CN111967462B - Method and device for acquiring region of interest - Google Patents
Method and device for acquiring region of interest Download PDFInfo
- Publication number
- CN111967462B CN111967462B CN202010340713.8A CN202010340713A CN111967462B CN 111967462 B CN111967462 B CN 111967462B CN 202010340713 A CN202010340713 A CN 202010340713A CN 111967462 B CN111967462 B CN 111967462B
- Authority
- CN
- China
- Prior art keywords
- dimension
- image
- frame
- medical image
- filtered
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 39
- 238000001914 filtration Methods 0.000 claims abstract description 51
- 230000011218 segmentation Effects 0.000 claims abstract description 43
- 238000009826 distribution Methods 0.000 claims description 30
- 238000013527 convolutional neural network Methods 0.000 claims description 14
- 238000003860 storage Methods 0.000 claims description 8
- 206010035664 Pneumonia Diseases 0.000 description 13
- 238000003745 diagnosis Methods 0.000 description 13
- 238000004590 computer program Methods 0.000 description 9
- 238000010586 diagram Methods 0.000 description 8
- 230000003902 lesion Effects 0.000 description 6
- 230000006870 function Effects 0.000 description 5
- 210000004072 lung Anatomy 0.000 description 4
- 238000012986 modification Methods 0.000 description 4
- 230000004048 modification Effects 0.000 description 4
- 230000009286 beneficial effect Effects 0.000 description 3
- 238000003062 neural network model Methods 0.000 description 3
- 241000711573 Coronaviridae Species 0.000 description 2
- 230000004075 alteration Effects 0.000 description 2
- 238000005520 cutting process Methods 0.000 description 2
- 239000005337 ground glass Substances 0.000 description 2
- 238000003384 imaging method Methods 0.000 description 2
- 210000004556 brain Anatomy 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000001926 lymphatic effect Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 210000000056 organ Anatomy 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
-
- 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
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/03—Recognition of patterns in medical or anatomical images
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Multimedia (AREA)
- Apparatus For Radiation Diagnosis (AREA)
- Magnetic Resonance Imaging Apparatus (AREA)
Abstract
The invention mainly discloses a method and a device for acquiring a region of interest. The method comprises the following steps: dividing the three-dimensional medical image by using a first dimension dividing plane to obtain a plurality of frames of first medical images in a first dimension; determining a region of interest in each frame of the first medical image in the first dimension to obtain a plurality of frames of the first image in the first dimension; taking each frame of first medical image in the first dimension as a guide image, and filtering each frame of first image in the first dimension corresponding to the first medical image to obtain multi-frame filtered first images in the first dimension; obtaining a region of interest based on the multi-frame filtered first image in the first dimension and the multi-frame nth medical image in the nth dimension; the multi-frame N-th medical image in the N dimension is obtained by segmenting the three-dimensional medical image through an N-th segmentation plane, and N is a natural number greater than or equal to 2. By adopting the scheme provided by the invention, the accuracy of the obtained region of interest is improved.
Description
Technical Field
The present invention relates to the field of medical technology, and in particular, to a method and apparatus for acquiring a region of interest, a computer device, and a computer readable storage medium.
Background
In the field of medical technology, it is often involved to extract a region of interest from a medical image, where the region of interest may be a focal region, an organ region, a lymphatic region, etc. As for the lesion area, specifically, a pneumonia lesion area (solid lesion and ground glass lesion) or the like may be mentioned. In practical applications, accurate extraction or delineating of the region of interest is critical for subsequent diagnosis. Such as: for patients infected with pneumonia, a plurality of pneumonia areas are displayed on a lung CT image, the pneumonia areas are accurately outlined, and further, the proportion of the pneumonia areas to the whole lung is determined according to the outlined pneumonia areas, so that an imaging doctor can be helped to quickly judge whether the patients are infected with novel coronaviruses.
Therefore, how to accurately determine the region of interest, thereby facilitating the diagnosis of doctors and improving the diagnosis efficiency and the diagnosis accuracy becomes one of the problems to be solved in the present day.
Disclosure of Invention
The invention provides a method, a device, computer equipment and a computer readable storage medium for acquiring an area of interest, which are used for accurately determining the area of interest, are beneficial to diagnosis of doctors and greatly improve the diagnosis efficiency and the diagnosis accuracy of the doctors.
The invention provides a method for acquiring a region of interest, which comprises the following steps:
dividing the three-dimensional medical image by using a first dimension dividing plane to obtain a plurality of frames of first medical images in a first dimension;
determining a region of interest in each frame of the first medical image in the first dimension to obtain a plurality of frames of the first image in the first dimension;
taking each frame of first medical image in the first dimension as a guide image, and filtering each frame of first image in the first dimension corresponding to the first medical image to obtain multi-frame filtered first images in the first dimension;
obtaining a region of interest based on the multi-frame filtered first image in the first dimension and the multi-frame nth medical image in the nth dimension; the multi-frame N-th medical image in the N dimension is obtained by segmenting the three-dimensional medical image through an N-th segmentation plane, and N is a natural number greater than or equal to 2.
Optionally, the obtaining the region of interest based on the multi-frame filtered first image in the first dimension and the multi-frame nth medical image in the nth dimension includes:
dividing the three-dimensional image formed by the multi-frame filtered first image in the first dimension by using a second dimension dividing plane to obtain a multi-frame second image in the second dimension;
Taking each frame of second medical image in the second dimension as a guide image, and filtering each frame of second image in the second dimension corresponding to the guide image to obtain multi-frame filtered second images in the second dimension;
and obtaining the region of interest based on the multi-frame filtered second image in the second dimension.
Optionally, the obtaining the region of interest based on the multi-frame filtered second image at the second dimension includes:
dividing the three-dimensional image formed by the multi-frame filtered second image in the second dimension by using an M-th dimension dividing plane to obtain a multi-frame M-th image in the M-th dimension;
the region of interest is contained in a multi-frame mth image in the mth dimension;
wherein the M-th dimension splitting plane is not parallel to the second dimension splitting plane, and M is a natural number greater than or equal to 1.
Optionally, the obtaining the region of interest based on the multi-frame filtered first image in the first dimension and the multi-frame nth medical image in the nth dimension includes:
dividing the three-dimensional image formed by the multi-frame filtered first image in the first dimension by using a second dimension dividing plane to obtain a multi-frame second image in the second dimension;
Taking each frame of second medical image in the second dimension as a guide image, and filtering each frame of second image in the second dimension corresponding to the guide image to obtain multi-frame filtered second images in the second dimension;
dividing the three-dimensional image formed by the multi-frame filtered second image in the second dimension by using a third dimension dividing plane to obtain a multi-frame third image in the third dimension;
taking each frame of third medical image in the third dimension as a guide image, and filtering each frame of third image in the third dimension corresponding to the guide image to obtain multi-frame filtered third images in the third dimension;
and obtaining the region of interest based on the multi-frame filtered third image in the third dimension.
