CN116363181A - Feature-based CT image and ultrasonic image liver registration method - Google Patents
Feature-based CT image and ultrasonic image liver registration method Download PDFInfo
- Publication number
- CN116363181A CN116363181A CN202310291088.6A CN202310291088A CN116363181A CN 116363181 A CN116363181 A CN 116363181A CN 202310291088 A CN202310291088 A CN 202310291088A CN 116363181 A CN116363181 A CN 116363181A
- Authority
- CN
- China
- Prior art keywords
- image
- liver
- slices
- ultrasonic
- slice
- 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
- 210000004185 liver Anatomy 0.000 title claims abstract description 104
- 238000000034 method Methods 0.000 title claims abstract description 54
- 239000011159 matrix material Substances 0.000 claims abstract description 26
- 230000009466 transformation Effects 0.000 claims abstract description 18
- 238000002604 ultrasonography Methods 0.000 claims description 17
- 238000012952 Resampling Methods 0.000 claims description 8
- 238000004364 calculation method Methods 0.000 claims description 7
- 238000004422 calculation algorithm Methods 0.000 claims description 5
- 238000001914 filtration Methods 0.000 claims description 5
- 230000011218 segmentation Effects 0.000 claims description 5
- 238000013527 convolutional neural network Methods 0.000 claims description 4
- 238000003708 edge detection Methods 0.000 claims description 3
- 230000001131 transforming effect Effects 0.000 abstract description 3
- 238000007781 pre-processing Methods 0.000 abstract description 2
- 238000012216 screening Methods 0.000 abstract description 2
- 230000004927 fusion Effects 0.000 description 5
- 210000000056 organ Anatomy 0.000 description 4
- 238000011524 similarity measure Methods 0.000 description 4
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 description 3
- 238000011161 development Methods 0.000 description 3
- 230000018109 developmental process Effects 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 230000004069 differentiation Effects 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000002679 ablation Methods 0.000 description 1
- 238000003759 clinical diagnosis Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 238000002059 diagnostic imaging Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 238000007499 fusion processing Methods 0.000 description 1
- 201000007270 liver cancer Diseases 0.000 description 1
- 208000014018 liver neoplasm Diseases 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000002980 postoperative effect Effects 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 238000011426 transformation method Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
- G06T7/33—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods
- G06T7/344—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods involving models
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformations in the plane of the image
- G06T3/40—Scaling of whole images or parts thereof, e.g. expanding or contracting
- G06T3/4007—Scaling of whole images or parts thereof, e.g. expanding or contracting based on interpolation, e.g. bilinear interpolation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10072—Tomographic images
- G06T2207/10081—Computed x-ray tomography [CT]
-
- 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
- G06T2207/10—Image acquisition modality
- G06T2207/10132—Ultrasound image
-
- 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
- G06T2207/20084—Artificial neural networks [ANN]
-
- 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
- G06T2207/30056—Liver; Hepatic
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Data Mining & Analysis (AREA)
- Radiology & Medical Imaging (AREA)
- Quality & Reliability (AREA)
- Life Sciences & Earth Sciences (AREA)
- Medical Informatics (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Ultra Sonic Daignosis Equipment (AREA)
Abstract
The invention discloses a feature-based CT image and ultrasonic image liver registering method, which comprises the steps of firstly, carrying out image preprocessing on CT and extracting outline features of liver organs from the CT image; collecting ultrasonic images and extracting outline features of liver organs from the ultrasonic images; then, according to the area characteristics of the liver organs, carrying out quick rough matching on the CT image and the ultrasonic image, and screening out a plurality of candidate CT slices; and registering the candidate CT slices with the ultrasonic image one by one to obtain a space transformation matrix, transforming the candidate CT slices under an ultrasonic image coordinate system, and calculating the image similarity between the transformed candidate CT slices and the two-dimensional ultrasonic to obtain CT slices which are accurately matched with the two-dimensional ultrasonic. The invention can improve the accuracy and efficiency of registration.
Description
Technical Field
The invention relates to the technical field of medical image registration, in particular to a feature-based CT image and ultrasonic image liver registration method.
