Nothing Special   »   [go: up one dir, main page]

CN105741241B - Tumor region image enchancing method and system based on synthesis enhancing image - Google Patents

Tumor region image enchancing method and system based on synthesis enhancing image Download PDF

Info

Publication number
CN105741241B
CN105741241B CN201610054646.7A CN201610054646A CN105741241B CN 105741241 B CN105741241 B CN 105741241B CN 201610054646 A CN201610054646 A CN 201610054646A CN 105741241 B CN105741241 B CN 105741241B
Authority
CN
China
Prior art keywords
image
tumor
roi
covering
window
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.)
Expired - Fee Related
Application number
CN201610054646.7A
Other languages
Chinese (zh)
Other versions
CN105741241A (en
Inventor
谌先敢
刘海华
高智勇
陆雪松
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.)
South Central Minzu University
Original Assignee
South Central University for Nationalities
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 South Central University for Nationalities filed Critical South Central University for Nationalities
Priority to CN201610054646.7A priority Critical patent/CN105741241B/en
Publication of CN105741241A publication Critical patent/CN105741241A/en
Application granted granted Critical
Publication of CN105741241B publication Critical patent/CN105741241B/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/70—Denoising; Smoothing
    • 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
    • G06T2207/20—Special algorithmic details
    • G06T2207/20092—Interactive image processing based on input by user
    • G06T2207/20104—Interactive definition of region of interest [ROI]
    • 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
    • G06T2207/30096—Tumor; Lesion

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Image Processing (AREA)
  • Apparatus For Radiation Diagnosis (AREA)

Abstract

The invention discloses a kind of tumor region image enchancing methods and system based on synthesis enhancing image, are related to field of medical image processing.Present invention firstly provides " synthesis enhancing image " this concepts, in original CT or MR images, noise reduction process and enhancing processing are carried out respectively to the ROI region of the whole tumor regions of covering, obtain the noise-reduced image and enhancing image of ROI region, fusion is weighted to noise-reduced image and enhancing image, obtains synthesis enhancing image.The present invention can enhance the ROI region in original CT or MR images, synthesize weighted factor different in enhancing image by selecting, so that the surface of tumour and smeared out boundary in original CT or MR images is become apparent from, doctor is facilitated to observe the surface and boundary of tumour；Threshold segmentation method, energy Accurate Segmentation is used to go out tumor region on synthesis enhancing image.

Description

Tumor region image enhancement method and system based on synthesis enhanced image
Technical Field
The invention relates to the field of medical image processing, in particular to a tumor region image enhancement method and system based on a synthesized enhanced image.
Background
In the process of actually diagnosing tumor-related diseases, doctors can help to clinically confirm the disease condition by measuring and analyzing the boundary, cross-sectional area and volume of the tumor in medical images such as CT (Computed Tomography) or MR (Magnetic Resonance), and the like, and at this time, the boundary of the tumor in the medical images such as CT or MR needs to be segmented. The accurate segmentation of the tumor boundary is very important for treatment planning, and at present, the part of work mainly depends on manual delineation and has low accuracy. Due to noise and ambiguity in medical images such as CT or MR, the boundaries of many tumors are blurred, and it is difficult to accurately segment the boundaries of tumors by using conventional image segmentation methods, such as threshold segmentation on the original medical images such as CT or MR.
Disclosure of Invention
The invention aims to overcome the defects of the background technology, and provides a tumor region image enhancement method and a tumor region image enhancement system based on a synthesized enhanced image, which can enhance the ROI (region of interest) covering all tumor regions in an original CT (computed tomography) or MR (magnetic resonance) image, and can ensure that the surface and fuzzy boundary of a tumor in the original CT or MR image become clear and visible by selecting different weighting factors in the synthesized enhanced image, thereby facilitating a doctor to observe the surface and the boundary of the tumor; the threshold segmentation method is adopted on the synthesized enhanced image, so that the tumor region can be accurately segmented, and the accuracy is obviously higher than that of the threshold segmentation directly carried out on the original CT or MR image.
The invention provides a tumor region image enhancement method based on a synthesis enhanced image, which comprises the following steps:
A. selecting an elliptical ROI (region of interest) covering all tumor regions from a frame of original CT (computed tomography) image or MR (magnetic resonance) image containing tumors; setting all pixels outside the ROI to be zero, and adjusting the window width and window level of the pixel value inside the ROI to meet the requirement of human observation;
B. performing noise reduction treatment on the ROI in the image after window width and window level adjustment by adopting an anisotropic diffusion method to obtain a noise reduction image of the ROI covering all tumor regions;
C. enhancing the ROI area in the image after window width and window level adjustment by adopting a multi-scale enhancement method to obtain an enhanced image of the ROI area covering all tumor areas;
D. and D, performing weighted fusion on the noise-reduced image of the ROI area covering all the tumor areas obtained in the step B and the enhanced image of the ROI area covering all the tumor areas obtained in the step C to obtain a synthesized enhanced image of the ROI area covering all the tumor areas.
On the basis of the above technical solution, the process of adjusting the window width and window level of the pixel value inside the ROI in step a is as follows: centering on the centroid of the ellipse, a rectangular region is selected, the length r of whichlengthLess than the major axis of the ellipse, the width r of the rectanglewidthA minor axis smaller than the ellipse; selecting the maximum gray value in the rectangle as the maximum value C of the window width in the window technologymaxSelecting the minimum gray value in the rectangle as the window techniqueMinimum value of intraoperative window width Cmin(ii) a Then arbitrarily selecting one of the following three formulas to map the pixel values within the ellipse:
wherein, IoriFor the original CT or MR image, α is the coefficient controlling the overall brightness of the image, γ1For adjustable coefficients, by varying gamma1I is the image after window width and level adjustment.
On the basis of the technical scheme, in the step A, gamma is1When less than 1, equation (3) converts a narrow range of input values to a wider range of output values.
On the basis of the technical scheme, the anisotropic diffusion method in the step B comprises the following steps: removing noise by using an anisotropic diffusion model, wherein the formula of the anisotropic diffusion model is as follows:
where div () is the divergence operator, c (x, y, t) is the diffusion coefficient, which is chosen as a function of the image gradient, t is time,represents the gradient operator and the gradient operator, respectively,it is the gradient of the diffusion coefficient c,is the gradient of image I, Δ represents the laplacian, Δ I is the laplacian of image I;
solving the partial differential equation expressed by the formula (4) whose solution is expressed as Irn，IrnAnd B, obtaining a noise reduction image of the ROI covering the whole tumor area in the step B.
On the basis of the above technical solution, in step B, the function of the diffusion coefficient c (x, y, t) in the formula of the anisotropic diffusion model has two expression modes:
wherein,k is a constant, the two diffusion coefficient functions take the mode of the gradient of the image I as the basis of diffusion speed, and the diffusion coefficient is small at the position with large gradient, so that the purpose of protecting the edge is achieved.
On the basis of the technical scheme, the multi-scale enhancement method in the step C comprises the following steps:
enhancement processing is carried out on ROI areas covering the whole tumor area under different scales:
first, an ROI region covering the entire tumor region is downsampled by the formula (7), and a downsampled image I is obtainedre；
Ire＝REDUCE(I) (7)
Wherein, REDUCE () represents a down-sampling operation;
then, the downsampled image I is processed by the formula (8)reCarrying out zooming enhancement processing;
Iu＝(Ire-Ire*G)γ2(8)
wherein, IuIs a down-sampled enhanced image of reduced size, IreIs a down-sampled image, G is a Gaussian kernel, Ire-IreG represents the high frequency component of the image, γ2To control the coefficients of differentiation between tumor and non-tumor regions, IuSize of and down-sampled image IreAre the same and are all smaller than the image I after window width and window level adjustment;
then, the down-sampled enhanced image I with reduced size obtained by the formula (8) is processed by the formula (9)uPerforming upsampling to obtain Iex：Iex＝EXPAND(Iu) (9)
Wherein, IexFor the enhanced image of the ROI area covering the entire tumor area obtained in step C, IexIs the same as the size of the image I after window width and window level adjustment; EXPAND () represents an upsample operation.
On the basis of the technical scheme, the EXPAND () in the step C is obtained by an extended interpolation operator, if the size of the EXPAND () needs to be extended by one time, the pixel which is doubled in the horizontal direction and the vertical direction is enhanced by one time, namely, a value needs to be inserted between any two pixels of each line, and a line needs to be inserted between each two lines; the interpolation operator adopts a double cubic interpolation method.
On the basis of the technical scheme, a pixel-level weighted average fusion method is adopted in the step D for weighted fusion, and the formula is as follows:
Ien＝w·Irn+(1-w)·Iex(10)
wherein, IrnFor the noise-reduced image of the ROI area covering the entire tumor area obtained in step B, IexFor the enhanced image of the ROI area covering the whole tumor area obtained in step C, w is a weighting factor, IenA composite enhanced image of the ROI region covering the entire tumor region; by selecting different weighting factors w in equation (10), the physician can easily observe the surface and boundary of the tumor.
The invention also provides a tumor region image enhancement system based on the synthesized enhanced image, which comprises an ROI region selection unit, a window width and window level adjustment unit, a noise reduction unit, an enhancement unit and a weighted fusion unit, wherein:
the ROI area selection unit is used for: selecting an elliptical ROI (region of interest) covering all tumor regions from a frame of original CT (computed tomography) image or MR (magnetic resonance) image containing tumors;
the window width and window level adjusting unit is used for: setting all pixels outside the ROI to be zero, and adjusting the window width and window level of the pixel value inside the ROI to meet the requirement of human observation;
the noise reduction unit is used for: performing noise reduction treatment on the ROI in the image after window width and window level adjustment by adopting an anisotropic diffusion method to obtain a noise reduction image of the ROI covering all tumor regions;
the enhancement unit is configured to: enhancing the ROI area in the image after window width and window level adjustment by adopting a multi-scale enhancement method to obtain an enhanced image of the ROI area covering all tumor areas;
the weighted fusion unit is configured to: and performing weighted fusion on the noise-reduced image of the ROI covering all the tumor regions and the enhanced image of the ROI covering all the tumor regions to obtain a synthesized enhanced image of the ROI covering all the tumor regions.
On the basis of the technical scheme, the window width and window level adjusting unit carries out adjustment on pixel values in the ROI areaThe process of adjusting the window width and the window level comprises the following steps: centering on the centroid of the ellipse, a rectangular region is selected, the length r of whichlengthLess than the major axis of the ellipse, the width r of the rectanglewidthA minor axis smaller than the ellipse; selecting the maximum gray value in the rectangle as the maximum value C of the window width in the window technologymaxSelecting the minimum gray value in the rectangle as the minimum value C of the window width in the window technologymin(ii) a Arbitrarily selecting one of the following three formulas to map the pixel values within the ellipse:
wherein, IoriFor the original CT or MR image, α is the coefficient controlling the overall brightness of the image, γ1For adjustable coefficients, by varying gamma1I is the image after window width and level adjustment.
On the basis of the technical scheme, gamma is1When less than 1, equation (3) converts a narrow range of input values to a wider range of output values.
On the basis of the technical scheme, the anisotropic diffusion method adopted by the noise reduction unit is as follows: removing noise by using an anisotropic diffusion model, wherein the formula of the anisotropic diffusion model is as follows:
wherein,div () is the divergence operator, c (x, y, t) is the diffusion coefficient, which is chosen as a function of the image gradient, t is time,represents the gradient operator and the gradient operator, respectively,it is the gradient of the diffusion coefficient c,is the gradient of image I, Δ represents the laplacian, Δ I is the laplacian of image I;
solving the partial differential equation expressed by the formula (4) whose solution is expressed as Irn，IrnA noise-reduced image of the ROI area covering the entire tumor area.
On the basis of the technical scheme, the function of the diffusion coefficient c (x, y, t) in the formula of the anisotropic diffusion model has two expression modes:
wherein,k is a constant, the two diffusion coefficient functions take the mode of the gradient of the image I as the basis of diffusion speed, and the diffusion coefficient is small at the position with large gradient, so that the purpose of protecting the edge is achieved.
On the basis of the technical scheme, the multi-scale enhancement method adopted by the enhancement unit comprises the following steps:
enhancement processing is carried out on ROI areas covering the whole tumor area under different scales:
first, an ROI region covering the entire tumor region is downsampled by the formula (7), and a downsampled image I is obtainedre；
Ire＝REDUCE(I) (7)
Wherein, REDUCE () represents a down-sampling operation;
then, the downsampled image I is processed by the formula (8)reCarrying out zooming enhancement processing;
Iu＝(Ire-Ire*G)γ2(8)
wherein, IuIs a down-sampled enhanced image of reduced size, IreIs a down-sampled image, G is a Gaussian kernel, Ire-IreG represents the high frequency component of the image, γ2To control the coefficients of differentiation between tumor and non-tumor regions, IuSize of and down-sampled image IreAre the same and are all smaller than the image I after window width and window level adjustment;
then, the down-sampled enhanced image I with reduced size obtained by the formula (8) is processed by the formula (9)uPerforming upsampling to obtain Iex：Iex＝EXPAND(Iu) (9)
Wherein, IexFor the enhanced image of the ROI area covering the entire tumor area obtained in step C, IexIs the same as the size of the image I after window width and window level adjustment; EXPAND () represents an upsample operation.
On the basis of the technical scheme, the EXPAND () is obtained by an extended interpolation operator, if the size of the EXPAND () is extended by one time, the pixel of one time is enhanced in the horizontal direction and the vertical direction, namely a value needs to be inserted between any two pixels of each line, and a line needs to be inserted between every two lines; the interpolation operator adopts a double cubic interpolation method.
On the basis of the technical scheme, the weighted fusion unit performs weighted fusion by adopting a pixel-level weighted average fusion method, and the formula is as follows:
Ien＝w·Irn+(1-w)·Iex(10)
wherein, IrnTo de-noise images of the ROI area covering the entire tumor area, IexFor an enhanced image of the ROI area covering the whole tumor area, w is a weighting factor, IenA composite enhanced image of the ROI region covering the entire tumor region; by selecting different weighting factors w in equation (10), the physician can easily observe the surface and boundary of the tumor.
Compared with the prior art, the invention has the following advantages:
the invention firstly proposes the concept of 'synthesizing enhanced images', in original CT or MR images, the ROI covering all tumor regions are respectively subjected to noise reduction processing and enhancement processing to obtain noise reduction images and enhanced images of the ROI, and the noise reduction images and the enhanced images are subjected to weighted fusion to obtain synthesized enhanced images. The boundaries of the tumor in the composite enhanced image are clearer compared to the original CT or MR image, while the information on the tumor surface is not much lost. The method can enhance the ROI area covering all tumor areas in the original CT or MR image, and the surface and fuzzy boundary of the tumor in the original CT or MR image become clear and visible by selecting different weighting factors in the synthesized enhanced image, thereby facilitating the doctor to observe the surface and boundary of the tumor; the threshold segmentation method is adopted on the synthesized enhanced image, so that the tumor region can be accurately segmented, and the accuracy is obviously higher than that of the threshold segmentation directly carried out on the original CT or MR image.
Drawings
Fig. 1 is a flowchart of a method for enhancing an image of a tumor region based on a synthetically enhanced image according to an embodiment of the present invention.
Fig. 2 is an image after window width and level adjustment in an embodiment of the present invention.
Fig. 3 is a region of interest in an image after window width level adjustment in an embodiment of the invention.
FIG. 4 is a noise-reduced image of the ROI area in an embodiment of the present invention.
FIG. 5 is an enhanced image of the ROI area in an embodiment of the present invention.
FIG. 6 is a composite enhanced image of the ROI area in an embodiment of the present invention.
Detailed Description
The invention is described in further detail below with reference to the figures and the embodiments.
Aiming at the problem that the boundary of a tumor in an original CT or MR image is fuzzy and difficult to segment, the embodiment of the invention provides a tumor region image enhancement method based on a synthesized enhanced image, which is shown in figure 1 and comprises the following steps:
A. selecting an elliptical ROI (Region Of Interest) covering all tumor regions from a frame Of original CT image or MR image containing tumors; setting all pixels outside the ROI area to be zero, and adjusting the window width and window level of the pixel value inside the ROI area to meet the requirement of human observation.
In practical applications, the ROI region is selected first, and then the window width level is adjusted, but since the pixel value of the original medical image is between-1000 and +1000HU (Hounsfield Unit, hansen feld Unit) and cannot be displayed on a normal display, only the image after the window width level is adjusted is displayed in the embodiment of the present invention. The image after window width and window level adjustment is shown in FIG. 2; the ROI region in the window-width window-level adjusted image is shown in fig. 3.
B. Performing noise reduction processing on the ROI region in the image after window width window level adjustment by using an Anisotropic diffusion method (Anisotropic diffusion), to obtain a noise-reduced image covering the ROI region of all tumor regions, as shown in fig. 4;
C. performing enhancement processing on the ROI area in the image after window width and window level adjustment by adopting a multi-scale enhancement method to obtain an enhanced image covering the ROI area of all tumor areas, which is shown in FIG. 5;
D. and (3) performing weighted fusion on the noise-reduced image of the ROI area covering the whole tumor area obtained in the step (B) and the enhanced image of the ROI area covering the whole tumor area obtained in the step (C) to obtain a synthesized enhanced image of the ROI area covering the whole tumor area, which is shown in fig. 6.
The boundaries of the tumor in the composite enhanced image are clearer compared to the original CT or MR image, while the information on the tumor surface is not much lost. The composite enhanced image can be used as an input of a threshold segmentation method, and the tumor region can be accurately segmented by adopting the threshold segmentation method on the composite enhanced image, and the accuracy is obviously higher than that of the original CT or MR image subjected to threshold segmentation directly. Meanwhile, the surface and the fuzzy boundary of the tumor in the original CT or MR image can be clearly seen by selecting different weighting factors in the synthesized enhanced image, so that a doctor can conveniently observe the surface and the boundary of the tumor.
The process of adjusting the window width and window level of the pixel value inside the ROI in the step A comprises the following steps: centering on the centroid of the ellipse, a rectangular region is selected, the length r of whichlengthLess than the major axis of the ellipse, the width r of the rectanglewidthA minor axis smaller than the ellipse; selecting the maximum gray value in the rectangle as the maximum value C of the window width in the window technologymaxSelecting the minimum gray value in the rectangle as the minimum value C of the window width in the window technologymin(ii) a Then arbitrarily selecting one of the following three formulas to map the pixel values within the ellipse:
wherein, IoriFor the original CT or MR image, α is the coefficient controlling the overall brightness of the image, γ1For adjustable coefficients, by varying gamma1Can be mapped in different ways, gamma1Equation (3) may convert a narrow range of input values to a wide range of output values when the value of (a) is less than 1, I being the image after window level adjustment.
The anisotropic diffusion method in the step B comprises the following steps: removing noise by using an anisotropic diffusion model, wherein the formula of the anisotropic diffusion model is as follows:
where div () is the divergence operator, c (x, y, t) is the diffusion coefficient, which is usually chosen as a function of the image gradient, t is time,represents the gradient operator and the gradient operator, respectively,it is the gradient of the diffusion coefficient c,is the gradient of image I, Δ represents the laplacian, and Δ I is the laplacian of image I.
The above formula for the anisotropic diffusion model is proposed by two scholars, Perona and Malik (two names), and also proposed by Perona and Malik: the function of the diffusion coefficient c (x, y, t) has two expressions:
k is a constant, representing the modulus of the gradient of image I. The two diffusion coefficient functions take the mode of the gradient of the image I as the basis of the diffusion speed, and the diffusion coefficient is small at the position with large gradient, so that the purpose of protecting the edge is achieved.
Solving the partial differential equation expressed by the formula (4) whose solution is expressed as Irn，IrnAnd B, obtaining a noise reduction image of the ROI covering the whole tumor area in the step B.
The multi-scale enhancement method in the step C comprises the following steps:
in order to obtain clear tumor boundaries, enhancement processing at different scales is performed on ROI regions covering the entire tumor region:
first, an ROI region covering the entire tumor region is downsampled by the formula (7), and a downsampled image I is obtainedre；
Ire＝REDUCE(I) (7)
Wherein, REDUCE () represents a down-sampling operation;
then, the downsampled image I is processed by the formula (8)reCarrying out zooming enhancement processing;
Iu＝(Ire-Ire*G)γ2(8)
wherein, IuIs a down-sampled enhanced image of reduced size, IreIs the down-sampled image, G is a gaussian kernel, the number of rows and columns of which can be adjusted, typically between 20 and 90.
Ire-IreG represents the high frequency component of the image, in which the tumor boundary is not clear, and gamma is calculated for the high frequency component to further enlarge the difference between the tumor region and the non-tumor region2Power of gamma2To control the coefficient of differentiation between tumor and non-tumor regions, due to IreThe size of the down-sampled image is smaller than that of the image I after window width and window level adjustment, and the down-sampled enhanced image I with reduced size is obtained through the formula (8)u，IuSize of and down-sampled image IreAre the same and are all smaller than the image I after window width and window level adjustment.
Then, the down-sampled enhanced image I with reduced size obtained by the formula (8) is processed by the formula (9)uPerforming upsampling to obtain Iex：Iex＝EXPAND(Iu) (9)
Wherein, IexFor the enhanced image of the ROI area covering the entire tumor area obtained in step C, IexIs the same as the size of the image I after window width and window level adjustment; EXPAND () represents an upsample operation.
EXPAND () is obtained by an extended interpolation operator, if the size is extended by one time, the pixels are enhanced by one time in the horizontal direction and the vertical direction, namely a value needs to be inserted between any two pixels of each line, and a line needs to be inserted between every two lines; the interpolation operator adopts a Bicubic interpolation method (Bicubic interpolation).
In the step D, a pixel-level weighted average fusion method is adopted for weighted fusion, and the formula is as follows:
Ien＝w·Irn+(1-w)·Iex(10)
wherein, IrnFor the noise-reduced image of the ROI area covering the entire tumor area obtained in step B, IexFor the enhanced image of the ROI area covering the whole tumor area obtained in step C, w is a weighting factor, IenThe image is enhanced for the synthesis of the ROI area covering the whole tumor area. The physician can easily observe the surface and the boundary of the tumor by selecting different weighting factors w in equation (10).
The reason for creating the synthetic enhanced image in step D is: the noise in the noise-reduced image in step B is reduced to some extent with respect to the original image, but the blurred boundary of the tumor is not clear, which is not favorable for segmentation. Although the boundary of the enhanced image in the step C becomes clear, the information on the tumor surface is lost more due to the scale transformation, which is not beneficial for the doctor to observe and diagnose. In order to simultaneously consider the segmentation of the tumor image and the observation and diagnosis of a doctor, the noise-reduced image of the ROI area covering all the tumor areas obtained in the step B and the enhanced image of the ROI area covering all the tumor areas obtained in the step C are subjected to weighted fusion to obtain a synthesized enhanced image of the ROI area covering all the tumor areas, the tumor boundary of the synthesized enhanced image is clear, and the information on the surface of the tumor is not lost much.
Next, taking a CT image as an example, the results obtained by the processing of the method are demonstrated.
The value range of the pixel value in the original CT image is between-1000 HU and +1000HU, so that the effect on a common display is poor. Thus, only images after adjustment outside the window width are provided below.
The image after window width and window level adjustment in step a is shown in fig. 2, and the ROI region in the image after window width and window level adjustment is shown in fig. 3, where the tumor boundary is blurred.
Referring to fig. 4, the noise of the ROI area in step B is reduced to some extent, but the tumor boundary is not improved significantly, as compared to fig. 3.
The enhanced image of the ROI region in step C is shown in fig. 5, where the boundary of the tumor becomes clear in fig. 5 compared to fig. 3, but the information on the tumor surface is more lost.
The synthetic enhanced image of the ROI region in step D is shown in fig. 6, and in comparison with fig. 4 and 5, fig. 6 is a compromise of fig. 4 and 5 in which the tumor boundary becomes clear and the tumor surface information is not much lost.
The embodiment of the invention also provides a tumor region image enhancement system based on the synthesized enhanced image, which comprises an ROI region selection unit, a window width and window level adjustment unit, a noise reduction unit, an enhancement unit and a weighted fusion unit, wherein:
the ROI area selection unit is used for: selecting an elliptical ROI (Region Of Interest) covering all tumor regions from a frame Of original CT image or MR image containing tumors;
the window width and window level adjusting unit is used for: setting all pixels outside the ROI area to be zero, and adjusting the window width and window level of the pixel value inside the ROI area to meet the requirement of human observation.
In practical application, the ROI region is selected first, and then the window width level is adjusted, but since the pixel value of the original medical image is between-1000 and +1000HU, and cannot be displayed on a common display, only the image after the window width level is adjusted is displayed in the embodiment of the present invention. The image after window width and window level adjustment is shown in FIG. 2; the ROI region in the window-width window-level adjusted image is shown in fig. 3.
The noise reduction unit is used for: performing noise reduction processing on the ROI region in the image after window width window level adjustment by using an Anisotropic diffusion method (Anisotropic diffusion), to obtain a noise-reduced image covering the ROI region of all tumor regions, as shown in fig. 4;
the enhancement unit is used for: performing enhancement processing on the ROI area in the image after window width and window level adjustment by adopting a multi-scale enhancement method to obtain an enhanced image covering the ROI area of all tumor areas, which is shown in FIG. 5;
the weighted fusion unit is used for: the noise-reduced image of the ROI region covering all the tumor regions and the enhanced image of the ROI region covering all the tumor regions are weighted and fused to obtain a synthesized enhanced image of the ROI region covering all the tumor regions, as shown in fig. 6.
The process of adjusting the window width and window level of the pixel value in the ROI area by the window width and window level adjusting unit comprises the following steps: centering on the centroid of the ellipse, a rectangular region is selected, the length r of whichlengthLess than the major axis of the ellipse, the width r of the rectanglewidthA minor axis smaller than the ellipse; selecting the maximum gray value in the rectangle as the maximum value C of the window width in the window technologymaxSelecting the minimum gray value in the rectangle as the minimum value C of the window width in the window technologymin(ii) a Then arbitrarily selecting one of the following three formulas to map the pixel values within the ellipse:
wherein, IoriFor the original CT or MR image, α is the coefficient controlling the overall brightness of the image, γ1For adjustable coefficients, by varying gamma1Can be mapped in different ways, gamma1Equation (3) may convert a narrow range of input values to a wide range of output values when the value of (a) is less than 1, I being the image after window level adjustment.
The anisotropic diffusion method adopted by the noise reduction unit comprises the following steps: removing noise by using an anisotropic diffusion model, wherein the formula of the anisotropic diffusion model is as follows:
where div () is the divergence operator, c (x, y, t) is the diffusion coefficient, which is usually chosen as a function of the image gradient, t is time,represents the gradient operator and the gradient operator, respectively,it is the gradient of the diffusion coefficient c,is the gradient of image I, Δ represents the laplacian, and Δ I is the laplacian of image I.
The above formula for the anisotropic diffusion model is proposed by two scholars, Perona and Malik (two names), and also proposed by Perona and Malik: the function of the diffusion coefficient c (x, y, t) has two expressions:
wherein,k is a constant, representing the modulus of the gradient of image I. The two diffusion coefficient functions take the mode of the gradient of the image I as the basis of the diffusion speed, and the diffusion coefficient is small at the position with large gradient, so that the purpose of protecting the edge is achieved.
Solving the partial differential equation expressed by the formula (4) whose solution is expressed as Irn，IrnA noise-reduced image of the ROI area covering the entire tumor area.
The multi-scale enhancement method adopted by the enhancement unit comprises the following steps:
in order to obtain clear tumor boundaries, enhancement processing at different scales is performed on ROI regions covering the entire tumor region:
first, an ROI region covering the entire tumor region is downsampled by the formula (7), and a downsampled image I is obtainedre；
Ire＝REDUCE(I) (7)
Wherein, REDUCE () represents a down-sampling operation;
then, the downsampled image I is processed by the formula (8)reCarrying out zooming enhancement processing;
Iu＝(Ire-Ire*G)γ2(8)
wherein, IuIs a down-sampled enhanced image of reduced size, IreIs the down-sampled image, G is a gaussian kernel, the number of rows and columns of which can be adjusted, typically between 20 and 90.
Ire-IreG represents the high frequency component of the image, in which the tumor boundary is not clear, and gamma is calculated for the high frequency component to further enlarge the difference between the tumor region and the non-tumor region2Power of gamma2To control the coefficient of differentiation between tumor and non-tumor regions, due to IreThe size of the down-sampled image is smaller than that of the image I after window width and window level adjustment, and the down-sampled enhanced image I with reduced size is obtained through the formula (8)u，IuSize of and down-sampled image IreAre the same and are all smaller than the image I after window width and window level adjustment.
Then, the down-sampled enhanced image I with reduced size obtained by the formula (8) is processed by the formula (9)uPerforming upsampling to obtain Iex：Iex＝EXPAND(Iu) (9)
Wherein, IexFor enhanced images of ROI regions covering the entire tumor region, IexIs the same as the size of the image I after window width and window level adjustment; EXPAND () represents an upsample operation.
EXPAND () is obtained by an extended interpolation operator, if the size is extended by one time, the pixels are enhanced by one time in the horizontal direction and the vertical direction, namely a value needs to be inserted between any two pixels of each line, and a line needs to be inserted between every two lines; the interpolation operator adopts a Bicubic interpolation method (Bicubic interpolation).
The weighted fusion unit performs weighted fusion by adopting a pixel-level weighted average fusion method, and the formula is as follows:
Ien＝w·Irn+(1-w)·Iex(10)
wherein, IrnTo de-noise images of the ROI area covering the entire tumor area, IexFor an enhanced image of the ROI area covering the whole tumor area, w is a weighting factor, IenThe image is enhanced for the synthesis of the ROI area covering the whole tumor area. The physician can easily observe the surface and the boundary of the tumor by selecting different weighting factors w in equation (10).
Therefore, the tumor region image enhancement method and system based on the synthesized enhanced image can be used for enabling the surface and the fuzzy boundary of the tumor in the original CT or MR image to be clearly visible, and facilitating the doctor to observe the surface and the boundary of the tumor. The tumor region can be accurately segmented by adopting a threshold segmentation method on the synthesized enhanced image.
Various modifications and variations of the embodiments of the present invention may be made by those skilled in the art, and they are also within the scope of the present invention, provided they are within the scope of the claims of the present invention and their equivalents.
What is not described in detail in the specification is prior art that is well known to those skilled in the art.

Claims (14)

1. A method for enhancing an image of a tumor region based on a synthetically enhanced image, comprising the steps of:
A. selecting an elliptical ROI (region of interest) covering all tumor regions from a frame of original CT (computed tomography) image or MR (magnetic resonance) image containing tumors; setting all pixels outside the ROI to be zero, and adjusting the window width and window level of the pixel value inside the ROI to meet the requirement of human observation;
B. performing noise reduction treatment on the ROI in the image after window width and window level adjustment by adopting an anisotropic diffusion method to obtain a noise reduction image of the ROI covering all tumor regions;
C. enhancing the ROI area in the image after window width and window level adjustment by adopting a multi-scale enhancement method to obtain an enhanced image of the ROI area covering all tumor areas;
the multi-scale enhancement method in the step C comprises the following steps:
enhancement processing is carried out on ROI areas covering the whole tumor area under different scales:
first, an ROI region covering the entire tumor region is downsampled by the formula (7), and a downsampled image I is obtainedre；
Ire＝REDUCE(I) (7)
Wherein, REDUCE () represents a down-sampling operation;
then, the downsampled image I is processed by the formula (8)reCarrying out zooming enhancement processing;
wherein, IuIs a down-sampled enhanced image of reduced size, IreIs a down-sampled image, G is a Gaussian kernel, Ire-IreG represents the high frequency component of the image, γ2To control the coefficients of differentiation between tumor and non-tumor regions, IuSize of and down-sampled image IreAre the same and are all smaller than the image I after window width and window level adjustment;
then, the down-sampled enhanced image I with reduced size obtained by the formula (8) is processed by the formula (9)uPerforming upsampling to obtain Iex：Iex＝EXPAND(Iu) (9)
Wherein, IexFor the enhanced image of the ROI area covering the entire tumor area obtained in step C, IexIs the same as the size of the image I after window width and window level adjustment; expandid () represents an upsample operation;
D. and D, performing weighted fusion on the noise-reduced image of the ROI area covering all the tumor areas obtained in the step B and the enhanced image of the ROI area covering all the tumor areas obtained in the step C to obtain a synthesized enhanced image of the ROI area covering all the tumor areas.
2. The method of image enhancement of a tumor region based on synthetically enhanced images as set forth in claim 1, wherein: the process of adjusting the window width and window level of the pixel value inside the ROI in the step A comprises the following steps: centering on the centroid of the ellipse, a rectangular region is selected, the length r of whichlengthLess than the major axis of the ellipse, the width r of the rectanglewidthA minor axis smaller than the ellipse; selecting the maximum gray value in the rectangle as the maximum value C of the window width in the window technologymaxSelecting the minimum gray value in the rectangle as the minimum value C of the window width in the window technologymin(ii) a Then arbitrarily selecting one of the following three formulas to map the pixel values within the ellipse:
wherein, IoriFor the original CT or MR image, α is the coefficient controlling the overall brightness of the image, γ1For adjustable coefficients, by varying gamma1I is the image after window width and level adjustment.
3. A method of image enhancement of a tumor region based on a synthetically enhanced image as claimed in claim 2 wherein: in step A,. gamma.1When less than 1, equation (3) converts a narrow range of input values to a wider range of output values.
4. A method of image enhancement of a tumor region based on a synthetically enhanced image as claimed in claim 2 wherein: the anisotropic diffusion method in the step B comprises the following steps: removing noise by using an anisotropic diffusion model, wherein the formula of the anisotropic diffusion model is as follows:
where div () is the divergence operator, c (x, y, t) is the diffusion coefficient, which is chosen as a function of the image gradient, t is time,represents the gradient operator and the gradient operator, respectively,it is the gradient of the diffusion coefficient c,is the gradient of image I, Δ represents the laplacian, Δ I is the laplacian of image I;
solving the partial differential equation expressed by the formula (4) whose solution is expressed as Irn，IrnAnd B, obtaining a noise reduction image of the ROI covering the whole tumor area in the step B.
5. The method of image enhancement of a tumor region based on a synthetically enhanced image as set forth in claim 4, wherein: in step B, the function of the diffusion coefficient c (x, y, t) in the formula of the anisotropic diffusion model has two expression modes:
wherein,k is a constant, the two diffusion coefficient functions take the mode of the gradient of the image I as the basis of diffusion speed, and the diffusion coefficient is small at the position with large gradient, so that the purpose of protecting the edge is achieved.
6. The method of image enhancement of a tumor region based on synthetically enhanced images as set forth in claim 1, wherein: the EXPAND () in the step C is obtained by an extended interpolation operator, if the size is extended by one time, the pixel of one time is enhanced in the horizontal direction and the vertical direction, namely a value needs to be inserted between any two pixels of each line, and a line needs to be inserted between every two lines; the interpolation operator adopts a double cubic interpolation method.
7. The method of image enhancement of a tumor region based on synthetically enhanced images as set forth in claim 1, wherein: in the step D, a pixel-level weighted average fusion method is adopted for weighted fusion, and the formula is as follows:
Ien＝w·Irn+(1-w)·Iex(10)
wherein, IrnFor the noise-reduced image of the ROI area covering the entire tumor area obtained in step B, IexFor the enhanced image of the ROI area covering the whole tumor area obtained in step C, w is a weighting factor, IenA composite enhanced image of the ROI region covering the entire tumor region; by selecting different weighting factors w in equation (10), the physician can easily observe the surface and boundary of the tumor.
8. A tumor region image enhancement system based on a synthetically enhanced image, characterized by: the system comprises an ROI region selection unit, a window width and window level adjustment unit, a noise reduction unit, an enhancement unit and a weighted fusion unit, wherein:
the ROI area selection unit is used for: selecting an elliptical ROI (region of interest) covering all tumor regions from a frame of original CT (computed tomography) image or MR (magnetic resonance) image containing tumors;
the window width and window level adjusting unit is used for: setting all pixels outside the ROI to be zero, and adjusting the window width and window level of the pixel value inside the ROI to meet the requirement of human observation;
the noise reduction unit is used for: performing noise reduction treatment on the ROI in the image after window width and window level adjustment by adopting an anisotropic diffusion method to obtain a noise reduction image of the ROI covering all tumor regions;
the enhancement unit is configured to: enhancing the ROI area in the image after window width and window level adjustment by adopting a multi-scale enhancement method to obtain an enhanced image of the ROI area covering all tumor areas;
the multi-scale enhancement method adopted by the enhancement unit comprises the following steps:
enhancement processing is carried out on ROI areas covering the whole tumor area under different scales:
first, an ROI region covering the entire tumor region is downsampled by the formula (7), and a downsampled image I is obtainedre；
Ire＝REDUCE(I) (7)
Wherein, REDUCE () represents a down-sampling operation;
then, the downsampled image I is processed by the formula (8)reCarrying out zooming enhancement processing;
wherein, IuIs a down-sampled enhanced image of reduced size, IreIs a down-sampled image, G is a Gaussian kernel, Ire-IreG represents the high frequency component of the image, γ2To control the coefficients of differentiation between tumor and non-tumor regions, IuSize of and down-sampled image IreAre the same and are all smaller than the image I after window width and window level adjustment;
then, the down-sampled enhanced image I with reduced size obtained by the formula (8) is processed by the formula (9)uPerforming upsampling to obtain Iex：Iex＝EXPAND(Iu) (9)
Wherein, IexFor the enhanced image of the ROI area covering the entire tumor area obtained in step C, IexIs the same as the size of the image I after window width and window level adjustment; expandid () represents an upsample operation;
the weighted fusion unit is configured to: and performing weighted fusion on the noise-reduced image of the ROI covering all the tumor regions and the enhanced image of the ROI covering all the tumor regions to obtain a synthesized enhanced image of the ROI covering all the tumor regions.
9. A composite enhanced image based tumor region image enhancement system as claimed in claim 8 wherein: the window width and window level adjusting unit adjusts the window width and window level of the pixel value in the ROI region by the following steps: centering on the centroid of the ellipse, a rectangular region is selected, the length r of whichlengthLess than the major axis of the ellipse, the width r of the rectanglewidthA minor axis smaller than the ellipse; selecting the maximum gray value in the rectangle as the maximum value C of the window width in the window technologymaxSelecting the minimum gray value in the rectangle as the minimum value C of the window width in the window technologymin(ii) a Arbitrarily selecting one of the following three formulas to map the pixel values within the ellipse:
wherein, IoriFor the original CT or MR image, α is the coefficient controlling the overall brightness of the image, γ1For adjustable coefficients, by varying gamma1I is the image after window width and level adjustment.
10. A composite enhanced image based tumor region image enhancement system as claimed in claim 9 wherein: gamma ray1When less than 1, equation (3) converts a narrow range of input values to a wider range of output values.
11. A composite enhanced image based tumor region image enhancement system as claimed in claim 9 wherein: the anisotropic diffusion method adopted by the noise reduction unit comprises the following steps: removing noise by using an anisotropic diffusion model, wherein the formula of the anisotropic diffusion model is as follows:
where div () is the divergence operator, c (x, y, t) is the diffusion coefficient, which is chosen as a function of the image gradient, t is time,represents the gradient operator and the gradient operator, respectively,it is the gradient of the diffusion coefficient c,is the gradient of image I, Δ represents the laplacian, Δ I is the laplacian of image I;
solving the partial differential equation expressed by the formula (4) whose solution is expressed as Irn，IrnA noise-reduced image of the ROI area covering the entire tumor area.
12. A composite enhanced image based tumor region image enhancement system as claimed in claim 11 wherein: the function of the diffusion coefficient c (x, y, t) in the formula of the anisotropic diffusion model has two expression modes:
wherein,k is a constant, the two diffusion coefficient functions take the mode of the gradient of the image I as the basis of diffusion speed, and the diffusion coefficient is small at the position with large gradient, so that the purpose of protecting the edge is achieved.
13. A composite enhanced image based tumor region image enhancement system as claimed in claim 8 wherein: the EXPAND () is obtained by an extended interpolation operator, if the size is extended by one time, the pixel of one time is enhanced in the horizontal direction and the vertical direction, namely a value needs to be inserted between any two pixels of each line, and a line needs to be inserted between every two lines; the interpolation operator adopts a double cubic interpolation method.
14. A composite enhanced image based tumor region image enhancement system as claimed in claim 8 wherein: the weighted fusion unit performs weighted fusion by adopting a pixel-level weighted average fusion method, and the formula is as follows:
Ien＝w·Irn+(1-w)·Iex(10)
wherein, IrnTo de-noise images of the ROI area covering the entire tumor area, IexFor an enhanced image of the ROI area covering the whole tumor area,w is a weighting factor, IenA composite enhanced image of the ROI region covering the entire tumor region; by selecting different weighting factors w in equation (10), the physician can easily observe the surface and boundary of the tumor.
CN201610054646.7A 2016-01-27 2016-01-27 Tumor region image enchancing method and system based on synthesis enhancing image Expired - Fee Related CN105741241B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201610054646.7A CN105741241B (en) 2016-01-27 2016-01-27 Tumor region image enchancing method and system based on synthesis enhancing image

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201610054646.7A CN105741241B (en) 2016-01-27 2016-01-27 Tumor region image enchancing method and system based on synthesis enhancing image

Publications (2)

Publication Number Publication Date
CN105741241A CN105741241A (en) 2016-07-06
CN105741241B true CN105741241B (en) 2018-08-17

Family

ID=56247645

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201610054646.7A Expired - Fee Related CN105741241B (en) 2016-01-27 2016-01-27 Tumor region image enchancing method and system based on synthesis enhancing image

Country Status (1)

Country Link
CN (1) CN105741241B (en)

Families Citing this family (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106600550B (en) * 2016-11-29 2020-08-11 深圳开立生物医疗科技股份有限公司 Ultrasonic image processing method and system
CN106709929A (en) * 2016-12-30 2017-05-24 上海联影医疗科技有限公司 Method and device for displaying interesting region of medical image
CN106600568B (en) * 2017-01-19 2019-10-11 东软医疗系统股份有限公司 A kind of low-dose CT image de-noising method and device
US11100632B2 (en) * 2018-04-13 2021-08-24 Elekta, Inc. Image synthesis using adversarial networks such as for radiation therapy
CN110163862B (en) * 2018-10-22 2023-08-25 腾讯科技（深圳）有限公司 Image semantic segmentation method and device and computer equipment
CN109492587B (en) * 2018-11-12 2021-06-22 浙江宇视科技有限公司 Image processing method and device
CN110135247B (en) * 2019-04-03 2021-09-24 深兰科技（上海）有限公司 Data enhancement method, device, equipment and medium in pavement segmentation
CN110162953A (en) * 2019-05-31 2019-08-23 Oppo（重庆）智能科技有限公司 Biometric discrimination method and Related product
CN110706791B (en) * 2019-09-30 2020-10-16 杭州依图医疗技术有限公司 Medical image processing method and device
CN111402261B (en) * 2020-02-25 2021-06-08 四川大学青岛研究院 Improved skull segmentation method based on Hessian matrix enhancement
CN111489314A (en) * 2020-04-16 2020-08-04 东软医疗系统股份有限公司 Image enhancement method and device and terminal equipment

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1980321A (en) * 2005-12-09 2007-06-13 逐点半导体（上海）有限公司 Picture reinforcing treatment system and treatment method
CN101393644A (en) * 2008-08-15 2009-03-25 华中科技大学 Hepatic portal vein tree modeling method and system thereof
CN101587586A (en) * 2008-05-20 2009-11-25 株式会社理光 Device and method for processing images
CN101706843A (en) * 2009-11-16 2010-05-12 杭州电子科技大学 Interactive film Interpretation method of mammary gland CR image
CN101727658A (en) * 2008-10-14 2010-06-09 深圳迈瑞生物医疗电子股份有限公司 Image processing method and device
CN103150714A (en) * 2013-03-12 2013-06-12 华东师范大学 Method and device for real-time interactive enhancement of magnetic resonance image

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1980321A (en) * 2005-12-09 2007-06-13 逐点半导体（上海）有限公司 Picture reinforcing treatment system and treatment method
CN101587586A (en) * 2008-05-20 2009-11-25 株式会社理光 Device and method for processing images
CN101393644A (en) * 2008-08-15 2009-03-25 华中科技大学 Hepatic portal vein tree modeling method and system thereof
CN101727658A (en) * 2008-10-14 2010-06-09 深圳迈瑞生物医疗电子股份有限公司 Image processing method and device
CN101706843A (en) * 2009-11-16 2010-05-12 杭州电子科技大学 Interactive film Interpretation method of mammary gland CR image
CN103150714A (en) * 2013-03-12 2013-06-12 华东师范大学 Method and device for real-time interactive enhancement of magnetic resonance image

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
基于CR图像的多尺度对比增强算法研究;葛日波;《公安海警高等专科学校学报》;20060815;第5卷(第3期);第16-17页 *
改进的各向异性扩散图像去噪模型;郑满满;《中国优秀硕士学位论文全文数据库 信息科技辑》;20150115(第1期);第9-10页 *

Also Published As

Publication number Publication date
CN105741241A (en) 2016-07-06

Similar Documents

Publication Publication Date Title
CN105741241B (en) Tumor region image enchancing method and system based on synthesis enhancing image
US10997701B1 (en) System and method for digital image intensity correction
EP2545527B1 (en) Probabilistic refinement of model-based segmentation
US9697603B2 (en) Medical image data processing system and method for vessel segmentation using pre- and post-contrast data
JP5753791B2 (en) Method for providing denominated predetermined resolution medical image, system for providing predetermined denoised predetermined resolution medical image
US7860331B2 (en) Purpose-driven enhancement filtering of anatomical data
US20130202177A1 (en) Non-linear resolution reduction for medical imagery
Sreeja et al. An improved feature based image fusion technique for enhancement of liver lesions
CN109949318B (en) Full convolution neural network epileptic focus segmentation method based on multi-modal image
Bhairannawar Efficient medical image enhancement technique using transform HSV space and adaptive histogram equalization
CN106911904B (en) Image processing method, image processing system and imaging system
JP2014014673A (en) Method for acquisition of angiographic image
US20160035102A1 (en) Method and system for computing digital tomosynthesis images
JP2016174905A (en) Ultrasound diagnostic device, medical image processor and medical image processing program
Wen et al. A novel Bayesian-based nonlocal reconstruction method for freehand 3D ultrasound imaging
Dawood et al. The importance of contrast enhancement in medical images analysis and diagnosis
WO2022207238A1 (en) Methods and systems for biomedical image segmentation based on a combination of arterial and portal image information
US8805122B1 (en) System, method, and computer-readable medium for interpolating spatially transformed volumetric medical image data
US8224046B2 (en) Visual enhancement of interval changes using rigid and non-rigid transformations
Yang et al. Techniques and software tool for 3D multimodality medical image segmentation
US20110123089A1 (en) Method and System for Enhancing Contrast of Spatially-Localized Phenomena in Mammography Image
Macho et al. Segmenting Teeth from Volumetric CT Data with a Hierarchical CNN-based Approach.
Tamilselvan et al. Brain tumor detection from clinical CT and MRI images using WT-FCM algorithm
Guo et al. LRENet: a location-related enhancement network for liver lesions in CT images
Kang et al. Interactive Liver Segmentation Algorithm Based on Geodesic Distance and V-Net

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into 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
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20180817

Termination date: 20210127