CN112101108A (en) | Left-right-to-pass sign identification method based on pole position characteristics of graph | |
| CN103914829A (en) | Method for detecting edge of noisy image | |
| CN114863492A (en) | Method and device for repairing low-quality fingerprint image | |
| CN112686872B (en) | Wood counting method based on deep learning | |
| CN111178210B (en) | Image identification and alignment method for cross mark | |
| CN104408726A (en) | Triangle-based image contour matching algorithm | |
| CN108717699B (en) | Ultrasonic image segmentation method based on continuous minimum segmentation | |
| CN111563410A (en) | Processing method for detecting motion speed of foam image | |
| CN112446894B (en) | Image segmentation method based on direction space | |
| CN112991395B (en) | Vision tracking method based on foreground condition probability optimization scale and angle | |
| CN110490885B (en) | Improved adaptive threshold value binarization method and VIN code character segmentation method |
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 |