Background
With the rapid development of computer technology and medical imaging technology, the processing of medical images by means of image processing technology has shown important application value and development prospect in clinic. Different medical images contain different diagnostic information, different medical images are required to be fused and compared clinically, and the information contained in the different images is comprehensively utilized, so that doctors can know the illness state in an omnibearing and multi-angle manner, and the diagnostic result is more accurate. The image registration technology is widely applied to aspects of diagnosis of clinical diseases, determination of treatment schemes, evaluation of treatment effects and the like as a precondition for realizing image fusion, and is a popular research direction.
The medical image registration technology is an important component of medical image processing, and registration plays a very important role in clinical diagnosis, surgical navigation, surgical planning, postoperative evaluation and the like by fusion comparison of information of different medical images. In particular, along with the rapid development of medical image technology, the multi-mode medical image registration plays a very important role in a plurality of clinical applications, so the method has important theoretical significance and clinical value for the research of the multi-mode medical image registration algorithm.
Medical image registration is classified into a gray-scale based registration method and a feature-based registration method. The gray level-based registration method directly utilizes the gray level of the whole image to measure the similarity between two images, and adopts a search method (such as a maximum mutual information method, a correlation method, a conditional entropy method, a joint entropy method and the like) to search for the point with the maximum or minimum value of the similarity measure so as to determine the transformation model parameters between the two images. The gray level-based registration method has the defects of large calculated amount, long registration time, sensitivity to scaling rotation, distortion and the like, neglecting spatial related information of the image and the like. Feature-based registration methods generally extract a feature of two images, and commonly used image features include point features, straight line segments, edges, contours, closed regions, statistical moments, and the like. The feature-based registration method has the advantages of simple operation, high registration speed, high precision and the like, but also has the disadvantages of needing manual intervention, difficult acquisition of feature points and the like.
Disclosure of Invention
The invention aims to solve the problems that the existing feature-based registration method needs manual intervention and the acquisition of feature points is difficult, and provides a feature-based CT image and ultrasonic image liver registration method.
In order to solve the problems, the invention is realized by the following technical scheme:
a feature-based CT image and ultrasonic image liver registration method comprises the following steps:
step 1, acquiring CT images of human livers, and extracting outline features of liver organs from each CT slice of the CT images;
step 2, collecting an ultrasonic image of the liver of a human body, and extracting outline features of liver organs from the ultrasonic image;
step 3, performing rough registration on all CT slices of the CT image and the ultrasonic image based on the contour features of the liver organ extracted in the steps 1 and 2 to obtain rough registered CT slices, namely;
step 3.1, calculating the area of the outline features of the liver organ of each CT slice of the CT image to obtain the surface area of the liver organ of each CT slice; simultaneously calculating the area of the outline features of the liver organ of the ultrasonic image to obtain the surface area of the liver organ of the ultrasonic image;
step 3.2, firstly calculating absolute values of differences of surface areas of liver organs of every two adjacent CT slices of the CT image, and then calculating average values of the absolute values of the differences of the surface areas of the liver organs of every two adjacent CT slices to serve as average differences of the CT image;
step 3.3, calculating the absolute value of the difference value between the surface area of the liver organ of each CT slice of the CT image and the surface area of the liver organ of the ultrasonic image as a comparison difference value of each CT slice;
step 3.4, comparing the comparison difference value of each CT slice of the CT image with the average difference value of the CT image, and reserving CT slices of which the comparison difference value of the CT slices in the CT image is smaller than the average difference value of the CT image, thereby obtaining coarse registered CT slices;
step 4, based on the contour features of the liver organs extracted in the steps 1 and 2, performing fine registration on the roughly registered CT slices of the CT images obtained in the step 3 and the ultrasonic images to obtain fine registered CT slices, namely;
step 4.1, performing image space transformation on each of the roughly registered CT slices by taking the mean square error of the binary matrix of the CT slice image and the binary matrix of the ultrasonic image as an objective function so as to transform each of the roughly registered CT slices into an ultrasonic space coordinate system to obtain the roughly registered CT slices after space transformation;
and 4.2, calculating the overlapping rate of the spatially transformed coarse registration CT slices and the ultrasonic image by using dice coefficients, and taking the spatially transformed coarse registration CT slices with the highest overlapping rate as the fine registration CT slices.
The specific process of the step 1 is as follows:
step 1.1, carrying out window adjustment on each CT slice;
step 1.2, resampling each CT slice subjected to window adjustment by using a bilinear interpolation method;
and 1.3, inputting each resampled CT slice into a UNet convolutional neural network to obtain a binary mask of the liver organ, and extracting the outline features of the liver organ of each CT slice.
The specific process of the step 2 is as follows:
step 2.1, window adjustment is carried out on the ultrasonic image;
step 2.2, resampling the ultrasonic image subjected to window adjustment by using a bilinear interpolation method;
and 2.3, obtaining a binary mask, namely a contour feature of the liver organ by adopting a pixel gradient segmentation method for the resampled ultrasonic image, thereby extracting the contour feature of the liver organ of the ultrasonic image.
The specific process of the step 2.3 is as follows: denoising the resampled ultrasonic image by using twice median filtering; extracting outline edge characteristics of the liver organs by adopting a Canny edge detection algorithm; and binarizing the outline edge image, and filling the closed region to obtain a binary mask of the liver organ of the ultrasonic image.
In the above step 4.2, the calculation formula of the overlapping ratio Dice is:
wherein I is US Is a binary matrix of ultrasonic images, I' CT Is a binary matrix of spatially transformed coarsely registered CT slices.
Compared with the prior art, the invention has the following characteristics:
1. the method for segmenting the liver organ features extracts the CT slice and the liver organ binary mask of the ultrasonic image, and the common feature description of the binary mask enables the CT slice to be related to the pixel value of the liver organ in the two-dimensional ultrasonic, so that the multi-mode registration problem is simplified into a single-mode registration problem.
2. The coarse registration is added before the fine registration, and the CT slices similar to the liver areas in the ultrasonic images can be efficiently searched by comparing the liver areas in the ultrasonic images with the liver areas in the CT slices, so that most CT slices are eliminated, the subsequent registration times are greatly reduced, and the calculation efficiency of the whole frame is improved.
3. Considering the essence of binary image registration, namely registration based on pixel value correlation between two images, compared with a gray level image, the correlation of the binary image on the pixel value is more obvious, the invention adopts the mean square error MSE as the similarity measure of the registration frame, and can simultaneously consider the calculation efficiency and the registration effect.
Drawings
Fig. 1 is a schematic diagram of a feature-based CT image and ultrasound image liver registration method.
Detailed Description
The present invention will be further described in detail with reference to specific examples in order to make the objects, technical solutions and advantages of the present invention more apparent.
A feature-based CT image and ultrasound image liver registration method, as shown in fig. 1, comprising the steps of:
step 1, acquiring CT images of the liver of a human body in a preoperative stage, and extracting outline features of liver organs from each CT slice of the CT images.
The CT image consists of a plurality of CT slices, each CT slice is a two-dimensional image, and the CT image formed by all CT slices is a three-dimensional image. For this purpose, feature extraction needs to be performed on each CT slice of the CT image, so as to obtain the contour features of the liver organ of each CT slice, which specifically includes the following steps:
and step 1.1, carrying out window adjustment on the CT slice. Due to liver registration, the window width and level were adjusted to 400Hu and 60Hu to increase the differentiation of the liver from surrounding organs.
And 1.2, resampling the CT slice subjected to window adjustment by using a bilinear interpolation method, and restoring the real size of the organ in the physical space.
And 1.3, inputting the resampled CT slice into a UNet convolutional neural network to obtain a binary mask of the liver organ. Because the resolution of the CT slice is obviously different from the pixel value of the region where the liver organ is located and the surrounding tissues, the present invention adopts the existing UNet convolutional neural network to divide the liver region of the CT slice, namely, the CT slice subjected to window adjustment and resampling is input into the UNet network to obtain the binary mask of the liver organ, and the binary mask of the liver organ is the outline feature of the liver organ.
And 2, acquiring an ultrasonic image of the liver of the human body in the intraoperative stage, and extracting the outline features of the liver organ from the ultrasonic image.
The ultrasonic image of the liver of the human body consists of one ultrasonic image and is a two-dimensional image. For this purpose, only the feature extraction is needed for the ultrasonic image, and the outline features of the liver organ of the ultrasonic image are obtained, which comprises the following specific procedures:
and 2.1, performing window adjustment on the ultrasonic image. Due to liver registration, the window width and level were adjusted to 400Hu and 60Hu to increase the differentiation of the liver from surrounding organs.
And 2.2, resampling the ultrasonic image subjected to window adjustment by using a bilinear interpolation method, and restoring the real size of the organ in the physical space.
And 2.3, obtaining a binary mask of the liver organ by adopting a pixel gradient segmentation method for the resampled ultrasonic image. The process of the pixel gradient segmentation method specifically comprises the following steps: firstly, because noise shadows exist in an ultrasonic image, the method uses twice median filtering to remove noise, the first filtering uses small-size kernels to remove the whole white noise of the image, and meanwhile ensures gradient information of the contour edge as much as possible, and the second filtering uses large-size kernels to remove the noise on the contour edge. And extracting the outline edge characteristics of the liver organ by adopting a Canny edge detection algorithm, binarizing the outline edge image, and filling the closed region to obtain an ultrasonic image liver organ binary mask, wherein the liver organ binary mask is the outline characteristics of the liver organ.
And 3, performing coarse registration on the three-dimensional CT image and the two-dimensional ultrasonic image to obtain CT slices in the three-dimensional CT image, which are in coarse registration with the two-dimensional ultrasonic image.
The method is characterized in that the CT slice and the liver organ binary mask of the ultrasonic image are extracted based on the segmentation method of the liver organ characteristics, and the CT slice is related to the pixel value of the liver organ in the two-dimensional ultrasonic by the common characteristic description of the binary mask, so that the multi-mode registration problem is simplified into the binary image registration problem.
In order to improve the registration speed from three-dimensional CT to two-dimensional ultrasound, the invention provides a novel method for quickly and roughly matching slices based on the area characteristics of liver organs, and because the outline characteristics of the liver organs are binary masks, the surface area of the outline characteristics can be obtained by counting the number of pixel points of the outline characteristics, and thus the matching relationship between CT slices in a three-dimensional CT image and a two-dimensional ultrasound image can be preliminarily determined by comparing liver area similarities between three-dimensional CT and two-dimensional ultrasound one by one without registration alignment. The three-dimensional CT-two-dimensional ultrasonic rough matching flow is as follows:
step 3.1, calculating the area of the outline feature of the liver organ of each CT slice of the CT image to obtain the surface area of the liver organ of each CT sliceWhere i=1, 2, …, n, n is the number of CT slices in the CT image. Simultaneously, calculating the area of the outline features of the liver organ of the ultrasonic image to obtain the surface area S of the liver organ of the ultrasonic image US 。
Firstly taking adjacent CT slices one by one, and calculating absolute value of difference value of surface area of liver organ of every two adjacent CT slices of CT image
Then calculate the average value of the absolute value of the difference of the surface areas of liver organs of every two adjacent CT slices as the average difference of CT images
Calculating the absolute value of the difference between the surface area of the liver organ of each CT slice of the CT image and the surface area of the liver organ of the ultrasound image
Step 3.4, comparing the difference value of each CT slice of the CT imageAverage difference from CT images, respectively->Comparison is performed: if->When the CT slice is in the rough registration, reserving the ith CT slice, wherein the reserved CT slice is the CT slice in the rough registration; otherwise, the ith CT slice is deleted.
In the rough matching process, the liver organ area difference between the ultrasonic image and the CT slice is compared with the liver organ area difference average value of the CT slices, so that the CT slice to be selected closest to the ultrasonic image can be efficiently searched, most CT slices are eliminated, the subsequent registration times are greatly reduced, and the calculation efficiency of the whole frame is improved.
And 4, performing fine registration on the three-dimensional CT image and the two-dimensional ultrasonic image to obtain a CT slice in the three-dimensional CT image, which is subjected to fine registration with the two-dimensional ultrasonic image.
After a plurality of reserved CT slices to be selected (CT slices with rough registration) are obtained through three-dimensional CT-two-dimensional ultrasonic rough matching, the CT slices with rough registration and ultrasonic are required to be subjected to fine registration one by one. The fine registration may employ existing image registration methods. The precise registration method adopted by the invention comprises the following specific processes:
and 4.1, firstly taking the mean square error of the binary matrix of the CT slice image and the binary matrix of the ultrasonic image as an objective function, updating registration parameters by the Adam optimizer according to the gradient of the objective function until the objective function converges to a global optimal value or the iteration number reaches a set value, obtaining a rigid transformation matrix of the CT slice to the ultrasonic image, and transforming each roughly registered CT slice into an ultrasonic space coordinate system by utilizing the rigid transformation matrix of the CT slice to the ultrasonic image to obtain the roughly registered CT slice after space transformation.
Considering the nature of binary image registration, i.e. registration based on pixel value correlation between two images, binary images are more pronounced in pixel value correlation than gray scale images. Based on the characteristics, the method adopts the mean square error MSE of the binary matrixes of the CT slice image and the ultrasonic image as the similarity measure of the registration frame, and can simultaneously consider the calculation efficiency and the registration effect. Based on the selected similarity measure, the invention adopts a rigid transformation method to carry out image space transformation on CT slice-ultrasound.
Let ultrasound images be the stationary image and CT slices be the moving image. The rigid transformation matrix of the CT slice to the ultrasound image is obtained by iterative optimization of equation (1). The Adam optimizer updates the registration parameters according to the gradient of the objective function MSE of equation (2) until the objective function converges to the global optimum or the number of iterations reaches a set maximum.
Wherein I is US Is a binary matrix of ultrasonic images, I CT Is a binary matrix of CT slices before spatial transformation,for CT slice to ultrasound imageRigid transformation matrix, MSE (I CT ,I US ) The mean square error of the binary matrix of the CT slice and the ultrasonic image is N, which is the number of pixels (the number of pixels of the CT slice and the ultrasonic image is the same as N because the actual sizes of liver organs in the CT slice and the ultrasonic image are restored through image resampling in the steps above)>For the ith gray value in the binary matrix of the CT slice,/th gray value>Is the gray value of the ith element in the binary matrix of the ultrasonic image.
After the rigid transformation matrix between the ultrasound and the CT slices to be selected is obtained by calculation of the Adam optimizer, the rigid transformation matrix can be utilized to transform the roughly registered CT slices to an ultrasound space coordinate system, namely
Wherein I is CT Binary matrix for spatially transformed coarsely registered CT slices, I' CT For a binary matrix of spatially transformed coarsely registered CT slices,is a rigid transformation matrix of CT slices into ultrasound images.
And 4.2, calculating the overlapping rate of the spatially transformed coarse registered CT slices and the ultrasonic image, and taking the spatially transformed coarse registered CT slices with the highest overlapping rate as the finely registered CT slices.
In order to measure the similarity of liver organ characteristics between two images, the overlapping rate of the binary mask of the CT slice and the ultrasonic image of the coarse registration after the space transformation is calculated by using Dice Coefficient (DC), and the formula is as follows:
wherein I is US Is a binary matrix of ultrasonic images, I' CT Is a binary matrix of spatially transformed coarsely registered CT slices.
CT slice I 'with highest overlap ratio Dice' CT Namely, the CT slice (the precisely registered CT slice) which is closest to the ultrasonic image, so that the precise matching relation between the three-dimensional CT image and the two-dimensional ultrasonic image is determined.
Aiming at the lack of medical image guidance with strong instantaneity and high resolution in liver cancer puncture ablation operation based on ultrasonic guidance, the invention aims to register and fuse a three-dimensional CT image obtained before operation with a real-time two-dimensional ultrasonic image in operation, combines the advantages of high resolution of the CT image and strong instantaneity of the ultrasonic image, and adopts a CT-ultrasonic fusion image guidance mode to clearly and real-time locate the focus position, thereby guiding a clinician to complete puncture operation and improving puncture precision. In the preoperative stage, the CT is subjected to image preprocessing and the outline features of liver organs are extracted from the CT; in the intraoperative stage, acquiring an ultrasonic image in real time and extracting outline features of liver organs from the ultrasonic image; carrying out quick rough matching on the CT image and the ultrasonic image according to the area characteristics of the liver organ, and screening out a plurality of candidate CT slices; registering the candidate CT slices and the ultrasonic image one by one to obtain a space transformation matrix, transforming the candidate CT slices under an ultrasonic image coordinate system, and obtaining CT slices which are accurately matched with the two-dimensional ultrasonic by calculating the image similarity between the transformed candidate CT slices and the two-dimensional ultrasonic. Therefore, the image fusion technology can be adopted to fuse the precisely registered CT slice with the ultrasonic image, and the image fusion process is not the focus of the invention and can be realized by adopting the existing fusion algorithm.
It should be noted that, although the examples described above are illustrative, this is not a limitation of the present invention, and thus the present invention is not limited to the above-described specific embodiments. Other embodiments, which are apparent to those skilled in the art from consideration of the specification and practice of the invention disclosed herein, are considered to be within the scope of the invention as claimed.
Claims (5)
1. The feature-based CT image and ultrasonic image liver registration method is characterized by comprising the following steps:
step 1, acquiring CT images of human livers, and extracting outline features of liver organs, namely binary masks, from each CT slice of the CT images;
step 2, collecting an ultrasonic image of the liver of a human body, and extracting outline features of liver organs, namely binary masks, from the ultrasonic image;
step 3, performing rough registration on all CT slices of the CT image and the ultrasonic image based on the contour features of the liver organ extracted in the steps 1 and 2 to obtain rough registered CT slices, namely;
step 3.1, calculating the area of the outline features of the liver organ of each CT slice of the CT image to obtain the surface area of the liver organ of each CT slice; simultaneously calculating the area of the outline features of the liver organ of the ultrasonic image to obtain the surface area of the liver organ of the ultrasonic image;
step 3.2, firstly calculating absolute values of differences of surface areas of liver organs of every two adjacent CT slices of the CT image, and then calculating average values of the absolute values of the differences of the surface areas of the liver organs of every two adjacent CT slices to serve as average differences of the CT image;
step 3.3, calculating the absolute value of the difference value between the surface area of the liver organ of each CT slice of the CT image and the surface area of the liver organ of the ultrasonic image as a comparison difference value of each CT slice;
step 3.4, comparing the comparison difference value of each CT slice of the CT image with the average difference value of the CT image, and reserving CT slices of which the comparison difference value of the CT slices in the CT image is smaller than the average difference value of the CT image, thereby obtaining coarse registered CT slices;
step 4, performing fine registration on the roughly registered CT slices of the CT image obtained in the step 3 and the ultrasonic image to obtain fine registered CT slices, namely;
step 4.1, performing image space transformation on each of the roughly registered CT slices by taking the mean square error of the binary matrixes of the CT slice images and the ultrasonic images as an objective function so as to transform each of the roughly registered CT slices into an ultrasonic space coordinate system, thereby obtaining the roughly registered CT slices after space transformation;
and 4.2, calculating the overlapping rate of the spatially transformed coarse registration CT slices and the ultrasonic image by using dice coefficients, and taking the spatially transformed coarse registration CT slices with the highest overlapping rate as the fine registration CT slices.
2. The feature-based CT image and ultrasound image liver registration method of claim 1, wherein the specific procedure of step 1 is as follows:
step 1.1, carrying out window adjustment on each CT slice;
step 1.2, resampling each CT slice subjected to window adjustment by using a bilinear interpolation method;
and 1.3, inputting each resampled CT slice into a UNet convolutional neural network to obtain a binary mask of the liver organ, and extracting the outline features of the liver organ of each CT slice.
3. The feature-based CT image and ultrasound image liver registration method of claim 1, wherein the specific procedure of step 2 is as follows:
step 2.1, window adjustment is carried out on the ultrasonic image;
step 2.2, resampling the ultrasonic image subjected to window adjustment by using a bilinear interpolation method;
and 2.3, obtaining a binary mask, namely a contour feature of the liver organ by adopting a pixel gradient segmentation method for the resampled ultrasonic image, thereby extracting the contour feature of the liver organ of the ultrasonic image.
4. A feature-based CT image and ultrasound image liver registration method as claimed in claim 3, wherein the specific procedure of step 2.3 is as follows: denoising the resampled ultrasonic image by using twice median filtering; extracting outline edge characteristics of the liver organs by adopting a Canny edge detection algorithm; and binarizing the outline edge image, and filling the closed region to obtain a binary mask of the liver organ of the ultrasonic image.
5. The feature-based CT image and ultrasound image liver registration method according to claim 1, wherein in step 4.2, the calculation formula of the overlap ratio Dice is:
wherein I is US Is a binary matrix of ultrasonic images, I' CT Is a binary matrix of spatially transformed coarsely registered CT slices.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310291088.6A CN116363181A (en) | 2023-03-23 | 2023-03-23 | Feature-based CT image and ultrasonic image liver registration method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310291088.6A CN116363181A (en) | 2023-03-23 | 2023-03-23 | Feature-based CT image and ultrasonic image liver registration method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN116363181A true CN116363181A (en) | 2023-06-30 |
Family
ID=86941210
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202310291088.6A Pending CN116363181A (en) | 2023-03-23 | 2023-03-23 | Feature-based CT image and ultrasonic image liver registration method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN116363181A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116580033A (en) * | 2023-07-14 | 2023-08-11 | 卡本(深圳)医疗器械有限公司 | Multi-mode medical image registration method based on image block similarity matching |
CN118037794A (en) * | 2024-04-15 | 2024-05-14 | 卡本(深圳)医疗器械有限公司 | Intelligent registration system for multi-mode multi-body-position medical images |
-
2023
- 2023-03-23 CN CN202310291088.6A patent/CN116363181A/en active Pending
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116580033A (en) * | 2023-07-14 | 2023-08-11 | 卡本(深圳)医疗器械有限公司 | Multi-mode medical image registration method based on image block similarity matching |
CN116580033B (en) * | 2023-07-14 | 2023-10-31 | 卡本(深圳)医疗器械有限公司 | Multi-mode medical image registration method based on image block similarity matching |
CN118037794A (en) * | 2024-04-15 | 2024-05-14 | 卡本(深圳)医疗器械有限公司 | Intelligent registration system for multi-mode multi-body-position medical images |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106934821B (en) | Conical beam CT and CT image registration method based on ICP algorithm and B spline | |
WO2022052303A1 (en) | Method, apparatus and device for registering ultrasound image and ct image | |
JP5584441B2 (en) | Association method between volume data | |
CN114119549B (en) | Multi-mode medical image three-dimensional point cloud registration optimization method | |
Wein et al. | Automatic bone detection and soft tissue aware ultrasound–CT registration for computer-aided orthopedic surgery | |
CN112509022A (en) | Non-calibration object registration method for preoperative three-dimensional image and intraoperative perspective image | |
CN112215844A (en) | MRI (magnetic resonance imaging) multi-mode image segmentation method and system based on ACU-Net | |
CN116580068B (en) | Multi-mode medical registration method based on point cloud registration | |
CN106780491B (en) | Initial contour generation method adopted in segmentation of CT pelvic image by GVF method | |
CN116363181A (en) | Feature-based CT image and ultrasonic image liver registration method | |
CN115830016A (en) | Medical image registration model training method and equipment | |
Maekawa et al. | Model-based registration for pneumothorax deformation analysis using intraoperative cone-beam CT images | |
CN115578320A (en) | Full-automatic space registration method and system for orthopedic surgery robot | |
Heitz et al. | Statistical shape model generation using nonrigid deformation of a template mesh | |
CN109816665B (en) | Rapid segmentation method and device for optical coherence tomography image | |
EP1956552B1 (en) | Visual enhancement of interval changes using a temporal subtraction technique | |
Hsu | A hybrid approach for brain image registration with local constraints | |
Urschler et al. | Assessing breathing motion by shape matching of lung and diaphragm surfaces | |
WO2021213053A1 (en) | System and method for estimating motion of target inside tissue on basis of soft tissue surface deformation | |
Erdt et al. | Computer aided segmentation of kidneys using locally shape constrained deformable models on CT images | |
CN111166373B (en) | Positioning registration method, device and system | |
CN114862799A (en) | Full-automatic brain volume segmentation algorithm for FLAIR-MRI sequence | |
Jia et al. | Active contour model with shape constraints for bone fracture detection | |
CN110335348B (en) | Pelvic and femoral medical image modeling method based on RFR-SSMs | |
Roy et al. | MDL-IWS: multi-view deep learning with iterative watershed for pulmonary fissure segmentation |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination |