CN102693533A - Medical digital image mosaicing method - Google Patents
Medical digital image mosaicing method Download PDFInfo
- Publication number
- CN102693533A CN102693533A CN2012100635443A CN201210063544A CN102693533A CN 102693533 A CN102693533 A CN 102693533A CN 2012100635443 A CN2012100635443 A CN 2012100635443A CN 201210063544 A CN201210063544 A CN 201210063544A CN 102693533 A CN102693533 A CN 102693533A
- Authority
- CN
- China
- Prior art keywords
- image
- window
- pixel
- medical digital
- digital images
- 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.)
- Pending
Links
Images
Landscapes
- Image Processing (AREA)
Abstract
The invention relates to the technical field of radiation imaging, and in particular relates to a medical digital image mosaicing method. The method comprises the following steps: carrying out brightness equalization, noise point removal and sharpening processing on a to-be-mosaiced image; and on this basis, carrying out feature extraction, feature matching, error matching removal, image transformation and image fusion on the to-be-mosaiced image, thereby achieving an effect of image mosaicing. According to the invention, the mosaicing under poor conditions can be processed through a strong preprocessing ability, and higher mosaicing accuracy can be achieved through strong accuracy control.
Description
Technical field
The invention belongs to the radiant image technical field, relate in particular to a kind of medical digital images joining method.
Background technology
The digital picture splicing is through splicing a kind of technology of obtaining bigger field-of-view image to having two width of cloth or several images with overlapping region.At biomedical sector, particularly in diagnosis of the orthopaedics of DR image and the treatment high clinical value is arranged.
Traditional biomedical collecting device is because the restriction of himself hardware condition, can not disposable high resolving power, no shifting ground all takes medical image.Though disposable the collecting of main equipment ability arranged, and expense is high, popularity is little.
Summary of the invention
Deficiencies such as the resolution to mentioning the existence of traditional images collecting device in the above-mentioned background technology is low, the precision of images is not high, be difficult for popularizing the present invention proposes a kind of medical digital images joining method.
Technical scheme of the present invention is that a kind of medical digital images joining method is characterized in that this method may further comprise the steps:
Step 1: to wanting spliced image to carry out luminance proportion, carrying out noise removal and image sharpening processing;
Step 2: on the basis of step 1, spliced image is carried out feature extraction, characteristic matching, mating removal and image transformation and image co-registration realizes image mosaic by mistake to wanting.
The method of said luminance proportion is the histogram matching method.
The method that said noise is removed is the traversal entire image; Getting a window around each pixel adds up; If the value of this pixel is bigger more than 1.5 times than the mean value of the pixel in the window; Can think that then this pixel is a noise, its pixel value is replaced with the average in its window.
The computing formula of the average in the said window is:
Wherein:
x
N/2Pixel value average for the window intermediate point;
x
jPixel value for j in window point;
N is a window interior pixel number.
The formula of said image sharpening is:
Wherein, sharpness is the sharpening extent index.
The method of said characteristic matching is an optical flow method.
The method that said mistake coupling is removed is relevant matches method or Ransac method.
The present invention be directed to a kind of joining method that the splicing problem of medical digital images is proposed.This method can be handled the splicing under the poor condition through stronger pre-service ability, and can reach higher splicing precision through stronger precision control.
Description of drawings
Fig. 1 is the image split-joint method block diagram.
Embodiment
Below in conjunction with accompanying drawing, preferred embodiment is elaborated.Should be emphasized that following explanation only is exemplary, rather than in order to limit scope of the present invention and application thereof.
(1) system chart
The method that the present invention proposes mainly comprises two steps: at first carries out pre-service, carries out the selection of feature extraction and unique point then remaining stitching image, and as shown in Figure 1.
(2) main modular
1, pre-service
Because in practical application; Two width of cloth to be spliced or multiple image might be from different shooting conditions; Perhaps different collecting devices; Perhaps under different environment and time, collect, cause between the image to be matched apparent in view difference being arranged, if do not carry out the requirement that splicing might have just been satisfied not in some pre-service.The present invention is directed to various complicated situations has proposed: the preprocessing process of luminance proportion, noise removal and image sharpening, the experiment proof can improve the precision of splicing greatly.
1) luminance proportion
Suppose that two width of cloth images splice,, then can directly influence the extraction of back unique point and the coupling of unique point if two width of cloth brightness of image are widely different.So need equilibrium be carried out in the brightness of two width of cloth figure, generally adopted histogram matching to realize.
2) noise is removed
Medical image, especially DR image inevitably have some and cross generation bright or that cross dark noise in the process of gathering, the precision and the speed of these noises meeting effect characteristics point couplings at random.The method that noise is removed is the traversal entire image; Getting a window (3 * 3 pixels or 5 * 5 pixels) around each pixel adds up; If the value of this pixel is bigger more than 1.5 times than the mean value of the pixel in the window; Then can think a noise, its pixel value is replaced with the average in its window.It is following that noise is removed expression formula:
X={x
1,…x
j,…x
n},
Wherein:
X is the pixel value of each point in the window;
x
jPixel value for j in window point;
x
N/2Pixel value average for the window intermediate point;
N is a window interior pixel number, if this value is all big more a lot of than other pixel values, then the value of this pixel replaces with other all pixel averages.
3) image sharpening
The effect of image sharpening is to be used for strengthening the edge, makes characteristic more obvious, thereby in the subsequent characteristics extraction module, can extract more and than the unique point of horn of plenty.In addition, the image that has edge in gatherer process is obvious inadequately, and unique point is not outstanding, can cause characteristic extracting module to extract less than unique point, thereby causes the splicing failure.Sharpening method is following:
Wherein, sharpness is the sharpening extent index, and span is 1--100.
Its ultimate principle is each pixel on the traversing graph picture; Around each pixel, get a window, suppose that window interior pixel number is n, if the mean value difference of the pixel in this pixel value and the window is bigger; Then amplify this difference through sharpness; Especially on the edge of the time, this difference is big more, and then the edge is sharp keen more.If at flat site, because central point pixel value and surrounding pixel value are comparatively approaching, and do not have the effect of sharpening, so this effect that is actually a kind of edge sharpening.
Remove and the pre-service of three steps of image sharpening through luminance proportion, noise, removed the not enough significant disadvantages of brightness, noise and unique point, eliminated the error of subsequent treatment to greatest extent, also enlarge the scope of sliceable image.
2, the removal of feature extraction and mistake coupling
After pre-processing module, next be exactly carry out feature extraction, characteristic matching, mistake is mated removal, image transformation and image co-registration and is realized splicing.
1) feature extraction
In the image mosaic technology, adopting more characteristic is the angle point characteristic of image, like the Harris angle point.The advantage of Harris angle point is to calculate simply, extracts the some characteristic evenly and rationally, and less to influences such as the rotation of image, viewpoint changes.Because the Harris angle point is the technology of a comparative maturity, be not described in detail at this.
2) characteristic matching
After the angle point feature extraction, mate the angle point that finds correspondence in two width of cloth images through light stream, promptly find pairing point in second width of cloth image, thereby can confirm the matching relationship of each point according to the point in first width of cloth image.
Light stream is the instantaneous velocity of the pixel motion of space motion object on the observation imaging surface.The research of light stream is to utilize the time domain of the pixel intensity data in the image sequence to change and correlativity is confirmed " motion " of location of pixels separately, promptly studies the relation of object structures in gradation of image variation and the scene in time and motion thereof.Because optical flow method is a proven technique comparatively, just be not described in detail at this.
Because there are a lot of similar zones in medical image especially DR image,, need these errors be got rid of so can there be certain error in the light stream coupling.
3) the mistake coupling is removed
This paper will mate removal through two kinds of method synthesis by mistake, thereby can eliminate the influence of not having coupling basically.
Step 1: the relevant matches method, right through the point that obtains a series of couplings after the light stream coupling, verify these to whether related coefficient is the highest, if not the highest, then to think to mate right by mistake.Concrete way is in Fig. 1, to get an angle point; Around point, get the window size of certain limit, calculate the size of the related coefficient of each point among this point and Fig. 1, get with Fig. 1 in the point of maximum related coefficient; Point if not the coupling of light stream before is right, thinks that then the mistake match point is right.
Step 2:Ransac method after process relevant matches method is removed certain error matching, is further removed error matching through Ransac, improves the precision of coupling.The Ransac algorithm is the random sampling consistency algorithm; The principle of this algorithm be from these centering picked at random some the point right, then according to these the point to calculating a kind of Perspective transformation model, draw model parameter; Through model parameter image is carried out conversion; Calculate the least mean-square error of two width of cloth images, repeat this process then always, it is right with those points to obtain that minimum group model parameter of square error at last.
Through after the above-mentioned steps, it is right can to remove error matching point basically, and can draw the basic transformation model parameter of two width of cloth images.
4) image transformation and image co-registration
Obtain after the image transformation model parameter image being carried out conversion, need two width of cloth images be merged after the conversion, if be left intact; Because the overlapping region gray-scale value is different; Thereby can produce apparent in view splicing boundary, so need carry out certain processing, simpler method is to carry out weighting fusion; More far then weight is low more from the distance of overlay region central point for image, thereby realizes the transition of overlay region.Specifically just be not described in detail at this.
The above; Be merely the preferable embodiment of the present invention, but protection scope of the present invention is not limited thereto, any technician who is familiar with the present technique field is in the technical scope that the present invention discloses; The variation that can expect easily or replacement all should be encompassed within protection scope of the present invention.Therefore, protection scope of the present invention should be as the criterion with the protection domain of claim.
Claims (7)
1. medical digital images joining method is characterized in that this method may further comprise the steps:
Step 1: to wanting spliced image to carry out luminance proportion, carrying out noise removal and image sharpening processing;
Step 2: on the basis of step 1, spliced image is carried out feature extraction, characteristic matching, mating removal and image transformation and image co-registration realizes image mosaic by mistake to wanting.
2. a kind of medical digital images joining method according to claim 1, the method that it is characterized in that said luminance proportion is the histogram matching method.
3. a kind of medical digital images joining method according to claim 1; It is characterized in that the method that said noise is removed is the traversal entire image; Getting a window around each pixel adds up; If the value of this pixel is bigger more than 1.5 times than the mean value of the pixel in the window, can think that then this pixel is a noise, its pixel value is replaced with the average in its window.
4. a kind of medical digital images joining method according to claim 3 is characterized in that the computing formula of the average in the said window is:
Wherein:
x
N/2Pixel value average for the window intermediate point;
x
jPixel value for j in window point;
N is a window interior pixel number.
5. a kind of medical digital images joining method according to claim 4 is characterized in that the formula of said image sharpening is:
Wherein, sharpness is the sharpening extent index.
6. a kind of medical digital images joining method according to claim 1, the method that it is characterized in that said characteristic matching is an optical flow method.
7. a kind of medical digital images joining method according to claim 1 is characterized in that the method that said mistake coupling is removed is relevant matches method or Ransac method.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN2012100635443A CN102693533A (en) | 2012-03-12 | 2012-03-12 | Medical digital image mosaicing method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN2012100635443A CN102693533A (en) | 2012-03-12 | 2012-03-12 | Medical digital image mosaicing method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN102693533A true CN102693533A (en) | 2012-09-26 |
Family
ID=46858937
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN2012100635443A Pending CN102693533A (en) | 2012-03-12 | 2012-03-12 | Medical digital image mosaicing method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN102693533A (en) |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103605959A (en) * | 2013-11-15 | 2014-02-26 | 武汉虹识技术有限公司 | A method for removing light spots of iris images and an apparatus |
CN103942615A (en) * | 2014-04-15 | 2014-07-23 | 广东电网公司信息中心 | Noisy point removing method |
CN107316275A (en) * | 2017-06-08 | 2017-11-03 | 宁波永新光学股份有限公司 | A kind of large scale Microscopic Image Mosaicing algorithm of light stream auxiliary |
CN108449580A (en) * | 2018-04-13 | 2018-08-24 | 任阿毛 | Field Monitoring System based on audio frequency component analysis |
CN109523495A (en) * | 2018-10-15 | 2019-03-26 | 北京东软医疗设备有限公司 | Image processing method and device, equipment and storage medium |
CN110619652A (en) * | 2019-08-19 | 2019-12-27 | 浙江大学 | Image registration ghost elimination method based on optical flow mapping repeated area detection |
CN111915587A (en) * | 2020-07-30 | 2020-11-10 | 北京大米科技有限公司 | Video processing method, video processing device, storage medium and electronic equipment |
CN112446262A (en) * | 2019-09-02 | 2021-03-05 | 深圳中兴网信科技有限公司 | Text analysis method, text analysis device, text analysis terminal and computer-readable storage medium |
WO2021103481A1 (en) * | 2019-11-29 | 2021-06-03 | Shanghai United Imaging Healthcare Co., Ltd. | Systems and methods for medical imaging |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7127085B2 (en) * | 2002-08-22 | 2006-10-24 | Korea Advanced Institute Of Science And Technology | Image processing method for automatic image registration and correction |
KR100691855B1 (en) * | 2005-12-29 | 2007-03-12 | 고려대학교 산학협력단 | Apparatus for extracting features from image information |
CN100550054C (en) * | 2007-12-17 | 2009-10-14 | 电子科技大学 | A kind of image solid matching method and device thereof |
CN101667293A (en) * | 2009-09-24 | 2010-03-10 | 哈尔滨工业大学 | Method for conducting high-precision and steady registration on diversified sensor remote sensing images |
-
2012
- 2012-03-12 CN CN2012100635443A patent/CN102693533A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7127085B2 (en) * | 2002-08-22 | 2006-10-24 | Korea Advanced Institute Of Science And Technology | Image processing method for automatic image registration and correction |
KR100691855B1 (en) * | 2005-12-29 | 2007-03-12 | 고려대학교 산학협력단 | Apparatus for extracting features from image information |
CN100550054C (en) * | 2007-12-17 | 2009-10-14 | 电子科技大学 | A kind of image solid matching method and device thereof |
CN101667293A (en) * | 2009-09-24 | 2010-03-10 | 哈尔滨工业大学 | Method for conducting high-precision and steady registration on diversified sensor remote sensing images |
Non-Patent Citations (4)
Title |
---|
刘强: "视频拼接实现过程的研究", 《硕士学位论文全文数据库》, 31 December 2008 (2008-12-31), pages 9 * |
刘贵喜等: "一种稳健的特征点配准算法", 《光学学报》, vol. 28, no. 03, 15 March 2008 (2008-03-15) * |
林学晶: "视频图像拼接技术研究", 《硕士学位论文全文数据库》, 31 December 2010 (2010-12-31), pages 1 - 5 * |
王娟等: "图像拼接技术综述", 《计算机应用研究》, vol. 25, no. 07, 15 July 2008 (2008-07-15), pages 1940 * |
Cited By (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103605959A (en) * | 2013-11-15 | 2014-02-26 | 武汉虹识技术有限公司 | A method for removing light spots of iris images and an apparatus |
CN103942615A (en) * | 2014-04-15 | 2014-07-23 | 广东电网公司信息中心 | Noisy point removing method |
CN103942615B (en) * | 2014-04-15 | 2018-03-27 | 广东电网有限责任公司信息中心 | Noise elimination method |
CN107316275A (en) * | 2017-06-08 | 2017-11-03 | 宁波永新光学股份有限公司 | A kind of large scale Microscopic Image Mosaicing algorithm of light stream auxiliary |
CN108449580A (en) * | 2018-04-13 | 2018-08-24 | 任阿毛 | Field Monitoring System based on audio frequency component analysis |
CN109523495B (en) * | 2018-10-15 | 2022-04-01 | 北京东软医疗设备有限公司 | Image processing method and device, equipment and storage medium |
CN109523495A (en) * | 2018-10-15 | 2019-03-26 | 北京东软医疗设备有限公司 | Image processing method and device, equipment and storage medium |
CN110619652A (en) * | 2019-08-19 | 2019-12-27 | 浙江大学 | Image registration ghost elimination method based on optical flow mapping repeated area detection |
CN110619652B (en) * | 2019-08-19 | 2022-03-18 | 浙江大学 | Image registration ghost elimination method based on optical flow mapping repeated area detection |
CN112446262A (en) * | 2019-09-02 | 2021-03-05 | 深圳中兴网信科技有限公司 | Text analysis method, text analysis device, text analysis terminal and computer-readable storage medium |
WO2021103481A1 (en) * | 2019-11-29 | 2021-06-03 | Shanghai United Imaging Healthcare Co., Ltd. | Systems and methods for medical imaging |
CN111915587A (en) * | 2020-07-30 | 2020-11-10 | 北京大米科技有限公司 | Video processing method, video processing device, storage medium and electronic equipment |
CN111915587B (en) * | 2020-07-30 | 2024-02-02 | 北京大米科技有限公司 | Video processing method, device, storage medium and electronic equipment |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN102693533A (en) | Medical digital image mosaicing method | |
Zhang et al. | DRCDN: learning deep residual convolutional dehazing networks | |
Gao et al. | Gradient-based interpolation method for division-of-focal-plane polarimeters | |
Zhang et al. | Individual urban tree species classification using very high spatial resolution airborne multi-spectral imagery using longitudinal profiles | |
CN105894499A (en) | Binocular-vision-based rapid detection method for three-dimensional information of space object | |
CN101681512B (en) | Vein pattern management system, vein pattern registration device, vein pattern authentication device, vein pattern registration method, and vein pattern authentication method | |
US10873700B2 (en) | Scanning with frame averaging | |
CN109191416A (en) | Image interfusion method based on sparse dictionary study and shearing wave | |
Wang et al. | Simultaneous depth and spectral imaging with a cross-modal stereo system | |
CN109215042A (en) | A kind of photovoltaic battery panel hot spot effect detection system based on computer vision and its calculation method | |
CN104899834A (en) | Blurred image recognition method and apparatus based on SIFT algorithm | |
CN105387814A (en) | Automatic measurement system for surface parameters of prefabricated component of three-dimension braiding composite material | |
CN111583113A (en) | Infrared image super-resolution reconstruction method based on generation countermeasure network | |
CN109376641A (en) | A kind of moving vehicle detection method based on unmanned plane video | |
CN101510295B (en) | Design method for machine vision system based on PCIe and Vision Assistan | |
CN109360150A (en) | A kind of joining method and device of the panorama depth map based on depth camera | |
CN110110793A (en) | Binocular image fast target detection method based on double-current convolutional neural networks | |
Asmare et al. | Image Enhancement by Fusion in Contourlet Transform. | |
CN102682439B (en) | Medical image fusion method based on multidirectional empirical mode decomposition | |
Baharav et al. | In situ height and width estimation of sorghum plants from 2.5 d infrared images | |
KR102331104B1 (en) | Novel image reconstruction system for nuclear medicine through training the neural network for improving the spatial resolution and image quality simultaneously based on structural image of phantoms | |
CN105205829B (en) | Substation's Infrared Image Segmentation based on improved two dimension Otsu | |
CN102521795A (en) | Cross matching fingerprint image scaling method based on global ridge distance | |
CN1570970A (en) | Image merging method based on maximum expectation value and discrete wavelet frame | |
CN103679726A (en) | Method for improving imaging quality of rock debris image |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20120926 |