Optionally, the obtaining the region of interest based on the multi-frame filtered third image under the third dimension includes:
dividing the three-dimensional image formed by the multi-frame filtered third image under the third dimension by using an M-th dimension dividing plane to obtain a multi-frame M-th image under the M-th dimension;
the region of interest is contained in a multi-frame mth image in the mth dimension;
wherein the M-th dimension splitting plane is not parallel to the third dimension splitting plane, and M is a natural number greater than or equal to 1.
Optionally, the first dimension splitting plane, the second dimension splitting plane and the third dimension splitting plane are not parallel.
Optionally, the filtering each frame of the first medical image in the first dimension with each frame of the first medical image in the first dimension as the guiding image to obtain a multi-frame filtered first image in the first dimension includes: obtaining a frame of filtered first image by the following formula:
wherein P is a first image, I is a first medical image, q is a filtered first image, I and j respectively represent pixel subscripts, and Wij is a filter kernel associated with the first medical image I.
Optionally, the determining the region of interest in each frame of the first medical image in the first dimension to obtain a plurality of frames of the first image in the first dimension includes:
inputting a first medical image layer formed by continuous multi-frame first medical images into a preset convolutional neural network to output a confidence distribution map corresponding to each frame of first medical image;
and determining the region of interest from the first medical images of each frame according to the confidence distribution map corresponding to the first medical images of each frame.
The invention also provides a device for acquiring the region of interest, which comprises:
The segmentation unit is used for segmenting the three-dimensional medical image by using a first-dimension segmentation plane to obtain a multi-frame first medical image in a first dimension;
a determining unit, configured to determine a region of interest in each frame of the first medical image in the first dimension to obtain a plurality of frames of the first image in the first dimension;
the filtering unit is used for filtering each frame of first medical image in the first dimension by taking each frame of first medical image in the first dimension as a guide image so as to obtain multi-frame filtered first images in the first dimension;
the acquisition unit is used for acquiring an interested region based on the multi-frame filtered first image in the first dimension and the multi-frame N medical image in the N dimension; the multi-frame N-th medical image in the N dimension is obtained by segmenting the three-dimensional medical image through an N-th segmentation plane, and N is a natural number greater than or equal to 2.
The invention also provides a computer device comprising at least one processor and at least one memory, wherein the memory stores a computer program which, when executed by the processor, enables the processor to perform the above-described method of acquiring a region of interest.
The invention also provides a computer readable storage medium, which when executed by a processor within a device, causes the device to perform the above-described method of acquiring a region of interest.
Compared with the prior art, the technical scheme of the invention has the following beneficial effects:
firstly, segmenting the three-dimensional medical image through a first dimension segmentation plane to obtain a multi-frame first medical image in a first dimension. A region of interest in each frame of the first medical image in the first dimension is determined to obtain a plurality of frames of the first image in the first dimension. And then, taking each frame of first medical image in the first dimension as a guide image, and filtering each frame of first image in the first dimension corresponding to the first medical image to obtain multi-frame filtered first images in the first dimension. Finally, obtaining a region of interest based on the multi-frame filtered first image in the first dimension and the multi-frame nth medical image in the nth dimension; the multi-frame N-th medical image in the N dimension is obtained by segmenting the three-dimensional medical image through an N-th segmentation plane, and N is a natural number greater than or equal to 2. Because each frame of first medical image in the first dimension is taken as a guiding image to filter each frame of first image in the first dimension so as to obtain a plurality of frames of filtered first images in the first dimension, the obtained each frame of filtered first image in the first dimension is similar to the corresponding first image in the first dimension, the texture of the obtained each frame of filtered first image in the first dimension is similar to the texture of the corresponding each frame of medical image in the first dimension, and furthermore, when an interested region is obtained based on the plurality of frames of filtered first images in the first dimension and the plurality of frames of N images in the N dimension, the accuracy of the obtained interested region is improved, and the diagnosis efficiency and the diagnosis accuracy of doctors are improved to a great extent while the diagnosis of the doctors are facilitated.
Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. The objectives and other advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and claims thereof as well as the appended drawings.
The technical scheme of the invention is further described in detail through the drawings and the embodiments.
Drawings
The accompanying drawings are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate the invention and together with the embodiments of the invention, serve to explain the invention. In the drawings:
FIG. 1 is a flowchart of a method for acquiring a region of interest according to a first embodiment of the present invention;
FIG. 2 is a schematic view of a first medical image in a first dimension according to a first embodiment of the present invention;
FIG. 3 is a schematic view of a first image according to a first embodiment of the present invention;
FIG. 4 is a view of an image of interest according to a first embodiment of the present invention;
FIG. 5 is a cross-sectional image of a segmented region of interest according to a first embodiment of the present invention;
FIG. 6 is a flowchart of a method for acquiring a region of interest according to a second embodiment of the present invention;
Fig. 7 is a schematic structural diagram of an apparatus for acquiring a region of interest according to the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be described in further detail below with reference to the accompanying drawings, and it is apparent that the described embodiments are only some embodiments of the present invention, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
The method for acquiring the region of interest comprises the following steps:
dividing the three-dimensional medical image by using a first dimension dividing plane to obtain a plurality of frames of first medical images in a first dimension;
determining a region of interest in each frame of the first medical image in the first dimension to obtain a plurality of frames of the first image in the first dimension;
taking each frame of first medical image in the first dimension as a guide image, and filtering each frame of first image in the first dimension corresponding to the first medical image to obtain multi-frame filtered first images in the first dimension;
obtaining a region of interest based on the multi-frame filtered first image in the first dimension and the multi-frame nth medical image in the nth dimension; the multi-frame N-th medical image in the N dimension is obtained by segmenting the three-dimensional medical image through an N-th segmentation plane, and N is a natural number greater than or equal to 2.
According to the technical scheme, the three-dimensional medical image can be segmented through segmentation planes with different dimensions, such as the segmentation planes from the first dimension to the N dimension. The first dimension may be a plane parallel to the transverse plane, a plane parallel to the coronal plane, or a plane parallel to the sagittal plane. In the invention, the three-dimensional medical image can be a three-dimensional chest medical image, a three-dimensional brain medical image and the like, and the region of interest can be a focus region, a region to be diagnosed, a region to be detected and the like. Taking a three-dimensional medical image as an example of a three-dimensional chest image, if a doctor needs to view a pneumonia area in the lung, the area of interest may be a real lesion area or a ground glass lesion area. In practical applications, the region of interest is determined by the actual needs of the physician.
The technical scheme of the present invention will be described in detail below with the three-dimensional medical image as a three-dimensional chest medical image and the region of interest as a pneumonia region.
Example 1
Fig. 1 is a flowchart of a method for acquiring a region of interest according to a first embodiment of the present invention, as shown in fig. 1, the method for acquiring a region of interest includes the following steps S11-S16:
S11: and cutting the three-dimensional medical image by using the first-dimension cutting plane to obtain a multi-frame first medical image in the first dimension.
S12: a region of interest in each frame of the first medical image in the first dimension is determined to obtain a plurality of frames of the first image in the first dimension.
S13: and taking each frame of first medical image in the first dimension as a guide image, and filtering each frame of first image in the first dimension corresponding to the guide image to obtain multi-frame filtered first images in the first dimension.
S14: and splitting the three-dimensional image formed by the multi-frame filtered first image in the first dimension by using a second dimension splitting plane to obtain a multi-frame second image in the second dimension.
S15: and taking each frame of second medical image in the second dimension as a guide image, and filtering each frame of second image in the second dimension corresponding to the guide image to obtain multi-frame filtered second images in the second dimension.
S16: and obtaining the region of interest based on the multi-frame filtered second image in the second dimension.
In this embodiment, the first dimension dividing plane is a plane parallel to the cross section, and the second dimension dividing plane is a plane parallel to the coronal plane. In other embodiments, the first dimension splitting plane may be a plane parallel to the coronal plane or the sagittal plane, or may be a splitting plane at an angle to the human body, and the second dimension splitting plane may be a plane parallel to the transverse plane or the sagittal plane, or may be a splitting plane at an angle to the human body, so that the first dimension splitting plane is a plane parallel to the transverse plane, and the second dimension splitting plane is a plane parallel to the coronal plane, which should not be taken as a limitation of the technical solution of the present invention.
And S11, segmenting the three-dimensional chest medical image with a first-dimension segmentation plane parallel to the cross section to obtain a plurality of frames of first medical images in the first dimension. Referring to fig. 3, fig. 3 is a schematic view of a first medical image in a first dimension, i.e. a cross-section of a three-dimensional chest medical image, according to an embodiment of the present invention.
And S12, determining a region of interest in each frame of the first medical image in the first dimension to obtain a plurality of frames of the first image in the first dimension, namely, dividing the region of interest from each frame of the first medical image. In this embodiment, the pneumonia area is divided from the cross section of each frame. Specifically, the region of interest may be segmented by a neural network model, or may be segmented from the first medical image by a thresholding method, a region growing method, a mode classification-based method, or the like.
In this embodiment, the region of interest is segmented from the first medical image (cross-sectional image) of each frame by a predetermined convolutional neural network model. In particular, the method comprises the steps of,
first, for each frame of the first medical image (each frame of the cross-sectional image), the first medical image layer corresponding to the first medical image layer is formed by the first medical image layer and one or more continuous frames of the first medical image layer located in front of and behind the first medical image layer. For example, for the fourth frame of the first medical image (the fourth frame of the cross-sectional image), the first medical image layer corresponding to the fourth frame of the first medical image (the fourth frame of the cross-sectional image) may be formed by taking the first frame of the first medical image (the cross-sectional image) and the three preceding and following three frames of the first medical image (the cross-sectional image), that is, the first medical image layer corresponding to the fourth frame of the first medical image (the fourth frame of the cross-sectional image) includes the first frame of the first medical image (the first frame of the cross-sectional image), the second frame of the first medical image (the second frame of the cross-sectional image), the third frame of the first medical image (the third frame of the cross-sectional image), the fourth frame of the first medical image (the fourth frame of the cross-sectional image), the fifth frame of the first medical image (the fifth frame of the cross-sectional image), the sixth frame of the first medical image (the sixth frame of the cross-sectional image) and the seventh frame of the first medical image (the seventh frame of the cross-sectional image). Similarly, for the first medical image layer corresponding to the fifth first medical image (fifth cross-sectional image), the second first medical image (second cross-sectional image), the third first medical image (third cross-sectional image), the fourth first medical image (fourth cross-sectional image), the fifth first medical image (fifth cross-sectional image), the sixth first medical image (sixth cross-sectional image), the seventh first medical image (seventh cross-sectional image), and the eighth first medical image (eighth cross-sectional image) are included. If the three-dimensional medical image is segmented by the first dimension segmentation plane to obtain a hundred frames of first medical images, a first medical image layer corresponding to a fourth frame of first medical image (fourth frame of cross-sectional image), a first medical image layer corresponding to a fifth frame of first medical image (fifth frame of cross-sectional image), a first medical image layer corresponding to a sixth frame of first medical image (sixth frame of cross-sectional image), … …, and a first medical image layer corresponding to a nineteenth seven frames of first medical image (seventeenth frame of cross-sectional image) (including a nineteenth fourteenth frame of first medical image, a nineteenth frame of first medical image, and a first hundred frames of first medical image) may be obtained.
And then, respectively inputting each first medical image layer into a preset convolutional neural network model to obtain a confidence distribution map of the first medical image corresponding to each first medical image layer. Specifically, taking the example that the first medical image layer includes seven frames of first medical images, presetting a convolutional neural network model to output seven frames of confidence distribution graphs corresponding to the seven frames of first medical images, and taking the confidence distribution graph of the middle frame in the seven frames of confidence distribution graphs as the confidence distribution graph of the first medical image corresponding to the first medical image layer. Such as: if the first medical image layer includes: and taking the confidence coefficient distribution map of the middle frame in the seven frames of confidence coefficient distribution maps output by the preset convolutional neural network model as the confidence coefficient distribution map of the fourth frame of first medical image (the fourth frame of cross-section image).
In this embodiment, the image layer is input to the preset convolutional neural network model, so that the confidence coefficient distribution of the first medical image of the nth frame is combined with the confidence coefficient information of the first three frames and the last three frames, and the accuracy of dividing the medical image of each frame can be improved.
For a certain frame of first medical image without the first three frames and the last three frames which can be referred, for example, the first frame of first medical image, the second frame of first medical image and the third frame of first medical image, the confidence distribution map of the first three frames output when the first medical image layer corresponding to the fourth frame of first medical image is input into the preset neural network model can be directly used as the confidence distribution map of the first frame of first medical image, the confidence distribution map of the second frame of first medical image and the confidence distribution map of the third frame of first medical image respectively. The multi-frame first medical image is formed into the first medical image layer and then is input into the preset convolutional neural network to obtain the confidence distribution map of each frame of the first medical image, so that the segmentation accuracy is improved to a great extent.
It should be noted that, the foregoing description is given by taking the first three frames including the first medical image of the M-th frame and the first medical image of the last three frames as an example, in practical application, the first medical image layer may also be the first medical image of the previous frame and the last frame including the first medical image of the M-th frame (three frames in total), or may also be the first medical image of the previous frame and the last frame including the first medical image of the M-th frame (five frames in total), and the first medical image layer specifically includes several frames of the first medical images, which are determined by factors such as segmentation precision, processing speed, and the like.
And finally, determining the region of interest from each frame of the first medical image based on the confidence distribution map of each frame of the first medical image. Specifically, according to the confidence distribution map of each frame of the first medical image, determining a region formed by one or more pixel points with confidence greater than a preset threshold value in each frame of the first medical image as a region of interest in each frame of the first medical image.
In other embodiments, the region of interest in the first medical image of each frame may be segmented without inputting the image layer into a predetermined convolutional neural network model. If the confidence distribution map of the first medical image of each frame can be directly input into the preset convolutional neural network model, the confidence distribution map of the first medical image of the frame can be directly output. Further, in order to increase the segmentation speed, only the first medical images of partial frames of the first medical images of the plurality of frames may be input into the preset convolutional neural network model, and the confidence distribution map of the first medical images of the rest frames may be determined through the confidence distribution map of the first medical images of the partial frames. For example, if the three-dimensional medical image is segmented by the first dimension segmentation plane to obtain a hundred frames of first medical images, the first medical images with odd frame numbers can be input into the preset convolutional neural network model, the confidence distribution map of the first medical images with odd frame numbers is output, and the confidence distribution map of the first medical images with even frame numbers can be obtained by the confidence distribution map of the first medical images with odd frame numbers. For example, the confidence coefficient distribution map of the first medical image of the second frame may be obtained by the confidence coefficient distribution maps of the first medical image of the first frame and the first medical image of the third frame, specifically, the confidence coefficient of each pixel point in the first medical image of the first frame and the confidence coefficient of each pixel point in the first medical image of the third frame are weighted and averaged, and the weighted and averaged confidence coefficient is used as the confidence coefficient of the corresponding pixel point in the first medical image of the second frame, so as to obtain the confidence coefficient distribution map of the first medical image of the second frame.
In this embodiment, the preset convolutional neural network model may be a U-NET neural network model using VGG as a backup, or may be a full convolutional neural network model.
And S13, taking each frame of first medical image in the first dimension as a guide image, and filtering each frame of first image in the first dimension corresponding to the first medical image to obtain multi-frame filtered first images in the first dimension. In this embodiment, each frame of cross-sectional image is used as a guiding image, and the image of the region of interest segmented on the cross-sectional image is filtered. Specifically, the first image of each frame in the first dimension may be filtered by the following formula:
q i ＝∑ j W ij (I) Pj formula (1)
Wherein P is a first image, I is a first medical image, q is a filtered first image, I and j respectively represent pixel subscripts, W ij Is a filter kernel associated with the first medical image I.
For filtering the corresponding first image P by the guiding image first medical image I, it is assumed that the filtered output first image (filtered first image q) is within the filtering window ω k There is a local linear relationship: i.e.
q i ＝a k I i +b k
For a defined window omega with radius i k ，(a k ，b k ) Will also be a uniquely determined constant coefficient. This ensures that in a local area, if the guiding image I (first medical image) has an edge, the output image q (filtered first image) also remains edge-wise, for adjacent pixelsTherefore, the output q is also obtained by solving the coefficients a and b. Meanwhile, it is considered that the non-edge region of the input image (first image P) is not smooth and is noise n, q is the case i ＝p i -n i Minimizing this noise, for each filter window, can be expressed as follows:
introducing a regularization parameter E to avoid a k Oversized, resulting in a loss function within the filter window
Respectively to a k And b k And (5) obtaining a deviation guide, and finally obtaining:
wherein mu k Andrespectively, the guiding image I (first medical image) is shown in window omega k The mean and variance of (ω) is the window ω k The number of middle pixels, < >>Is that the input image p (first image) is in window omega k Is a mean value of (b). Will a k And b k Substitution of q i ＝a k I i +b k The gray value of each pixel in the output image can be obtained.
The gray value of each pixel point in the filtered first image can be obtained through the formula, namely the filtered first image. And filtering each frame of first image under the first dimension through the formula so as to obtain multi-frame filtered first images.
S14 is executed: and splitting the three-dimensional image formed by the multi-frame filtered first image in the first dimension by using a second dimension splitting plane to obtain a multi-frame second image in the second dimension. In this embodiment, specifically, the three-dimensional image formed by the multi-frame filtered first image may be segmented through a plane parallel to the coronal plane, that is, the multi-frame filtered cross-sectional image obtained by dividing the multi-frame into the region of interest through the guide is segmented into the three-dimensional image, and then the multi-frame coronal plane image is obtained.
S15 is executed: and taking each frame of second medical image in the second dimension as a guide image, and filtering each frame of second image in the second dimension corresponding to the guide image to obtain multi-frame filtered second images in the second dimension. In this step, each frame of the second medical image in the second dimension may be obtained by slicing the three-dimensional medical image in a plane parallel to the coronal plane. Specifically, each frame of coronal image is taken as a guide image, and the multi-frame coronal image is obtained after the multi-frame cut after the guide filtering in the S14 is divided into the three-dimensional image composed of the cross-sectional images of the region of interest, and the guide filtering is performed. The procedure of guided filtering is similar to S13, see equation (1), except that the guided image I in this equation is the second medical science The image P is the second image, q is the filtered second image, W ij Is a filter kernel associated with the second medical image I. In addition, a k And b k The solution of (a) can be seen in the formula (2) and the formula (3), as long as the guiding image I in the formula (2) and the formula (3) is replaced by the second medical image, and the input image p is replaced by the second image, which is not described herein. Thus, the multi-frame filtered second image under the second dimension can be obtained through the formula.
And S16, obtaining the region of interest based on the multi-frame filtered second image in the second dimension. Generally speaking, in order to make a region of interest as clear as possible, a doctor may need to select different viewing angles or different dimensions to observe the region of interest, so in practical applications, after obtaining a multi-frame filtered second image at a second dimension, a final region of interest needs to be obtained according to practical requirements. In this embodiment, the three-dimensional image formed by the multi-frame filtered second image in the second dimension may be segmented by using an mth dimension segmentation plane, so as to obtain a multi-frame mth image in the mth dimension. The region of interest is included in the mth image of the plurality of frames in the mth dimension. Wherein the M-th dimension splitting plane is not parallel to the second dimension splitting plane, and M is a natural number greater than or equal to 1.
Specifically, as can be seen from the foregoing, in this embodiment, the multi-frame filtered coronal image (multi-frame filtered second image) is finally obtained, if the doctor needs to observe the region of interest from the cross-sectional angle, then the three-dimensional image formed by the multi-frame filtered coronal image needs to be segmented by a plane parallel to the cross-sectional plane, so as to obtain the multi-frame cross-sectional image, that is, the M-th dimension segmentation plane is taken as 1, and is not parallel to the second dimension segmentation plane (the plane parallel to the coronal plane). Finally, the segmented region of interest is displayed in the cross-sectional image of the corresponding frame. Referring to fig. 4, fig. 4 is an image of interest according to a first embodiment of the present invention, which is a frame of cross-sectional image obtained by slicing a three-dimensional image formed by a plurality of frames of filtered images in a plane parallel to the cross-sectional image. It can be seen from fig. 4 that the region of interest in fig. 4 is smooth in edge, complete in region of interest, and the noise present in fig. 3 is removed in fig. 4, relative to fig. 3.
In other embodiments, if the doctor wants to observe the region of interest from other dimensions, a similar method as described above may be used, i.e. a slicing plane of other dimensions, such as a plane parallel to the sagittal plane, is used to slice the three-dimensional image composed of the multi-frame filtered second image in the second dimension.
It should be noted that, if the doctor selects the viewing angle for viewing to observe the region of interest from the coronal image, it is not necessary to split the three-dimensional image formed by the multi-frame filtered second image in the second dimension after obtaining the multi-frame filtered second image.
In this embodiment, the three-dimensional medical image is segmented by using the first-dimensional segmentation plane to obtain a plurality of frames of first medical images, and the region of interest is segmented from the plurality of frames of first medical images to obtain the plurality of frames of first images. And performing guided filtering on the first image by taking the first medical image as a guide image, and segmenting a three-dimensional image formed by a plurality of frames of first images obtained after the guided filtering to obtain a second image. And then, segmenting the three-dimensional medical image by adopting a second-dimensional segmentation plane to obtain a plurality of frames of second medical images, and continuing to conduct guide filtering on the second images by taking the second medical images as guide images so as to obtain a plurality of frames of second images after guide filtering. Finally, according to actual requirements, the three-dimensional image formed by the multi-frame second image after guide filtering is segmented by an N-th dimension segmentation plane, and finally the image of the region of interest is obtained. And after the first image is subjected to guide filtering in the first dimension in a guide filtering mode, the three-dimensional image formed by the multi-frame first image subjected to guide filtering is segmented in the second dimension to obtain a second image, and the second image is subjected to guide filtering in the second dimension in a guide filtering mode continuously, so that the region of interest is obtained based on the multi-frame second image subjected to guide filtering, and the accuracy of the finally obtained region of interest is improved. When the area to be segmented is the pneumonia area, the accuracy of the pneumonia area obtained by segmentation is improved, on one hand, diagnosis by doctors is facilitated, on the other hand, the diagnosis efficiency and the accuracy of diagnosis by doctors are improved to a great extent, and in addition, when the proportion of the pneumonia area to the whole lung is calculated based on the pneumonia area, the method is also beneficial to the imaging doctor to quickly and accurately judge whether the patient is infected with the novel coronavirus.
In practical application, the segmented region of interest is usually displayed on the medical image, so in this embodiment, the M-th segmentation plane is adopted to segment the three-dimensional image formed by the multi-frame filtered second image in the second dimension, and after the multi-frame M-th image in the M-th dimension is obtained, the multi-frame M-th image and the multi-frame M-th medical image in the M-th dimension obtained by segmenting the three-dimensional medical image in the M-th dimension are correspondingly superimposed and displayed. Referring to fig. 5, fig. 5 is a cross-sectional image of a region of interest segmented according to the first embodiment of the present invention, that is, after a three-dimensional image formed by a plurality of frames of filtered second images in a second dimension is segmented by a plane parallel to the cross-section, the three-dimensional image is superimposed with a cross-sectional image corresponding to the three-dimensional medical image to obtain an image, where the corresponding image refers to that if the three-dimensional image formed by the plurality of frames of filtered second images in the second dimension is segmented by a plane parallel to the cross-section by a first thickness, a hundred-frame image is obtained, and the three-dimensional medical image is segmented by a plane parallel to the cross-section by a first thickness, and the corresponding image refers to that the frame number is the same.
Example two
Fig. 6 is a flowchart of a method for acquiring a region of interest according to a second embodiment of the present invention, as shown in fig. 6, the method for acquiring a region of interest includes the following steps S21 to S28:
s21, segmenting the three-dimensional medical image by using a first-dimension segmentation plane to obtain a multi-frame first medical image in a first dimension.
And S22, determining a region of interest in each frame of the first medical image in the first dimension to obtain a plurality of frames of the first image in the first dimension.
And S23, taking each frame of first medical image in the first dimension as a guide image, and filtering each frame of first image in the first dimension corresponding to the first medical image to obtain multi-frame filtered first images in the first dimension.
And S24, segmenting the three-dimensional image formed by the multi-frame filtered first image in the first dimension by using a second dimension segmentation plane to obtain a multi-frame second image in the second dimension.
And S25, taking each frame of second medical image in the second dimension as a guide image, and filtering each frame of second image in the second dimension corresponding to the guide image to obtain multi-frame filtered second images in the second dimension.
S26, segmenting the three-dimensional image formed by the multi-frame filtered second image in the second dimension by using the third dimension segmentation plane to obtain a multi-frame third image in the third dimension.
And S27, taking each frame of third medical image in the third dimension as a guide image, and filtering each frame of third image in the third dimension corresponding to the guide image to obtain multi-frame filtered third images in the third dimension.
And S28, obtaining the region of interest based on the multi-frame filtered third image under the third dimension.
In this embodiment, S21 to S25 are similar to S11 to S15 in the first embodiment, and will not be described here again. In the present embodiment, after the second images after the multi-frame filtering in the second dimension are obtained in S25, the three-dimensional images formed by the multi-frame filtered second images in the second dimension are continuously segmented in the third dimension segmentation plane, so as to obtain the multi-frame third images in the third dimension. Specifically, the third three-dimensional slicing plane in this embodiment may be a plane parallel to the sagittal plane, so that a multi-frame sagittal plane image may be obtained.
And then, continuing to take each frame of third medical image in the third dimension as a guide image, and filtering each frame of third image in the third dimension corresponding to the guide image to obtain multi-frame filtered third images in the third dimension. How to obtain multi-frame filtered third image under third dimension Referring to the first embodiment, the difference is that the guiding image I in the formula is the third medical image, P is the third image, q is the filtered third image, W ij Is a filter kernel associated with the third medical image I. In addition, a k And b k The solution of (a) can be seen in the formula (2) and the formula (3), as long as the guiding image I in the formula (2) and the formula (3) is replaced by a third medical image, and the input image p is replaced by a third image, which is not described herein. So far, the multi-frame filtered third image under the third dimension can be obtained through the formula. In this embodiment, the first dimension splitting plane, the second dimension splitting plane, and the third dimension splitting plane are not parallel to each other. In the present embodiment, the description is given taking the example that the first dimension splitting plane is a plane parallel to the cross section, the second dimension splitting plane is a plane parallel to the coronal plane, and the third dimension splitting plane is a plane parallel to the sagittal plane, but the first dimension splitting plane is a plane parallel to the cross section, the second dimension splitting plane is a plane parallel to the coronal plane, and the third dimension splitting plane is a plane parallel to the sagittal plane, which should not be taken as a limitation of the technical scheme of the present invention, as long as the first dimension splitting plane, the second dimension splitting plane, and the third dimension splitting plane are not parallel to each other.
Finally, S28 is executed to obtain the region of interest based on the multi-frame filtered third image in the third dimension. Similarly, in order to make the region of interest as clear as possible, the doctor may need to choose a different viewing angle or a different dimension to observe the region of interest, so in practical application, after obtaining the multi-frame filtered third image in the third dimension, the final region of interest needs to be obtained according to the actual requirement. In this embodiment, the three-dimensional image formed by the multi-frame filtered third image in the third dimension may be segmented by using an mth dimension segmentation plane, so as to obtain a multi-frame mth image in the mth dimension. The region of interest is included in the mth image of the plurality of frames in the mth dimension. Wherein the M-th dimension splitting plane is not parallel to the third dimension splitting plane, and M is a natural number greater than or equal to 1.
Specifically, as can be seen from the foregoing, in this embodiment, the multi-frame filtered sagittal image (multi-frame filtered third image) is finally obtained, if the doctor needs to observe the region of interest from the cross-sectional angle, then the three-dimensional image formed by the multi-frame filtered sagittal image needs to be segmented by a plane parallel to the cross-sectional angle, so as to obtain the multi-frame cross-sectional image, that is, M takes 1, and the mth dimension segmentation plane at this time is a plane parallel to the cross-sectional plane, which is not parallel to the second dimension segmentation plane (a plane parallel to the coronal plane) and the third dimension segmentation plane (a plane parallel to the sagittal plane). Finally, the segmented region of interest is displayed in the cross-sectional image of the corresponding frame.
In other embodiments, if the doctor wants to observe the region of interest from other dimensions, a similar method as described above may be used to segment the three-dimensional image composed of the multi-frame filtered third image in the third dimension by using a segmentation plane of other dimensions, such as a plane parallel to the coronal plane.
If the doctor selects the viewing angle for viewing to view the region of interest from the sagittal image, it is unnecessary to split the three-dimensional image formed by the multi-frame filtered third image in the third dimension after obtaining the multi-frame filtered third image.
In this embodiment, on the basis of the first embodiment, the three-dimensional image formed by the multi-frame filtered second image in the second dimension is continuously segmented by using the third dimension segmentation plane, so as to obtain a multi-frame third image in the third dimension. And taking each frame of third medical image in the third dimension as a guide image, and filtering each frame of third image in the third dimension corresponding to the guide image to obtain multi-frame filtered third images in the third dimension. And finally, acquiring the region of interest based on the multi-frame filtered third image under the third dimension, thereby greatly improving the accuracy of the finally acquired region of interest.
It should be noted that, in the present embodiment, the region of interest is obtained based on the multi-frame filtered third image under the third dimension, but the technical solution of the present invention is not limited to obtaining the region of interest based on the multi-frame filtered third image under the third dimension. After the multi-frame filtered third image in the third dimension is obtained, the three-dimensional image formed by the multi-frame filtered third image in the third dimension can be further segmented by a fourth dimension segmentation plane, so as to obtain a multi-frame fourth image in the fourth dimension. And taking each frame of fourth medical image under the fourth dimension as a guide image, filtering each frame of fourth image under the fourth dimension corresponding to the fourth medical image to obtain a multi-frame filtered fourth image under the fourth dimension, and segmenting a three-dimensional image formed by the multi-frame filtered fourth image under the fourth dimension by a fifth dimension segmentation plane to obtain a multi-frame fifth image under the fifth dimension. And filtering each frame of fifth medical image in the fifth dimension by taking each frame of fifth medical image in the fifth dimension as a guide image, so as to obtain a multi-frame filtered fifth image in the fifth dimension, … …, until each frame of N medical image in the N dimension is taken as a guide image, filtering each frame of N image in the N dimension, so as to obtain a multi-frame filtered N image in the N dimension, and finally obtaining an interested region based on the multi-frame filtered N image in the N dimension. The amount of N is determined according to actual requirements, such as the accuracy of the finally obtained region of interest, the speed of obtaining the region of interest and the like.
The present invention also provides a device for acquiring a region of interest, referring to fig. 7, fig. 7 is a schematic structural diagram of the device for acquiring a region of interest according to the present invention, as shown in fig. 7, where the device for acquiring a region of interest includes:
the segmentation unit 101 is configured to segment the three-dimensional medical image with a first dimension segmentation plane to obtain a multi-frame first medical image in the first dimension;
a determining unit 102, configured to determine a region of interest in each frame of the first medical image in the first dimension to obtain a plurality of frames of the first image in the first dimension;
a filtering unit 103, configured to filter each frame of the first medical image in the first dimension with each frame of the first medical image in the first dimension as a guide image, so as to obtain a plurality of frames of filtered first images in the first dimension;
an obtaining unit 104, configured to obtain a region of interest based on the multi-frame filtered first image in the first dimension and the multi-frame nth medical image in the nth dimension; the multi-frame N-th medical image in the N dimension is obtained by segmenting the three-dimensional medical image through an N-th segmentation plane, and N is a natural number greater than or equal to 2.
The implementation of the apparatus for acquiring a region of interest in this embodiment may refer to the implementation of the method for acquiring a region of interest described above, and will not be described herein.
Based on the same technical idea, an embodiment of the present invention provides a computer device, including at least one processor, and at least one memory, where the memory stores a computer program, and when the program is executed by the processor, enables the processor to perform the above-mentioned method for acquiring a region of interest.
Based on the same technical idea, an embodiment of the present invention provides a computer-readable storage medium, which when executed by a processor within an apparatus, enables the apparatus to perform the above-described method of acquiring a region of interest.
It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as a method, or as a computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
While preferred embodiments of the present invention have been described, additional variations and modifications in those embodiments may occur to those skilled in the art once they learn of the basic inventive concepts. It is therefore intended that the following claims be interpreted as including the preferred embodiments and all such alterations and modifications as fall within the scope of the invention.
It will be apparent to those skilled in the art that various modifications and variations can be made to the present invention without departing from the spirit or scope of the invention. Thus, it is intended that the present invention also include such modifications and alterations insofar as they come within the scope of the appended claims or the equivalents thereof.
Claims (8)
1. A method of acquiring a region of interest, comprising:
dividing the three-dimensional medical image by using a first dimension dividing plane to obtain a plurality of frames of first medical images in a first dimension;
determining a region of interest in each frame of the first medical image in the first dimension to obtain a plurality of frames of the first image in the first dimension;
taking each frame of first medical image in the first dimension as a guide image, and filtering each frame of first image in the first dimension corresponding to the first medical image to obtain multi-frame filtered first images in the first dimension;
obtaining a region of interest based on the multi-frame filtered first image in the first dimension and the multi-frame nth medical image in the nth dimension; the method comprises the steps that a plurality of frames of N medical images in the N dimension are obtained by segmenting the three-dimensional medical images through an N dimension segmentation plane, and N is a natural number greater than or equal to 2;
Wherein:
the obtaining the region of interest based on the multi-frame filtered first image in the first dimension and the multi-frame nth medical image in the nth dimension includes:
dividing the three-dimensional image formed by the multi-frame filtered first image in the first dimension by using a second dimension dividing plane to obtain a multi-frame second image in the second dimension;
taking each frame of second medical image in the second dimension as a guide image, and filtering each frame of second image in the second dimension corresponding to the guide image to obtain multi-frame filtered second images in the second dimension;
obtaining a region of interest based on the multi-frame filtered second image at the second dimension;
or alternatively
The obtaining the region of interest based on the multi-frame filtered first image in the first dimension and the multi-frame nth medical image in the nth dimension includes:
dividing the three-dimensional image formed by the multi-frame filtered first image in the first dimension by using a second dimension dividing plane to obtain a multi-frame second image in the second dimension;
taking each frame of second medical image in the second dimension as a guide image, and filtering each frame of second image in the second dimension corresponding to the guide image to obtain multi-frame filtered second images in the second dimension;
Dividing the three-dimensional image formed by the multi-frame filtered second image in the second dimension by using a third dimension dividing plane to obtain a multi-frame third image in the third dimension;
taking each frame of third medical image in the third dimension as a guide image, and filtering each frame of third image in the third dimension corresponding to the guide image to obtain multi-frame filtered third images in the third dimension;
and obtaining the region of interest based on the multi-frame filtered third image in the third dimension.
2. The method of claim 1, wherein the obtaining the region of interest based on the multi-frame filtered second image at the second dimension comprises:
dividing the three-dimensional image formed by the multi-frame filtered second image in the second dimension by using an M-th dimension dividing plane to obtain a multi-frame M-th image in the M-th dimension;
the region of interest is contained in a multi-frame mth image in the mth dimension;
wherein the M-th dimension splitting plane is not parallel to the second dimension splitting plane, and M is a natural number greater than or equal to 1.
3. The method of claim 1, wherein the obtaining the region of interest based on the multi-frame filtered third image at the third dimension comprises:
Dividing the three-dimensional image formed by the multi-frame filtered third image under the third dimension by using an M-th dimension dividing plane to obtain a multi-frame M-th image under the M-th dimension;
the region of interest is contained in a multi-frame mth image in the mth dimension;
wherein the M-th dimension splitting plane is not parallel to the third dimension splitting plane, and M is a natural number greater than or equal to 1.
4. The method of claim 1, wherein the first dimension splitting plane, the second dimension splitting plane, and the third dimension splitting plane are non-parallel.
5. The method of claim 1, wherein,
the filtering each frame of the first medical image in the first dimension corresponding to the first medical image in the first dimension to obtain a plurality of frames of filtered first images in the first dimension includes: obtaining a frame of filtered first image by the following formula:
wherein P is a first image, I is a first medical image, q is a filtered first image, I and j respectively represent pixel subscripts, W ij Is a filter kernel associated with the first medical image I.
6. The method of claim 1, wherein determining the region of interest in each frame of the first medical image in the first dimension to obtain a plurality of frames of the first image in the first dimension comprises:
For each frame of the first medical image, forming a first medical image layer corresponding to the first medical image layer by the first medical image layer and one or more continuous frames of the first medical image before and after the first medical image layer;
respectively inputting each first medical image layer into a preset convolutional neural network model to obtain a confidence distribution map of a first medical image corresponding to each first medical image layer;
and determining a region of interest from each frame of the first medical image based on the confidence distribution map of each frame of the first medical image.
7. An apparatus for acquiring a region of interest, comprising:
the segmentation unit is used for segmenting the three-dimensional medical image by using a first-dimension segmentation plane to obtain a multi-frame first medical image in a first dimension;
a determining unit, configured to determine a region of interest in each frame of the first medical image in the first dimension to obtain a plurality of frames of the first image in the first dimension;
the filtering unit is used for filtering each frame of first medical image in the first dimension by taking each frame of first medical image in the first dimension as a guide image so as to obtain multi-frame filtered first images in the first dimension;
the acquisition unit is used for acquiring an interested region based on the multi-frame filtered first image in the first dimension and the multi-frame N medical image in the N dimension; the method comprises the steps that a plurality of frames of N medical images in the N dimension are obtained by segmenting the three-dimensional medical images through an N dimension segmentation plane, and N is a natural number greater than or equal to 2;
The acquisition unit obtains the region of interest based on the multi-frame filtered first image in the first dimension and the multi-frame nth medical image in the nth dimension by:
dividing the three-dimensional image formed by the multi-frame filtered first image in the first dimension by using a second dimension dividing plane to obtain a multi-frame second image in the second dimension;
taking each frame of second medical image in the second dimension as a guide image, and filtering each frame of second image in the second dimension corresponding to the guide image to obtain multi-frame filtered second images in the second dimension;
obtaining a region of interest based on the multi-frame filtered second image at the second dimension;
or alternatively
The acquisition unit obtains the region of interest based on the multi-frame filtered first image in the first dimension and the multi-frame nth medical image in the nth dimension by:
dividing the three-dimensional image formed by the multi-frame filtered first image in the first dimension by using a second dimension dividing plane to obtain a multi-frame second image in the second dimension;
taking each frame of second medical image in the second dimension as a guide image, and filtering each frame of second image in the second dimension corresponding to the guide image to obtain multi-frame filtered second images in the second dimension;
Dividing the three-dimensional image formed by the multi-frame filtered second image in the second dimension by using a third dimension dividing plane to obtain a multi-frame third image in the third dimension;
taking each frame of third medical image in the third dimension as a guide image, and filtering each frame of third image in the third dimension corresponding to the guide image to obtain multi-frame filtered third images in the third dimension;
and obtaining the region of interest based on the multi-frame filtered third image in the third dimension.
8. A computer readable storage medium, which when executed by a processor within a device, causes the device to perform the method of acquiring a region of interest of any of claims 1-6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010340713.8A CN111967462B (en) | 2020-04-26 | 2020-04-26 | Method and device for acquiring region of interest |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010340713.8A CN111967462B (en) | 2020-04-26 | 2020-04-26 | Method and device for acquiring region of interest |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111967462A CN111967462A (en) | 2020-11-20 |
| CN111967462B true CN111967462B (en) | 2024-02-02 |
Family
ID=73358134
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010340713.8A Active CN111967462B (en) | 2020-04-26 | 2020-04-26 | Method and device for acquiring region of interest |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111967462B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112712507B (en) * | 2020-12-31 | 2023-12-19 | 杭州依图医疗技术有限公司 | Method and device for determining calcified region of coronary artery |
| EP4341896A1 (en) * | 2021-05-21 | 2024-03-27 | Qualcomm Incorporated | Face detection based filtering for image processing |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104541289A (en) * | 2013-07-24 | 2015-04-22 | 华为技术有限公司 | Interest point judgement method and interest point judgement device |
| CN105844687A (en) * | 2016-04-07 | 2016-08-10 | 北京雅森科技发展有限公司 | Apparatus and method for processing medical image |
| CN106408576A (en) * | 2016-09-08 | 2017-02-15 | 飞依诺科技(苏州)有限公司 | Automatic region of interest segmentation method and system based on three-dimensional ultrasonic image |
| CN109145944A (en) * | 2018-07-11 | 2019-01-04 | 哈尔滨工程大学 | A kind of classification method based on longitudinal depth of 3 D picture learning characteristic |
| CN109523560A (en) * | 2018-11-09 | 2019-03-26 | 成都大学 | A kind of three-dimensional image segmentation method based on deep learning |
| CN109543623A (en) * | 2018-11-26 | 2019-03-29 | 微医云(杭州)控股有限公司 | A kind of development of fetus condition predicting device based on Magnetic resonance imaging |
| CN110533029A (en) * | 2019-08-02 | 2019-12-03 | 杭州依图医疗技术有限公司 | Determine the method and device of target area in image |
| CN110782446A (en) * | 2019-10-25 | 2020-02-11 | 杭州依图医疗技术有限公司 | Method and device for determining volume of lung nodule |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11234666B2 (en) * | 2018-05-31 | 2022-02-01 | Canon Medical Systems Corporation | Apparatus and method for medical image reconstruction using deep learning to improve image quality in position emission tomography (PET) |
| CN109446951B (en) * | 2018-10-16 | 2019-12-10 | 腾讯科技（深圳）有限公司 | Semantic segmentation method, device and equipment for three-dimensional image and storage medium |
-
2020
- 2020-04-26 CN CN202010340713.8A patent/CN111967462B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104541289A (en) * | 2013-07-24 | 2015-04-22 | 华为技术有限公司 | Interest point judgement method and interest point judgement device |
| CN105844687A (en) * | 2016-04-07 | 2016-08-10 | 北京雅森科技发展有限公司 | Apparatus and method for processing medical image |
| CN106408576A (en) * | 2016-09-08 | 2017-02-15 | 飞依诺科技(苏州)有限公司 | Automatic region of interest segmentation method and system based on three-dimensional ultrasonic image |
| CN109145944A (en) * | 2018-07-11 | 2019-01-04 | 哈尔滨工程大学 | A kind of classification method based on longitudinal depth of 3 D picture learning characteristic |
| CN109523560A (en) * | 2018-11-09 | 2019-03-26 | 成都大学 | A kind of three-dimensional image segmentation method based on deep learning |
| CN109543623A (en) * | 2018-11-26 | 2019-03-29 | 微医云(杭州)控股有限公司 | A kind of development of fetus condition predicting device based on Magnetic resonance imaging |
| CN110533029A (en) * | 2019-08-02 | 2019-12-03 | 杭州依图医疗技术有限公司 | Determine the method and device of target area in image |
| CN110782446A (en) * | 2019-10-25 | 2020-02-11 | 杭州依图医疗技术有限公司 | Method and device for determining volume of lung nodule |
Non-Patent Citations (2)
| Title |
|---|
| 2D image classification for 3D anatomy localization: employing deep convolutional neural networks;Bob D. de Vos et al.;《Medical Imaging 2016: Image Processing》;第9784卷;第9784Y1-9784Y7页 * |
| 基于DICOM医学图像的虚拟切片提取技术研究与实现;吕晓琪 等;《计算机应用研究》;第30卷(第07期);第2226-2229页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111967462A (en) | 2020-11-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110796670B (en) | Dissection method and device for dissecting interbed artery | |
| CN111145206A (en) | Liver image segmentation quality evaluation method and device and computer equipment | |
| CN111967462B (en) | Method and device for acquiring region of interest | |
| CN113808146B (en) | Multi-organ segmentation method and system for medical image | |
| CN110782446B (en) | Method and device for determining volume of lung nodule | |
| CN111696126B (en) | Multi-view-angle-based multi-task liver tumor image segmentation method | |
| CN111815766A (en) | Processing method and system for reconstructing blood vessel three-dimensional model based on 2D-DSA image | |
| CN116503607B (en) | CT image segmentation method and system based on deep learning | |
| CN111524142B (en) | Automatic segmentation method for cerebrovascular images | |
| CN114332132A (en) | Image segmentation method and device and computer equipment | |
| CN116452618A (en) | Three-input spine CT image segmentation method | |
| CN102419864A (en) | Method and device for extracting skeletons of brain CT (computerized tomography) image | |
| KR102182357B1 (en) | Surgical assist device and method for 3D analysis based on liver cancer area in CT image | |
| CN117315210A (en) | Image blurring method based on stereoscopic imaging and related device | |
| CN116725563B (en) | Eyeball salience measuring device | |
| CN116030043A (en) | Multi-mode medical image segmentation method | |
| Wen et al. | A novel lesion segmentation algorithm based on U-Net network for Tuberculosis CT image | |
| CN113256754B (en) | Stacking projection reconstruction method for segmented small-area tumor mass | |
| Zhou et al. | Automatic segmentation algorithm of femur and tibia based on Vnet-C network | |
| CN113706541B (en) | Image processing method and device | |
| CN110533637B (en) | Method and device for detecting object | |
| CN111462067B (en) | Image segmentation method and device | |
| Zhu et al. | Automatic measurement of scoliosis based on an improved segmentation model | |
| CN112712507A (en) | Method and device for determining calcified area of coronary artery | |
| CN113362294A (en) | Puncture needle identification method, system and equipment containing puncture needle ultrasonic blood vessel image |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |