CN102915538B - A kind of bianry image automatic Selection of Image Threshold based on biological vision - Google Patents
A kind of bianry image automatic Selection of Image Threshold based on biological vision Download PDFInfo
- Publication number
- CN102915538B CN102915538B CN201210346068.6A CN201210346068A CN102915538B CN 102915538 B CN102915538 B CN 102915538B CN 201210346068 A CN201210346068 A CN 201210346068A CN 102915538 B CN102915538 B CN 102915538B
- Authority
- CN
- China
- Prior art keywords
- array
- image
- gray level
- index value
- apply
- 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
Landscapes
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention provides a kind of bianry image automatic Selection of Image Threshold based on biological vision, it is characterized in that: the method is specifically implemented as follows: the first step: apply for the shaping array a that a length is 256, pixel number corresponding each gray level of gray level image histogram is deposited in array a by gray level incremental order; Second step: apply for again one long be 256 shaping array b, by the difference Δ of a array adjacent element, deposit in successively in array b; The 3rd step: finding out Δ from array b is non-negative two the longest sections continuously, records the index value of hypomere tail in array b; The 4th step: draw according to two section pygochords that obtained in the 3rd step, try to achieve in array b corresponding to the index value that is worth minimum element between two index values, this index value is required segmentation threshold. The present invention is in medical image cell segmentation process, and speed is very fast, and effect is very good, has reached the ability of real-time processing.
Description
Technical field
The present invention relates to Computer Image Processing, field of medical image processing, is specifically a kind of based on biological visionBianry image automatic Selection of Image Threshold.
Background technology
It is in graphical analysis and computer vision science to be substantially the most also one of the most difficult technology that image is cut apart, and it is baseBetween the similitude of same interior of articles feature and different objects, digital picture is divided into not phase by the diversity of this featureHand over the process of the regional of (not overlapping). Choosing of feature can be used gray scale, form factor, texture etc. Image is cut apartMethod a lot, classical based on the histogrammic thresholding method of pixel, the region-growing method based on neighborhood etc. In recent years also notBreak and emerged many new methods: as the cutting techniques based on wavelet analysis and conversion, based on the skill of cutting apart of mathematical morphologyArt etc.
Automatic threshold is used for the background of gray level image to separate with target, in real image, and image object and the back of the bodyBetween scape, do not have distinct gray scale, therefore the selection of threshold value is extremely important, precision and figure that directly impact is cut apartThe correctness that picture is processed. The mathematical theory that mostly foundation is complicated of image partition method in the past and loaded down with trivial details computational process, therebyWhile causing computer to be processed, need to carry out the calculating of magnanimity, expend a large amount of time, can not reach the ability of real-time processing, andIn most cases also can not get desirable result, especially a difficult problem for medical image real-time analysis.
Summary of the invention
The object of the invention is to provides a kind of bianry image automatic Selection of Image Threshold based on biological vision at this, shouldMethod can solve occur in image segmentation algorithm at present consuming time, segmentation effect is not obvious, has solved medical image real-timeA difficult problem of analyzing. This method is mainly utilized current C PU to carry out the fastest " addition instruction " and " contrast instruction " and is completed binary mapAs obtaining of threshold value, support irregular discrete histogram, possess operation stable at a high speed, thinking is succinct.
The present invention is achieved in that a kind of bianry image automatic Selection of Image Threshold based on biological vision of structure, itsBe characterised in that: the method is specifically implemented as follows:
The first step: apply for the shaping array a that a length is 256, by corresponding each gray level of gray level image histogramPixel number deposits in array a by gray level incremental order;
Second step: apply for again one long be 256 shaping array b, by the difference Δ of a array adjacent element, deposit successively number inIn group b;
The 3rd step: finding out Δ from array b is non-negative two the longest sections continuously, records hypomere tail at array bIn index value,
Set a length n, when occurring that a certain Δ is for negative, and within the scope of the n of its front and back, Δ, all for just, is ignored current ΔBe 0;
The 4th step: draw according to two section pygochords that obtained in the 3rd step, try to achieve in array b corresponding to two index valuesBetween be worth the index value of minimum element, this index value is required segmentation threshold.
The present invention has adopted based on a kind of simple automatic Selection of Image Threshold that is applicable to medical image segmentation of biological vision,That is to say people and see two crests, when there is a trench centre, find this point of the lowest point, people first soCan, by contrast, find two wave crest points, then between wave crest point, find trough. Algorithm has adopted computer execution speedFast addition instruction and compare instruction, make that the method amount of calculation is little, computational speed is fast, it is rapid to obtain result, at medical imageThe effect of cell segmentation aspect is very good.
Beneficial effect of the present invention is: the method can solve occur in image segmentation algorithm at present consuming time, pointCut DeGrain, can solve a difficult problem for medical image real-time analysis, in medical image cell segmentation process, speed veryRapidly, effect is very good, has reached the ability of real-time processing. Be of a size of 768 × 576 gray level image for a width, while cutting apartBetween in 10ms.
Brief description of the drawings
Fig. 1 is schematic flow sheet of the present invention
Fig. 2 is the mathematical theory basis schematic diagram directly perceived of describing the method
Fig. 3 is that experiment correspondence is cut apart former figure
Fig. 4 is segmentation effect figure.
Detailed description of the invention
The invention provides a kind of bianry image automatic Selection of Image Threshold based on biological vision, the present invention has mainly solvedIn image segmentation algorithm, occur at present consuming time, segmentation effect is not obvious, has solved a difficult problem for medical image real-time analysis. ThisMethod is mainly utilized current C PU to carry out the fastest " addition instruction " and " contrast instruction " and is completed the obtaining of bianry image threshold value,Support irregular discrete histogram, possess operation stable at a high speed, thinking is succinct.
Be illustrated in fig. 2 shown below, our target is the value in the case of only knowing f (x), by one method fast and accuratelyFind the minimum point in function. Human eye just knows that by contrast minimum point is c point, and our algorithm has also utilized this spy justPoint. The mathematical theory of algorithm is described below:
The first step: function is in intervalUpper, function increases progressively, for arbitrarilyAnd，Function perseverance has,, exist in certain large region and have the poor of functional valuePerseverance is more than or equal to 0.
Second step: in like manner, function is in intervalIf on deposit equally，Exist in certain large region and have the difference perseverance of functional value to be more than or equal to 0.
The 3rd step: can search out two crest summit b by the first step and second step, d, so below only need to beIn interval, search out minimum point, i.e. c point.
Bianry image automatic Selection of Image Threshold based on biological vision of the present invention, realizes in the following way; AsShown in Fig. 1; The method is specifically implemented as follows:
The first step: apply for the shaping array a that a length is 256, by corresponding each gray level of gray level image histogramPixel number deposits in array a by gray level incremental order.
Second step: apply for again one long be 256 shaping array b, by the difference Δ of a array adjacent element, deposit successively number inIn group b.
The 3rd step: finding out Δ from array b is non-negative two the longest sections continuously, records hypomere tail at array bIn index value; (set a length n, when occurring that a certain Δ is for negative, and before and after it within the scope of n Δ all for just, by current ΔIgnore is 0); Why the reason of design is so herein: the grey level histogram of the image that we obtain on a certain interval notMay be to increase progressively continuously, always it is corresponding can not to be greater than transverse axis previous moment institute with the value of a moment longitudinal axis after the increase of transverse axisThe value of the longitudinal axis, this is called monotonic increase on mathematics, owing to there is so little fluctuation, we have added processing belowProcess, sets a length n, and when occurring that a certain Δ is for negative, and within the scope of the n of its front and back, Δ, all for just, is ignored current ΔBe 0; So just can get rid of due to histogram fluctuation cause threshold value select wrong may, thereby ensured to cut apart correctProperty.
The 4th step: draw according to two section pygochords that obtained in the 3rd step, try to achieve in array b corresponding to two index valuesBetween be worth the index value of minimum element, this index value is required segmentation threshold.
Showing through experiment, be of a size of 768 × 576 gray level image for a width, is IntelT5750 at a CPU,Operating system is WindowsXP, and translation and compiling environment is VS2010, and carries out image sliced time is 8MS, has reached real-time processing completelyAbility, cut apart former figure as shown in Figure 3, segmentation effect figure as shown in Figure 4.
Claims (1)
1. the bianry image automatic Selection of Image Threshold based on biological vision, is characterized in that: the method is concrete implement asUnder:
The first step: apply for the shaping array a that a length is 256, by pixel corresponding to each gray level of gray level image histogramCount and deposit in array a by gray level incremental order;
Second step: apply for again one long be 256 shaping array b, by the difference Δ of array a adjacent element, deposit successively array b inIn;
The 3rd step: finding out Δ from array b is continuously non-negative two the longest sections, records hypomere tail in array bIndex value; Set a length n, when occurring that a certain Δ is for negative, and within the scope of the n of its front and back, Δ, all for just, is ignored current ΔBe 0; Its object be to get rid of due to histogram fluctuation cause threshold value select wrong may, thereby ensured to cut apart correctProperty;
The 4th step: draw according to two section pygochords that obtained in the 3rd step, try to achieve in array b corresponding between two index valuesThe index value that is worth minimum element, this index value is required segmentation threshold.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210346068.6A CN102915538B (en) | 2012-09-18 | 2012-09-18 | A kind of bianry image automatic Selection of Image Threshold based on biological vision |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210346068.6A CN102915538B (en) | 2012-09-18 | 2012-09-18 | A kind of bianry image automatic Selection of Image Threshold based on biological vision |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102915538A CN102915538A (en) | 2013-02-06 |
| CN102915538B true CN102915538B (en) | 2016-05-18 |
Family
ID=47613891
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201210346068.6A Active CN102915538B (en) | 2012-09-18 | 2012-09-18 | A kind of bianry image automatic Selection of Image Threshold based on biological vision |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102915538B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101106716A (en) * | 2007-08-21 | 2008-01-16 | 北京大学软件与微电子学院 | A shed image division processing method |
| CN101236607A (en) * | 2008-03-03 | 2008-08-06 | 哈尔滨工程大学 | Rapid multi- threshold value dividing method for gray-scale image |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7298898B2 (en) * | 2003-04-15 | 2007-11-20 | Honeywell International Inc. | System and method for image segmentation |
-
2012
- 2012-09-18 CN CN201210346068.6A patent/CN102915538B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101106716A (en) * | 2007-08-21 | 2008-01-16 | 北京大学软件与微电子学院 | A shed image division processing method |
| CN101236607A (en) * | 2008-03-03 | 2008-08-06 | 哈尔滨工程大学 | Rapid multi- threshold value dividing method for gray-scale image |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102915538A (en) | 2013-02-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102915541B (en) | Multi-scale image segmenting method | |
| CN108205703B (en) | Multi-input multi-output matrix average value pooling vectorization implementation method | |
| CN105893585B (en) | A kind of bigraph (bipartite graph) model academic paper recommended method of combination tag data | |
| CN103310481A (en) | Point cloud reduction method based on fuzzy entropy iteration | |
| CN103559724A (en) | Method for synchronously tracking multiple cells in high-adhesion cell environment | |
| CN107945200A (en) | Image binaryzation dividing method | |
| CN111488917A (en) | Garbage image fine-grained classification method based on incremental learning | |
| JP2012042990A (en) | Image identification information adding program and image identification information adding apparatus | |
| CN112270259B (en) | SAR image ship target rapid detection method based on lightweight convolutional neural network | |
| CN105512206A (en) | Outlier detection method based on clustering | |
| CN113487610B (en) | Herpes image recognition method and device, computer equipment and storage medium | |
| CN103870834A (en) | Method for searching for sliding window based on layered segmentation | |
| CN102903121A (en) | Fusion algorithm based on moving target tracking | |
| CN103886585A (en) | Video tracking method based on rank learning | |
| CN103984746A (en) | Semi-supervised classification and regional distance measurement based SAR (Synthetic Aperture Radar) image identification method | |
| CN110659724B (en) | Target detection depth convolution neural network construction method based on target scale | |
| CN112766491A (en) | Neural network compression method based on Taylor expansion and data driving | |
| CN103198489A (en) | Automatic detection method of salient object based on salience density and edge response | |
| CN102163321B (en) | Image segmentation method based on lattice Boltzman model | |
| CN113096079B (en) | Image analysis system and construction method thereof | |
| Wang et al. | A novel fault diagnosis scheme for rolling bearing based on symbolic aggregate approximation and convolutional neural network with channel attention | |
| CN102915538B (en) | A kind of bianry image automatic Selection of Image Threshold based on biological vision | |
| CN103955946A (en) | Ant colony neighbor cell tracking system based on cooperation and competition mode and application thereof | |
| CN117636032A (en) | Multi-label image classification method based on multi-scale local features and difficult class mining | |
| CN102184524B (en) | Neighbourhood learning culture gene image segmentation method based on standard cut |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |