CN104809713A - CBCT panorama nonlinear sharpening enhancing method based on neighborhood information and Gaussian filter - Google Patents

CBCT panorama nonlinear sharpening enhancing method based on neighborhood information and Gaussian filter Download PDF

Info

Publication number
CN104809713A
CN104809713A CN201510201648.XA CN201510201648A CN104809713A CN 104809713 A CN104809713 A CN 104809713A CN 201510201648 A CN201510201648 A CN 201510201648A CN 104809713 A CN104809713 A CN 104809713A
Authority
CN
China
Prior art keywords
image
cbct
original image
panorama sketch
method based
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
Application number
CN201510201648.XA
Other languages
Chinese (zh)
Other versions
CN104809713B (en
Inventor
王远军
汪葛
万俊
王丽嘉
聂生东
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
University of Shanghai for Science and Technology
Original Assignee
University of Shanghai for Science and Technology
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by University of Shanghai for Science and Technology filed Critical University of Shanghai for Science and Technology
Priority to CN201510201648.XA priority Critical patent/CN104809713B/en
Publication of CN104809713A publication Critical patent/CN104809713A/en
Application granted granted Critical
Publication of CN104809713B publication Critical patent/CN104809713B/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Landscapes

  • Image Processing (AREA)
  • Facsimile Image Signal Circuits (AREA)

Abstract

The invention provides a CBCT nonlinear sharpening enhancing method based on neighborhood information and Gaussian filter. The method includes the following steps: reading a CBCT original image f (x, y), conducting convolution processing on the image to obtain a smooth image, conducting subtraction on the original image and the smoothened image to obtain an unsharp mask image, then looking through the original image f (x, y), calculating a weighting coefficient KW (x, y), finally adding the weighting portion of the unsharp mask image composed of the unsharp mask image and the weighting coefficient KW (x, y) on the original image to form a reinforced image. Compared with the CBCT panorama sharpening enhancing method in the prior art, the weighting coefficient KW (x, y) is arranged, the effect on the CBCT image integral contrast caused by soft tissue images in the original CBCT image and unapparent image boundary tissue enhancing effect caused by noise point amplification in the original image f (x, y) can be effectively restrained, and the method is simple in algorithm and high in computation speed and has good robustness.

Description

Based on the non-linear sharpening enhancement method of CBCT panorama sketch of neighborhood information and gaussian filtering
Technical field
The invention belongs to field of medical image processing, be specifically related to a kind of non-linear sharpening enhancement method of CBCT panorama sketch based on neighborhood information and gaussian filtering.
Background technology
CBCT (Cone Beam Computer Tomography), namely pencil-beam throws illuminated computerized tomography image scanner, is to start to be applied to a kind of novel imaging technique of clinical oral in the beginning of this century.Its principle is that x ray generator is annular DR (digital throwing shine) around throwing according to body with lower quantity of X-rays X (usual tube current is at about 10 milliamperes), then will around throwing according to body repeatedly (180 times-360 times, different and different according to product) numeral throw according to the data obtained in rear " commons factor " that " reconstruction " afterwards and then acquisition 3-D view in a computer.The maximum difference of it and spiral CT is its data for projection is two-dimentional, and be three-dimensional after reconstruction, and the data for projection of spiral CT is one dimension, data for projection is two-dimentional, and obtain three-dimensional data needs the multiple two dimension slicing of continuous sweep.Relative to traditional CT, pencil-beam x-ray utilization factor is high, and roentgen dose X is low, and spatial resolution is high, and cost is low, and floor area is little, scans more flexible.
The major defect of CBCT imaging technique is that density resolution is low, poor to soft tissue structure's imaging effect, and image boundary is organized greatly affected by noise.By can give prominence to the detailed information of image to the sharpening enhancement of image, obtain image boundary more clearly, facilitate doctor's delineating target area.
The enhancing algorithm of CBCT mainly comprises two kinds: the method based on spatial domain and the method based on frequency domain.It is all based on local pixel information that spatial domain method calculates at every turn, can not better embody integral image characteristic, have impact on the raising of CBCT integral image contrast to a certain extent, and the impact for the CBCT image of low contrast is comparatively serious.Although frequency domain method is to the effect of the sharpening enhancement of CBCT panorama sketch in overall contrast, comparatively spatial domain method is good, and calculate comparatively complicated, computing velocity is slow, has the ring effect of trace, cannot meet the CBCT image request of high resolving power, low contrast.
Summary of the invention
The present invention is for solving the problem and carrying out, spatial resolution for CBCT image is high, density resolution is low, the features such as low-density imaging of tissue is unintelligible, provide a kind of non-linear sharpening enhancement method of CBCT panorama sketch based on neighborhood information and gaussian filtering, utilize the spatial domain method of the neighborhood information of pixel to classics to improve.
Present invention employs following technical scheme:
CBCT panorama sketch sharpening enhancement method based on neighborhood information and gaussian filtering provided by the invention, has such feature, comprises the following steps:
Step one, reads width CBCT original image f (x, y);
Step 2, employing radius is R, and standard deviation is that the gauss low frequency filter of σ carries out process of convolution to CBCT original image, obtains the image f smoothly c(x, y);
Step 3, the image fc (x, y) after to be deducted by the original image f (x, y) in step one in step 2 level and smooth, obtains unsharp masking image f mask(x, y)=f (x, y)-ffc c((x, y);
Step 4, traversal original image f (x, y), with each pixel X kcentered by, calculate average value mu (x, y) and the meansquaredeviationσ (x, y) of pixel in its m*n neighborhood, and by formula α ( x , y ) = e | X k - u ( x , y ) | - σ ( x , y ) , KW ( x , y ) = α ( x , y ) , α ( x , y ) ≤ 1 1 1 + e α ( x , y ) , α ( x , y ) > 1 , Calculate weights COEFFICIENT K W (x, y),
Wherein, a (x, y) is noise intensity coefficient.
Step 5, CBCT original image adds described unsharp masking image f maskweight portion g (x, y)=f (x, the y)+K*f of (x, y) mask(x, y) * KW (x, y), the image after being enhanced.
CBCT panorama sketch sharpening enhancement method based on neighborhood information and gaussian filtering provided by the invention, it is characterized in that: in step 2, the value of radius R is preferably 3, and the value of standard deviation is preferably 20.
CBCT panorama sketch sharpening enhancement method based on neighborhood information and gaussian filtering provided by the invention, it is characterized in that: in step 2, gauss low frequency filter formula is wherein, H lP(u, v) is gauss low frequency filter function; D (u, v) is for each point (u, v) in picture frequency territory is to the distance of filter center, and σ is standard deviation.
CBCT panorama sketch sharpening enhancement method based on neighborhood information and gaussian filtering provided by the invention, it is characterized in that: in step 2, Convolution Formula is:
w ( x , y ) * f ( x , y ) = Σ s = - a a Σ t = - b b w ( s , t ) f ( x - s , y - t ) ,
Wherein, w (x, y) is gauss low frequency filter function, and f (x, y) is original image.
CBCT panorama sketch sharpening enhancement method based on neighborhood information and gaussian filtering provided by the invention, is characterized in that: the preferred 3*3 neighborhood of m*n neighborhood.
CBCT panorama sketch sharpening enhancement method based on neighborhood information and gaussian filtering provided by the invention, is characterized in that: the value of k described in step 5 scope is 1 ~ 4.
Invention effect and effect
The invention provides a kind of non-linear sharpening enhancement method of CBCT panorama sketch based on neighborhood information and gaussian filtering, comprise the following steps: read a width CBCT original image f (x, y), the image is smoothly obtained after process of convolution is carried out to it, original image and level and smooth after image do difference and obtain unsharp mask image, travel through original image f (x afterwards, y), calculate weights COEFFICIENT K W (x, y), finally on original image, add unsharp mask image and weights COEFFICIENT K W (x, the weight portion of the unsharp mask image y) formed, image after being enhanced.Compare with CBCT panorama sketch sharpening enhancement method of the prior art, weights COEFFICIENT K W (x is provided with in CBCT panorama sketch sharpening enhancement method in the present invention, y), effectively can suppress original image f (x, y) in, noise spot is exaggerated and soft-tissue image in the CBCT original image that causes and image boundary tissue enhancing DeGrain, and the phenomenon affecting CBCT integral image contrast occurs, simultaneously, the algorithm of the method is simple, fast operation, and there is good robustness.
Accompanying drawing explanation
Fig. 1 is the process flow diagram of the non-linear sharpening enhancement method of CBCT panorama sketch of the present invention;
Fig. 2 is the primitive oral cavity CBCT panorama sketch before enhancing of the present invention;
Fig. 3 is the oral cavity CBCT panorama sketch after directly utilizing unsharp masking operator to strengthen in the present invention;
Fig. 4 is the CBCT panorama sketch after utilizing the CBCT panorama sketch sharpening enhancement method in the present invention to strengthen.
Embodiment
Below in conjunction with accompanying drawing, the specific embodiment of the present invention is described.
Fig. 1 is the process flow diagram of the non-linear sharpening enhancement method of CBCT panorama sketch in the present embodiment.
As shown in Figure 1, the non-linear sharpening enhancement method of the CBCT panorama sketch in the present embodiment comprises the steps:
Step one (S 1), read width CBCT panorama original image f (x, y), read CBCT panorama original image f (x, y) of input, and be converted into double-length floating image;
Step 2 (S 2), employing radius is R, and standard deviation is that the gauss low frequency filter of σ carries out process of convolution to image, obtains the image f smoothly c(x, y).
The value of radius R is preferably 3, and the value of standard deviation is preferably 20.
Gauss low frequency filter formula is: wherein, H lP(u, v) is gauss low frequency filter function; D (u, v) is for each point (u, v) in picture frequency territory is to the distance of filter center, and σ is standard deviation.
Convolution Formula is: w ( x , y ) * f ( x , y ) = Σ s = - a a Σ t = - b b w ( s , t ) f ( x - s , y - t ) , Wherein, w (x, y) is gauss low frequency filter function, and f (x, y) is CBCT panorama original image;
Step 3 (S 3), traversal original image f (x, y), with each pixel X kcentered by, calculate average value mu (x, y) and the meansquaredeviationσ (x, y) of pixel in its 3*3 neighborhood, by formula α ( x , y ) = e | X k - u ( x , y ) | - σ ( x , y ) , KW ( x , y ) = α ( x , y ) , α ( x , y ) ≤ 1 1 1 + e α ( x , y ) , α ( x , y ) > 1 , Calculate weights COEFFICIENT K W (x, y).
The pixel of CBCT panoramic picture is mainly divided three classes: marginal point, image internal point and noise spot.If formula middle finger is several | X k-u (x, y) |-σ (x, y) > 0, then in image, this point is noise spot, otherwise is then edge and the internal point of image.Formula the scope of value is (0,2), and therefore, α (x, y) > 1 represents that this point is noise spot, and α (x, y)≤1 item represents that this point is non-noise point.
The object arranging weight coefficient is to prevent noise spot in original image f (x, y) to be exaggerated, therefore directly can not using α k(x, y) is as weight coefficient item.α kin (x, y) value larger key diagram picture, the noise intensity of this point is larger, in order to the weight coefficient value of restraint speckle point, utilizes formula KW ( x , y ) = α ( x , y ) , α ( x , y ) ≤ 1 1 1 + e α ( x , y ) , α ( x , y ) > 1 In the weights coefficient of noise spot is carried out inverse proportion function, made it much smaller than weights coefficient corresponding to non-noise point.
Step 4 (S 4): original image f (x, y) is deducted the image f smoothly c(x, y), produces unsharp masking image f mask(x, y)=f (x, y)-fcf c((xx, yy)).
Step 5 (S 5): on original image, add unsharp masking image f maskweight portion g (x, y)=f (x, the y)+K*f of (x, y) mask(x, y) * KW (x, y), the image after being enhanced.
Fig. 2 is the primitive oral cavity CBCT panorama sketch before the enhancing of the present embodiment.
Fig. 3 is the oral cavity CBCT panorama sketch after directly utilizing unsharp masking operator to strengthen in the present embodiment.
Fig. 4 is the oral cavity CBCT panorama sketch after utilizing the CBCT panorama sketch sharpening enhancement method in the present embodiment to strengthen.
As shown in Figures 2 to 4, the contrast of the oral cavity CBCT panorama sketch after directly utilizing unsharp masking operator to strengthen is better than the primitive oral cavity CBCT panorama sketch before strengthening.The image utilizing the method for the present embodiment to strengthen all is better than directly utilizing unsharp masking to strengthen the effect of image method in the enhancing effect of soft tissue and the contrast of whole image.
Embodiment effect and effect
The invention provides a kind of non-linear sharpening enhancement method of CBCT panorama sketch based on neighborhood information and gaussian filtering, comprise the following steps: read a width CBCT original image f (x, y), the image is smoothly obtained after process of convolution is carried out to it, original image and level and smooth after image do difference and obtain unsharp mask image, travel through original image f (x afterwards, y), calculate weights COEFFICIENT K W (x, y), finally on original image, add unsharp mask image and weights COEFFICIENT K W (x, the weight portion of the unsharp mask image y) formed, image after being enhanced.Compare with CBCT panorama sketch sharpening enhancement method of the prior art, weights COEFFICIENT K W (x is provided with in CBCT panorama sketch sharpening enhancement method in the present invention, y), effectively can suppress original image f (x, y) in, noise spot is exaggerated and soft-tissue image in the CBCT original image that causes and image boundary tissue enhancing DeGrain, and the phenomenon affecting CBCT integral image contrast occurs, simultaneously, the algorithm of the method is simple, fast operation, and there is good robustness.

Claims (6)

1., based on a CBCT panorama sketch sharpening enhancement method for neighborhood information and gaussian filtering, for improving the overall contrast of described CBCT image, it is characterized in that, comprise the following steps:
Step one, reads width CBCT original image f (x, y);
Step 2, employing radius is R, and standard deviation is that the gauss low frequency filter of σ carries out process of convolution to described CBCT original image, obtains the image f smoothly c(x, y);
Step 3, travels through described original image f (x, y), with each pixel X kcentered by, calculate average value mu (x, y) and the meansquaredeviationσ (x, y) of pixel in its m*n neighborhood, and by formula α ( x , y ) = e | X k - u ( x , y ) | - σ ( x , y ) , KW ( x , y ) = α ( x , y ) , α ( x , y ) ≤ 1 1 1 + e α ( x , y ) , α ( x , y ) > 1 , Calculate weights COEFFICIENT K W (x, y),
Wherein, described a (x, y) is noise intensity coefficient.
Step 4, the image f after to be deducted by the original image f (x, y) in described step one in described step 2 level and smooth c(x, y), obtains unsharp masking image f mask(x, y)=f (x, y)-ffc c((xx, yy));
Step 5, described CBCT original image adds described unsharp masking image f maskweight portion g (x, y)=f (x, the y)+K*f of (x, y) mask(x, y) * KW (x, y), the image after being enhanced.
2. the CBCT panorama sketch sharpening enhancement method based on neighborhood information and gaussian filtering according to claim 1, is characterized in that:
Wherein, in step 2, the value of described radius R is preferably 3, and the value of described standard deviation is preferably 20.
3. the CBCT panorama sketch sharpening enhancement method based on neighborhood information and gaussian filtering according to claim 1, is characterized in that:
Wherein, in step 2, described gauss low frequency filter formula is wherein, H lP(u, v) is gauss low frequency filter function; D (u, v) is for each point (u, v) in picture frequency territory is to the distance of filter center, and σ is standard deviation.
4. the CBCT panorama sketch sharpening enhancement method based on neighborhood information and gaussian filtering according to claim 1, is characterized in that:
Wherein, in step 2, described Convolution Formula is:
w ( x , y ) * f ( x , y ) = Σ s = - a a Σ t = - b b w ( s , t ) f ( x - s , y - t ) ,
Wherein, w (x, y) is Gaussian low pass wave function, and f (x, y) is original image.
5. the CBCT panorama sketch sharpening enhancement method based on neighborhood information and gaussian filtering according to claim 1, is characterized in that:
Wherein, the preferred 3*3 neighborhood of described m*n neighborhood.
6. the CBCT panorama sketch sharpening enhancement method based on neighborhood information and gaussian filtering according to claim 1, is characterized in that:
Wherein, the value of k described in step 5 scope is 1 ~ 4.
CN201510201648.XA 2015-04-24 2015-04-24 The non-linear sharpening enhancement method of CBCT panorama sketch based on neighborhood information and gaussian filtering Expired - Fee Related CN104809713B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201510201648.XA CN104809713B (en) 2015-04-24 2015-04-24 The non-linear sharpening enhancement method of CBCT panorama sketch based on neighborhood information and gaussian filtering

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201510201648.XA CN104809713B (en) 2015-04-24 2015-04-24 The non-linear sharpening enhancement method of CBCT panorama sketch based on neighborhood information and gaussian filtering

Publications (2)

Publication Number Publication Date
CN104809713A true CN104809713A (en) 2015-07-29
CN104809713B CN104809713B (en) 2017-09-12

Family

ID=53694518

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201510201648.XA Expired - Fee Related CN104809713B (en) 2015-04-24 2015-04-24 The non-linear sharpening enhancement method of CBCT panorama sketch based on neighborhood information and gaussian filtering

Country Status (1)

Country Link
CN (1) CN104809713B (en)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105225209A (en) * 2015-10-29 2016-01-06 Tcl集团股份有限公司 A kind of sharpening implementation method of non-homogeneous interpolation image and system
CN105894444A (en) * 2016-03-31 2016-08-24 深圳市菲森科技有限公司 Method and device for generating dental panoramic image on the basis of CBCT image
CN106911904A (en) * 2015-12-17 2017-06-30 通用电气公司 Image processing method, image processing system and imaging system
CN108024103A (en) * 2017-12-01 2018-05-11 重庆贝奥新视野医疗设备有限公司 Image sharpening method and device
CN115409833A (en) * 2022-10-28 2022-11-29 一道新能源科技(衢州)有限公司 Hot spot defect detection method of photovoltaic panel based on unsharp mask algorithm

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20090087121A1 (en) * 2007-10-01 2009-04-02 Samsung Electronics Co., Ltd. Method and apparatus for enhancing detail based on noise elimination, and method and apparatus for calculating noise weight
CN103391392A (en) * 2012-05-11 2013-11-13 索尼公司 Image enhancement apparatus and method
CN103473748A (en) * 2013-09-23 2013-12-25 中国矿业大学(北京) Method for enhancing underground coal mine image
CN103530854A (en) * 2013-10-15 2014-01-22 南通市牧井微电科技发展有限公司 Spatial-domain filtering-based CBCT (Cone Beam Computed Tomography) panoramic image sharpening method
JP5537164B2 (en) * 2010-01-08 2014-07-02 株式会社日立メディコ Image reconstruction apparatus and X-ray CT apparatus

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20090087121A1 (en) * 2007-10-01 2009-04-02 Samsung Electronics Co., Ltd. Method and apparatus for enhancing detail based on noise elimination, and method and apparatus for calculating noise weight
JP5537164B2 (en) * 2010-01-08 2014-07-02 株式会社日立メディコ Image reconstruction apparatus and X-ray CT apparatus
CN103391392A (en) * 2012-05-11 2013-11-13 索尼公司 Image enhancement apparatus and method
CN103473748A (en) * 2013-09-23 2013-12-25 中国矿业大学(北京) Method for enhancing underground coal mine image
CN103530854A (en) * 2013-10-15 2014-01-22 南通市牧井微电科技发展有限公司 Spatial-domain filtering-based CBCT (Cone Beam Computed Tomography) panoramic image sharpening method

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105225209A (en) * 2015-10-29 2016-01-06 Tcl集团股份有限公司 A kind of sharpening implementation method of non-homogeneous interpolation image and system
CN105225209B (en) * 2015-10-29 2018-11-30 Tcl集团股份有限公司 A kind of sharpening realization method and system of non-homogeneous interpolation image
CN106911904A (en) * 2015-12-17 2017-06-30 通用电气公司 Image processing method, image processing system and imaging system
CN106911904B (en) * 2015-12-17 2020-04-21 通用电气公司 Image processing method, image processing system and imaging system
CN105894444A (en) * 2016-03-31 2016-08-24 深圳市菲森科技有限公司 Method and device for generating dental panoramic image on the basis of CBCT image
CN105894444B (en) * 2016-03-31 2018-12-21 深圳市菲森科技有限公司 A kind of method and device based on CBCT video generation panoramic dental image
CN108024103A (en) * 2017-12-01 2018-05-11 重庆贝奥新视野医疗设备有限公司 Image sharpening method and device
CN115409833A (en) * 2022-10-28 2022-11-29 一道新能源科技(衢州)有限公司 Hot spot defect detection method of photovoltaic panel based on unsharp mask algorithm
CN115409833B (en) * 2022-10-28 2023-01-31 一道新能源科技(衢州)有限公司 Hot spot defect detection method of photovoltaic panel based on unsharp mask algorithm

Also Published As

Publication number Publication date
CN104809713B (en) 2017-09-12

Similar Documents

Publication Publication Date Title
Nomura et al. Projection‐domain scatter correction for cone beam computed tomography using a residual convolutional neural network
Kim et al. A performance comparison of convolutional neural network‐based image denoising methods: The effect of loss functions on low‐dose CT images
CN104809713A (en) CBCT panorama nonlinear sharpening enhancing method based on neighborhood information and Gaussian filter
CN104318536B (en) The bearing calibration of CT images and device
CN104574416B (en) A kind of low dosage power spectrum CT image de-noising methods
CN105913398A (en) Method and device for processing CT reconstructed image
JP2017196404A (en) Image processing apparatus, x-ray ct apparatus, and image processing method
CN104408758A (en) Low-dose processing method of energy spectrum CT image
WO2021169695A1 (en) Ct image generation method for attenuation correction of pet images
CN104751429B (en) A kind of low dosage power spectrum CT image processing methods based on dictionary learning
CN106725565A (en) A kind of cone-beam XCT imaging quality assessment methods under sparse projection
CN102024251A (en) System and method for multi-image based virtual non-contrast image enhancement for dual source CT
CN103679801A (en) Angiocarpy three-dimensional reconstruction method based on multi-view X-ray film
CN103961128A (en) Variable-focus cone-beam computed tomography (CT) imaging device
Li et al. Multienergy cone-beam computed tomography reconstruction with a spatial spectral nonlocal means algorithm
Ma et al. An encoder-decoder network for direct image reconstruction on sinograms of a long axial field of view PET
Li et al. Incorporation of residual attention modules into two neural networks for low‐dose CT denoising
Yuan et al. Head and neck synthetic CT generated from ultra‐low‐dose cone‐beam CT following Image Gently Protocol using deep neural network
CN103793890A (en) Method for recovering and processing energy spectrum CT images
Funama et al. Conditional generative adversarial networks to generate pseudo low monoenergetic CT image from a single-tube voltage CT scanner
CN114511497A (en) Imaging method and device applied to cone beam CT sparse sampling
JP2019536538A (en) Bone and hard plaque segmentation in spectral CT
van Gogh et al. INSIDEnet: Interpretable nonexpansive data‐efficient network for denoising in grating interferometry breast CT
Sechopoulos et al. Power spectrum analysis of the x‐ray scatter signal in mammography and breast tomosynthesis projections
TWI716762B (en) A method of using neural network to remove tooth image noise

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
EXSB Decision made by sipo to initiate substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20170912

Termination date: 20200424

CF01 Termination of patent right due to non-payment of annual fee