CN114332132A - Image segmentation method and device and computer equipment - Google Patents
Image segmentation method and device and computer equipment Download PDFInfo
- Publication number
- CN114332132A CN114332132A CN202111679357.3A CN202111679357A CN114332132A CN 114332132 A CN114332132 A CN 114332132A CN 202111679357 A CN202111679357 A CN 202111679357A CN 114332132 A CN114332132 A CN 114332132A
- Authority
- CN
- China
- Prior art keywords
- region
- interest
- segmentation
- image
- processed
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Abstract
The application relates to an image segmentation method, an image segmentation device and computer equipment. The method comprises the following steps: acquiring a medical image to be processed and a target part segmentation image in the medical image to be processed; the target part comprises at least one region of interest; acquiring an initial segmentation result of an interest region in the medical image to be processed according to the medical image to be processed; and (4) according to the target part segmentation image, carrying out false positive filtering on the initial segmentation result of the region of interest to obtain the target segmentation result of the region of interest in the target part. Therefore, through the target part segmentation image, false positive interference in the initial segmentation result of the region of interest in the target part can be effectively eliminated, and the accuracy of the segmentation result of the region of interest in the target part is improved.
Description
Technical Field
The present application relates to the field of medical image processing technologies, and in particular, to an image segmentation method, an image segmentation apparatus, and a computer device.
Background
Clinical identification of lung tumors relies primarily on the visual judgment and rough assessment of the severity of Computed Tomography (CT) images of the lungs by clinicians, the accuracy of which depends entirely on the individual experience of the clinician. And the lung CT image data shows explosive growth, thereby greatly enhancing the workload of doctors, and causing misdiagnosis and missed diagnosis easily in the disease detection process.
In order to improve the efficiency and accuracy of lung tumor identification, in the related technology, a deep learning technology can be used for carrying out image segmentation processing on a lung CT image, and determining the region where the lung tumor is located, so as to assist a doctor in diagnosis.
However, when the lung tumor region is segmented by using the deep learning technique in the related art, the segmentation result is inaccurate, which causes a problem of false positive in the lung tumor segmentation result.
Disclosure of Invention
In view of the above, it is desirable to provide an image segmentation method, an image segmentation apparatus, and a computer device, which can effectively eliminate the interference of false positive segmentation results of a target region and improve the accuracy of the segmentation results.
In a first aspect, the present application provides an image segmentation method. The method comprises the following steps:
acquiring a medical image to be processed and a target part segmentation image in the medical image to be processed; the target part comprises at least one region of interest;
acquiring an initial segmentation result of an interest region in the medical image to be processed according to the medical image to be processed;
and (4) according to the target part segmentation image, carrying out false positive filtering on the initial segmentation result of the region of interest to obtain the target segmentation result of the region of interest in the target part.
In one embodiment, acquiring a target region segmentation image in a medical image to be processed includes:
and segmenting the region of the target part in the medical image to be processed through a preset first segmentation model to obtain a segmentation image of the target part.
In one embodiment, the obtaining of the target segmentation result of the region of interest in the target region by performing false positive filtering on the initial segmentation result of the region of interest according to the target region segmentation image includes:
and if the initial segmentation result of the region of interest is not in the target part segmentation image, determining that the target segmentation result of the region of interest in the target part is false positive.
In one embodiment, obtaining an initial segmentation result of a region of interest in a medical image to be processed according to the medical image to be processed includes:
adjusting the window width and the window level of the medical image to be processed;
and segmenting the region of interest in the adjusted medical image to be processed to obtain an initial segmentation result of the region of interest.
In one embodiment, if the target region is a lung, the adjusting the window width and the window level of the medical image to be processed includes:
and adjusting the window width and the window level of the medical image to be processed according to the window parameters of the lung window.
In one embodiment, if the target region is a lung, the window width and the window level of the medical image to be processed are adjusted, including:
and adjusting the window width and the window position of the medical image to be processed according to the window parameters of the lung milled glass window.
In one embodiment, segmenting the region of interest in the adjusted medical image to be processed to obtain an initial segmentation result of the region of interest, includes:
and segmenting the region of interest in the adjusted medical image to be processed through a preset second segmentation model to obtain an initial segmentation result of the region of interest.
In one embodiment, the building process of the second segmentation model includes:
acquiring a plurality of sample medical images; the sample medical image comprises a target part, and at least one region of interest exists in the target part;
sampling target parts in the plurality of sample medical images according to a preset sampling proportion to obtain a plurality of first sample medical images, and sampling interested areas in the plurality of sample medical images to obtain a plurality of second sample medical images;
and training the initial second segmentation model according to the plurality of first sample medical images, the plurality of second sample medical images and a preset gold standard until an output result of the initial second segmentation model meets a preset convergence condition to obtain a second segmentation model.
In a second aspect, the present application further provides an image segmentation apparatus. The device includes:
the first acquisition module is used for acquiring a medical image to be processed and a target part segmentation image in the medical image to be processed; the target part comprises at least one region of interest;
the second acquisition module is used for acquiring an initial segmentation result of an interest region in the medical image to be processed according to the medical image to be processed;
and the detection module is used for carrying out false positive filtering on the initial segmentation result of the region of interest according to the segmentation image of the target part to obtain the target segmentation result of the region of interest in the target part.
In a third aspect, the present application also provides a computer device. The computer device comprises a memory in which a computer program is stored and a processor which, when executing the computer program, performs the steps of any of the method embodiments of the first aspect described above.
In a fourth aspect, the present application further provides a computer-readable storage medium. The computer-readable storage medium having stored thereon a computer program which, when being executed by a processor, carries out the steps of any of the method embodiments of the first aspect described above.
In a fifth aspect, the present application further provides a computer program product. The computer program product comprising a computer program that when executed by a processor performs the steps of any of the method embodiments of the first aspect described above.
According to the image segmentation method, the image segmentation device and the computer equipment, the medical image to be processed and the target part segmentation image in the medical image to be processed are obtained; the target part comprises at least one region of interest; acquiring an initial segmentation result of an interest region in the medical image to be processed according to the medical image to be processed; and (4) according to the target part segmentation image, carrying out false positive filtering on the initial segmentation result of the region of interest to obtain the target segmentation result of the region of interest in the target part. In the method, the whole segmentation image of the target part is accurately acquired from the medical image to be processed, and the segmentation image of the target part is obtained. Furthermore, false positive interference in the initial segmentation result of the region of interest in the target part can be effectively eliminated through the segmentation image of the target part, and the accuracy of the segmentation result of the region of interest in the target part is improved.
Drawings
FIG. 1 is a flow diagram illustrating a method for image segmentation in one embodiment;
FIG. 2 is a flow chart illustrating an image segmentation method according to another embodiment;
FIG. 3 is a schematic flow chart illustrating construction of a second segmentation model in one embodiment;
FIG. 4 is a flowchart illustrating an image segmentation method according to another embodiment;
FIG. 5 is a block diagram showing the structure of an image segmentation apparatus according to an embodiment;
FIG. 6 is a diagram illustrating an internal structure of a computer device according to an embodiment.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
Deep learning is a mainstream technology adopted by a current medical image segmentation task, is limited by factors such as training sample limitation and semantic information similarity, and has a false positive problem in a segmentation result, namely, the segmentation model erroneously identifies an out-of-focus region as a target region.
Based on this, in order to reduce the influence of false positive problems on the accuracy of the focus segmentation result and improve the performance of the product, the application provides an image segmentation method, an image segmentation device and computer equipment.
The image segmentation method provided by the present application can be applied to an image segmentation apparatus, which can be implemented in a software and/or hardware manner, and the apparatus can be integrated in a computer device with a medical image processing function, for example: terminals, servers, medical devices, etc.
The terminal may include, but is not limited to, software running in a physical device, such as an application program or a client installed on the device, and may also include, but is not limited to, a personal computer, a laptop, a smartphone, a tablet, and a portable wearable device installed with an application. The server may include, but is not limited to, at least one of a stand-alone server, a distributed server, a cloud server, and a server cluster.
The following describes in detail the technical solutions of the embodiments of the present application and how to solve the above technical problems with the embodiments of the present application by using the embodiments and with reference to the drawings. The following several specific embodiments may be combined with each other, and details of the same or similar concepts or processes may not be repeated in some embodiments. The image segmentation method provided by the embodiment of the present application may be executed by the computer device, a specific medical device, or an image segmentation apparatus provided by the present application. It is to be understood that the embodiments described are only some of the embodiments of the present application and not all of them.
In one embodiment, as shown in fig. 1, there is provided an image segmentation method, which is described by taking the method as an example applied to the computer device in fig. 1, and includes the following steps:
step 110: acquiring a medical image to be processed and a target part segmentation image in the medical image to be processed; the target site includes at least one region of interest therein.
The medical image to be processed is a medical image including information of a target portion of an object to be examined, and the medical image may be a two-dimensional image or a three-dimensional image. Specifically, the medical image to be processed can be obtained by using any one of the following medical imaging techniques: CT scanning technology, Magnetic Resonance Imaging (MRI) technology, X-Ray (X-Ray) Imaging technology, Positron Emission Tomography (PET) technology, and the like.
Further, the target region in the medical image to be processed is an organ or tissue of the subject where a lesion may exist. If the subject is a human, the target site may be: lung, liver, heart, spleen, brain, gallbladder, genital organ, etc. The region of interest is a region in which diseased tissue caused by disease is located in the target site.
In one possible implementation manner, the implementation procedure of step 110 may be: acquiring a medical image to be processed from a medical image database, identifying a target part from the medical image to be processed, and segmenting the target part to obtain a segmented image of the target part. The target part segmentation image is a global image of the target part, and the distribution condition of tissues in the target part and the boundary line of the target part can be accurately and clearly obtained from the target part segmentation image.
It should be noted that the medical image to be processed may be a real-time image obtained by scanning a target region of the object to be examined, or may be a medical image in a stored medical image database. Meanwhile, when medical images to be processed are obtained, the medical images to be processed can be obtained one by one or in batches, and the source and the obtaining mode of the medical images to be processed are not limited in the embodiment of the application.
Step 120: and acquiring an initial segmentation result of the region of interest in the medical image to be processed according to the medical image to be processed.
The initial segmentation result of the region of interest may include one region of interest segmentation image or a plurality of region of interest segmentation images. It should be understood that in case no region of interest is identified from the medical image to be processed, the initial segmentation result of the region of interest is empty.
In one possible implementation manner, the implementation procedure of step 120 may be: for the medical image to be processed, a pre-trained segmentation model is adopted to detect and identify the region of interest of a target part in the medical image to be processed, and image segmentation is carried out on the identified region of interest to obtain an initial segmentation result of the region of interest in the medical image to be processed.
The segmentation model may be any machine learning model, such as a deep learning model, a random forest model, a decision tree model, and the like. The segmentation processing network in the machine learning model may be a single neural network or a cascade network formed by at least two neural networks. The embodiment of the present application does not limit this.
Further, the segmentation model is obtained by training in advance by using the sample medical image, and the training process can be supervised learning, unsupervised learning or semi-supervised learning.
In the supervised learning mode training, a target part and an interested area in the target part are marked in advance in a sample medical image. In the training process, the initial segmentation model learns the physiological characteristics of the target part and the difference between the non-interesting region and the interesting region in the target part according to the pre-labeled sample medical image, and then outputs the interesting region segmentation result corresponding to the sample medical image, and when the output result meets the preset convergence condition, the initial segmentation model is trained to be finished, so that the trained segmentation model is obtained.
The segmentation model may be a model for segmenting a region of interest with respect to a specific part of the object to be examined, or may be a model for segmenting a region of interest with respect to a plurality of parts of the object to be examined, which is not limited in the embodiment of the present application.
Step 130: and (4) according to the target part segmentation image, carrying out false positive filtering on the initial segmentation result of the region of interest to obtain the target segmentation result of the region of interest in the target part.
Although the segmentation model is much higher than manual work in the identification rate of the region of interest, due to factors such as image imaging, interferents and artifacts, the marked region of interest may be a region of non-interest in the initial segmentation result of the region of interest output by the segmentation model, i.e. there is a false positive in the segmentation result. Therefore, after obtaining the initial segmentation result of the region of interest according to the segmentation model, a false positive removal operation is required to obtain a target segmentation result that can reflect the region of interest in the target portion.
However, as can be seen from the a priori knowledge and the pathological knowledge, in the medical image to be processed, the region of interest should be located within the target region, and the false positive result may exist outside the target region. Therefore, for the initial segmentation result of the region of interest, the region of interest segmentation image outside the target part can be filtered quickly and effectively by using the target part segmentation image, and false positive results are removed.
In one possible implementation manner, the implementation procedure of step 130 may be: and if the initial segmentation result of the region of interest is not in the target part segmentation image, determining that the target segmentation result of the region of interest in the target part is false positive. In this way, after removing false positive results from the initial segmentation result of the region of interest, the target segmentation result of the region of interest in the target portion is obtained.
As an example, a region-of-interest segmented image in an initial segmentation result of a region of interest is sequentially subjected to region matching with a target part segmented image, and if the region-of-interest segmented image is not in the target part segmented image, the region-of-interest segmented image is determined to be false positive. And if the region-of-interest segmentation image is in the target part segmentation image, determining the region-of-interest segmentation image as a target segmentation result of the region of interest in the target part.
In the image segmentation method, a target part segmentation image in a medical image to be processed is obtained, and false positive filtering is performed on an initial segmentation result of a region of interest through the target part segmentation image to obtain a target segmentation result of the region of interest. False positive interference of the segmentation result of the region of interest in the target part is effectively eliminated, and the accuracy of the segmentation result of the region of interest is improved.
In one embodiment, based on the technical idea of obtaining the initial segmentation result of the region of interest in the medical image to be processed through the segmentation model in the above embodiment, the process of obtaining the segmented image of the target portion in the medical image to be processed may be: and segmenting the region of the target part in the medical image to be processed through a preset first segmentation model to obtain a segmentation image of the target part.
Similarly, the first segmentation model may be any machine learning model. Taking supervised learning as an example, the process of training the first segmentation model is as follows: acquiring a sample medical image including a target part, wherein the target part is marked in the sample medical image in advance; inputting the sample medical image into an initial first segmentation model, training the initial first segmentation model, and if the error between the target part segmentation image output by the initial first segmentation model and the region where the marked target part is located is within a preset error range, converging the initial first segmentation model to obtain the trained first segmentation model.
Further, since the target portion segmentation image is used to filter false positive results in the initial segmentation result of the region of interest, the accuracy requirement is high. Therefore, when the first segmentation model is trained, the segmentation processing network of the first segmentation model can be optimized, and the first segmentation model is trained by periodically adopting new sample medical images so as to improve the accuracy of the segmentation result.
In this embodiment, the trained first segmentation model is used to obtain the target region segmentation image from the medical image to be processed, so as to improve the efficiency of obtaining the target region segmentation image.
Based on the above method embodiments, in an embodiment, as shown in fig. 2, an implementation process for obtaining an initial segmentation result of a region of interest in a medical image to be processed according to the medical image to be processed may include the following steps:
step 210: and adjusting the window width and the window level of the medical image to be processed.
In clinical practice, when a doctor in an imaging department observes a CT image to diagnose, the window width and the window level of the image can be adjusted to a plurality of ranges, and the images under different window widths and window levels can present different details, so that the diagnosis is more accurate. The method and the device have the advantages that when the region of interest is segmented, in order to improve the segmentation accuracy, the window width and the window level of the medical image to be processed can be adjusted, so that the region of interest is clearer, the boundary is more definite, and the segmentation effect is improved.
The medical image to be processed is a CT image, for example, and the window width refers to a CT value range displayed by the CT image. The texture in this CT value range is divided into 16 gray levels from white to black according to the density (because human eyes can only recognize 16 gray levels) for observing contrast. Window level (also called window center) refers to the mean or center value over a window width.
As an example, if the window width of a CT image is 100Hu, the CT value resolvable by the human eye is 100/16-6.25 Hu. That is, the difference between the CT values of two tissues in the medical image to be processed is more than 6.25Hu, so that the two tissues can be identified by human eyes. Further, if the window level is selected at 0 Hu; then all tissues in the range of 100Hu can be displayed and identified by human eyes, with the window level as the center (0Hu), upwards comprising +50Hu, and downwards comprising-50 Hu. Wherein tissues greater than +50Hu all appear white; tissues less than-50 Hu all appear black.
Therefore, the width of the window width directly affects the sharpness and contrast of the image. If a narrow window width is used, the displayed CT value range is small, the CT value represented by each gray scale is small in amplitude and strong in contrast, and the method is suitable for observing tissue structures (such as brain tissues) with close density. Conversely, if a wide window width is used, the displayed range of CT values is large, and the amplitude of the CT value represented by each gray level is large, the image contrast is poor, but the density is uniform, and the method is suitable for observing structures with large density differences (such as bones and soft tissues).
It should be noted that, when adjusting the window width and the window level, the window level should be close to the CT value of the target region to be observed in principle, and the window width should reflect the change range of the CT value of the tissue or lesion in the target region.
In one possible application scenario, if the target region of interest in the medical image to be processed is the lung. The adjustment process of step 210 may include the following (1) and/or (2).
(1) And adjusting the window width and the window level of the medical image to be processed according to the window parameters of the lung window.
And if the window width of the lung window is 1500Hu and the window position is-400 Hu, adjusting the window width of the medical image to be processed to be 1500Hu and the window position to be-400 Hu, wherein the adjusted medical image to be processed can present more detailed information of the lung.
(2) And adjusting the window width and the window position of the medical image to be processed according to the window parameters of the lung milled glass window.
Wherein, the window width of the lung frosted glass window is 800Hu, the window position is-600 Hu, the window width of the medical image to be processed is adjusted to be 800Hu, the window position is-600 Hu, and the adjusted medical image to be processed can clearly display the frosted glass window nodule information of the lung.
Optionally, the pixels in the medical image to be processed may be subjected to normalization, resampling, pixel expansion, and the like, so as to enrich the detail information of the medical image to be processed. The embodiment of the present application does not limit this.
Step 220: and segmenting the region of interest in the adjusted medical image to be processed to obtain an initial segmentation result of the region of interest.
In one possible implementation manner, the implementation procedure of step 220 may be: and segmenting the region of interest in the adjusted medical image to be processed through a preset second segmentation model to obtain an initial segmentation result of the region of interest.
The second segmentation model is a pre-trained deep learning model, the region of interest can be accurately determined from the medical image to be processed through the second segmentation model, the region of interest is segmented, an initial segmentation result of the region of interest is rapidly obtained, and segmentation efficiency of the region of interest is improved.
The second segmentation model and the first segmentation model may be the same machine learning model, and the difference is that the sample data used for training is different, and the configuration parameters of the models are different. Further, the second segmentation model and the first segmentation model may also be different machine learning models, and are trained using different sample data. The embodiment of the present application does not limit the specific type of the segmentation model.
In this embodiment, the medical images under different window widths and window levels can present different details, so that the adjusted medical images to be processed can present comprehensive characteristic information of the region of interest by adjusting the window width and the window level of the medical images to be processed, and the region of interest can be conveniently and accurately identified and efficiently segmented. In addition, when the region of interest is segmented from the adjusted medical image to be processed, the segmentation is realized through the second segmentation model, and the segmentation efficiency of the region of interest is improved.
Further, in one embodiment, as shown in fig. 3, the second segmentation model may be constructed by:
step 310: acquiring a plurality of sample medical images; the sample medical image includes a target site, and at least one region of interest is present in the target site.
Before the initial second segmentation model is trained by using a plurality of sample medical images, the window widths and the window levels of the plurality of sample medical images can be adjusted, so that the initial second segmentation model can obtain stronger adaptability on limited sample data, and for unseen sample medical images, a more accurate region-of-interest segmentation result can be output.
If the target region in the sample medical images is a lung, in one possible implementation, the adjusting the window widths and window levels of the plurality of sample medical images may be performed by: according to the window parameters of the lung window, adjusting the window width and the window level of each sample medical image, and according to the window parameters of the lung milled glass window, adjusting the window width and the window level of each sample medical image; and taking the adjusted plurality of sample medical images as training data of the initial second segmentation model.
Step 320: sampling target parts in the plurality of sample medical images according to a preset sampling proportion to obtain a plurality of first sample medical images, and adopting interested areas in the plurality of sample medical images to obtain a plurality of second sample medical images.
As an example, if the preset ratio 4: and 6, sampling the target part with a probability of 40% for each sample medical image to obtain a first sample medical image. Sampling is performed on the region of interest in the target region with a probability of 60% to obtain a second sample medical image.
Further, the first sample medical image and the second sample medical image are image blocks obtained by sampling from the sample medical images, and therefore, the size of the sampling range can be preset before sampling.
As an example, if the preset sampling range size is: 64x64x64, the first sample medical image is an image block of 64x64x64 pixels acquired from a random point in the target region as the center point, and the second sample medical image is an image block of 64x64x64 pixels acquired from a random point in the interest as the center point.
It should be noted that, since the first sample medical image is an image block sampled for the entire target portion, the image block has a global property and a larger visible range, but there may be irrelevant interference in the sampled image block. That is, the first medical image may include a region of interest in the target region, and may also be a region of non-interest in the target region. The second sample medical image is an image block sampled in the region of interest, has locality, and has a small visible range, but can obtain more accurate characteristic information.
Therefore, the sampling ratio can be adjusted according to the learning effect of the second model to adjust the number of the first sample medical images and the second sample medical images, so that the model can perform targeted learning according to the input sample medical images.
Step 330: and training the initial second segmentation model according to the plurality of first sample medical images, the plurality of second sample medical images and a preset gold standard until an output result of the initial second segmentation model meets a preset convergence condition to obtain a second segmentation model.
The gold standard is an interested area which is manually and preliminarily calibrated in a target part of the sample medical image.
In one possible implementation manner, the implementation procedure of step 330 may be: and inputting the plurality of first sample medical images and the plurality of second sample medical images into the initial second segmentation model, and training the initial second segmentation model. And if the region-of-interest segmentation result output by the initial second segmentation model and the gold standard are within the allowable error range, the output result of the initial second segmentation model meets the preset convergence condition, and the second segmentation model is obtained.
In the embodiment of the application, when the second segmentation model is trained, the window width and the window level of the sample medical image are adjusted, so that the initial second segmentation model can effectively learn under the condition that the contrast between the region of interest and the background is stronger, the recognition capability of the region of interest in the unknown medical image is enhanced, and the initial second segmentation model can obtain stronger adaptability on limited data. In addition, a plurality of first sample medical images and a plurality of second sample medical images are obtained according to a preset sampling proportion, the initial second segmentation model focuses on the region of interest, meanwhile, the characteristic information of the target part and the relation between the target part and the region of interest are learned, and false positive brought by difference is reduced. Therefore, the number of the first sample medical images and the number of the second sample medical images input into the initial second model are controlled by adjusting the window width and the window level of the sample medical images and by sampling proportion, the learning effect of the model is improved, and the segmentation result of the trained second segmentation model is more accurate when the region of interest is segmented.
Combining the above embodiments of the method, as shown in fig. 4, the present application further provides another image segmentation method, which is described by taking the method applied to the computer device in fig. 1 as an example, and includes the following steps:
step 410: acquiring a medical image to be processed;
step 420: segmenting the region of the target part in the medical image to be processed through a preset first segmentation model to obtain a target part segmentation image;
step 430: adjusting the window width and the window level of the medical image to be processed;
step 440: segmenting the region of interest in the adjusted medical image to be processed through a preset second segmentation model to obtain an initial segmentation result of the region of interest;
step 450: and if the initial segmentation result of the region of interest is not in the target part segmentation image, determining that the target segmentation result of the region of interest in the target part is false positive.
In the image segmentation method provided in this embodiment, the implementation principle and technical effect of each step are similar to those in the previous embodiments, and are not described herein again.
It is to be understood that, although the steps in the flowcharts according to the embodiments described above are sequentially displayed as indicated by arrows, the steps are not necessarily sequentially executed in the order indicated by the arrows. The steps are not performed in the exact order shown and described, and may be performed in other orders, unless explicitly stated otherwise. Moreover, at least a part of the steps in the flowcharts related to the embodiments described above may include multiple steps or multiple stages, which are not necessarily performed at the same time, but may be performed at different times, and the execution order of the steps or stages is not necessarily sequential, but may be rotated or alternated with other steps or at least a part of the steps or stages in other steps.
Based on the same inventive concept, the embodiment of the present application further provides an image segmentation apparatus for implementing the image segmentation method. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme described in the above method, so specific limitations in one or more embodiments of the image segmentation device provided below can be referred to the limitations of the image segmentation method in the above, and are not described herein again.
In one embodiment, as shown in fig. 5, there is provided an image segmentation apparatus 500, comprising: a first acquisition module 510, a second acquisition module 520, and a detection module 530, wherein:
a first obtaining module 510, configured to obtain a medical image to be processed and a target portion segmentation image in the medical image to be processed; the target part comprises at least one region of interest;
a second obtaining module 520, configured to obtain an initial segmentation result of an interest region in the medical image to be processed according to the medical image to be processed;
the detecting module 530 is configured to perform false positive filtering on the initial segmentation result of the region of interest according to the target portion segmentation image, so as to obtain a target segmentation result of the region of interest in the target portion.
In one embodiment, the first obtaining module 510 includes:
and the part segmentation unit is used for segmenting the region of the target part in the medical image to be processed through a preset first segmentation model to obtain a target part segmentation image.
In one embodiment, the detection module 530 includes:
and the false positive filtering unit is used for determining that the target segmentation result of the region of interest in the target part is false positive if the initial segmentation result of the region of interest is not in the target part segmentation image.
In one embodiment, the second obtaining module 520 includes:
the adjusting unit is used for adjusting the window width and the window level of the medical image to be processed;
and the region segmentation unit is used for segmenting the region of interest in the adjusted medical image to be processed to obtain an initial segmentation result of the region of interest.
In one embodiment, the target site is a lung, and the adjusting unit includes:
and the first adjusting subunit is used for adjusting the window width and the window level of the medical image to be processed according to the window parameters of the lung window.
In one embodiment, the target site is a lung, and the adjusting unit includes:
and the second adjusting subunit is used for adjusting the window width and the window level of the medical image to be processed according to the window parameters of the lung milled glass window.
In one embodiment, the region dividing unit includes:
and the segmentation subunit is used for segmenting the region of interest in the adjusted medical image to be processed through a preset second segmentation model to obtain an initial segmentation result of the region of interest.
In one embodiment, the apparatus 500 further comprises:
the third acquisition module is used for acquiring a plurality of sample medical images; the sample medical image comprises a target part, and at least one interested area exists in the target part;
the sampling module is used for sampling target parts in the plurality of sample medical images according to a preset sampling proportion to obtain a plurality of first sample medical images, and sampling interested areas in the plurality of sample medical images to obtain a plurality of second sample medical images;
and the modeling module is used for training the initial second segmentation model according to the plurality of first sample medical images, the plurality of second sample medical images and a preset gold standard until the output result of the initial second segmentation model meets a preset convergence condition to obtain the second segmentation model.
The respective modules in the image segmentation apparatus can be wholly or partially implemented by software, hardware and a combination thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.
In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as shown in fig. 6. The computer device includes a processor, a memory, a communication interface, a display screen, and an input device connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The communication interface of the computer device is used for carrying out wired or wireless communication with an external terminal, and the wireless communication can be realized through WIFI, an operator network, NFC (near field communication) or other technologies. The computer program is executed by a processor to implement an image segmentation method. The display screen of the computer equipment can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer equipment can be a touch layer covered on the display screen, a key, a track ball or a touch pad arranged on the shell of the computer equipment, an external keyboard, a touch pad or a mouse and the like.
Those skilled in the art will appreciate that the architecture shown in fig. 6 is merely a block diagram of some of the structures associated with the disclosed aspects and is not intended to limit the computing devices to which the disclosed aspects apply, as particular computing devices may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.
In one embodiment, a computer device is provided, comprising a memory and a processor, the memory having a computer program stored therein, the processor implementing the following steps when executing the computer program:
acquiring a medical image to be processed and a target part segmentation image in the medical image to be processed; the target part comprises at least one region of interest;
acquiring an initial segmentation result of an interest region in the medical image to be processed according to the medical image to be processed;
and (4) according to the target part segmentation image, carrying out false positive filtration on the initial segmentation result of the region of interest to obtain the target segmentation result of the region of interest in the target part.
The implementation principle and technical effect of the computer device provided by the above embodiment are similar to those of the above method embodiment, and are not described herein again.
In one embodiment, a computer-readable storage medium is provided, having a computer program stored thereon, which when executed by a processor, performs the steps of:
acquiring a medical image to be processed and a target part segmentation image in the medical image to be processed; the target part comprises at least one region of interest;
acquiring an initial segmentation result of an interest region in the medical image to be processed according to the medical image to be processed;
and (4) according to the target part segmentation image, carrying out false positive filtration on the initial segmentation result of the region of interest to obtain the target segmentation result of the region of interest in the target part.
The implementation principle and technical effect of the computer-readable storage medium provided by the above embodiments are similar to those of the above method embodiments, and are not described herein again.
In one embodiment, a computer program product is provided, comprising a computer program which, when executed by a processor, performs the steps of:
acquiring a medical image to be processed and a target part segmentation image in the medical image to be processed; the target part comprises at least one region of interest;
acquiring an initial segmentation result of an interest region in the medical image to be processed according to the medical image to be processed;
and (4) according to the target part segmentation image, carrying out false positive filtration on the initial segmentation result of the region of interest to obtain the target segmentation result of the region of interest in the target part.
The foregoing embodiments provide a computer program product, which has similar implementation principles and technical effects to those of the foregoing method embodiments, and will not be described herein again.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database or other medium used in the embodiments provided herein can include at least one of non-volatile and volatile memory. Non-volatile Memory may include Read-Only Memory (ROM), magnetic tape, floppy disk, flash Memory, optical storage, or the like. Volatile Memory can include Random Access Memory (RAM) or external cache Memory. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM), among others.
The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.
Claims (10)
1. A method of image segmentation, the method comprising:
acquiring a medical image to be processed and a target part segmentation image in the medical image to be processed; the target part comprises at least one region of interest;
acquiring an initial segmentation result of an interest region in the medical image to be processed according to the medical image to be processed;
and carrying out false positive filtering on the initial segmentation result of the region of interest according to the target part segmentation image to obtain a target segmentation result of the region of interest in the target part.
2. The method according to claim 1, wherein the acquiring of the target portion segmentation image in the medical image to be processed comprises:
and segmenting the region of the target part in the medical image to be processed through a preset first segmentation model to obtain a segmented image of the target part.
3. The method according to claim 1 or 2, wherein the performing false positive filtering on the initial segmentation result of the region of interest according to the target portion segmentation image to obtain the target segmentation result of the region of interest in the target portion comprises:
and if the initial segmentation result of the region of interest is not in the segmentation image of the target part, determining that the target segmentation result of the region of interest in the target part is false positive.
4. The method according to claim 1 or 2, wherein the obtaining an initial segmentation result of a region of interest in the medical image to be processed according to the medical image to be processed comprises:
adjusting the window width and the window level of the medical image to be processed;
and segmenting the region of interest in the adjusted medical image to be processed to obtain an initial segmentation result of the region of interest.
5. The method according to claim 4, wherein the target region is a lung, and the adjusting the window width and the window level of the medical image to be processed comprises:
and adjusting the window width and the window level of the medical image to be processed according to the window parameters of the lung window.
6. The method according to claim 4, wherein the target region is a lung, and the adjusting the window width and the window level of the medical image to be processed comprises:
and adjusting the window width and the window position of the medical image to be processed according to the window parameters of the lung milled glass window.
7. The method according to claim 4, wherein the segmenting the region of interest in the adjusted medical image to be processed to obtain an initial segmentation result of the region of interest comprises:
and segmenting the region of interest in the adjusted medical image to be processed through a preset second segmentation model to obtain an initial segmentation result of the region of interest.
8. The method of claim 7, wherein the second segmentation model is constructed by:
acquiring a plurality of sample medical images; the sample medical image comprises a target part, and at least one region of interest exists in the target part;
sampling target parts in the plurality of sample medical images according to a preset sampling proportion to obtain a plurality of first sample medical images, and sampling interested areas in the plurality of sample medical images to obtain a plurality of second sample medical images;
and training an initial second segmentation model according to the plurality of first sample medical images, the plurality of second sample medical images and a preset gold standard until an output result of the initial second segmentation model meets a preset convergence condition to obtain the second segmentation model.
9. An image segmentation apparatus, characterized in that the apparatus comprises:
the system comprises a first acquisition module, a second acquisition module and a third acquisition module, wherein the first acquisition module is used for acquiring a medical image to be processed and a target part segmentation image in the medical image to be processed; the target part comprises at least one region of interest;
the second acquisition module is used for acquiring an initial segmentation result of an interest region in the medical image to be processed according to the medical image to be processed;
and the detection module is used for carrying out false positive filtering on the initial segmentation result of the region of interest according to the target part segmentation image to obtain a target segmentation result of the region of interest in the target part.
10. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor, when executing the computer program, implements the steps of the method of any of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111679357.3A CN114332132A (en) | 2021-12-31 | 2021-12-31 | Image segmentation method and device and computer equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111679357.3A CN114332132A (en) | 2021-12-31 | 2021-12-31 | Image segmentation method and device and computer equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114332132A true CN114332132A (en) | 2022-04-12 |
Family
ID=81023232
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111679357.3A Pending CN114332132A (en) | 2021-12-31 | 2021-12-31 | Image segmentation method and device and computer equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114332132A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115131333A (en) * | 2022-07-20 | 2022-09-30 | 珠海横琴圣澳云智科技有限公司 | Method and device for detecting and segmenting image based on instance uncertainty |
| CN115482248A (en) * | 2022-09-22 | 2022-12-16 | 推想医疗科技股份有限公司 | Image segmentation method and device, electronic device and storage medium |
| CN116071375A (en) * | 2023-03-10 | 2023-05-05 | 福建自贸试验区厦门片区Manteia数据科技有限公司 | Image segmentation method and device, storage medium and electronic equipment |
-
2021
- 2021-12-31 CN CN202111679357.3A patent/CN114332132A/en active Pending
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115131333A (en) * | 2022-07-20 | 2022-09-30 | 珠海横琴圣澳云智科技有限公司 | Method and device for detecting and segmenting image based on instance uncertainty |
| CN115482248A (en) * | 2022-09-22 | 2022-12-16 | 推想医疗科技股份有限公司 | Image segmentation method and device, electronic device and storage medium |
| CN115482248B (en) * | 2022-09-22 | 2023-12-08 | 推想医疗科技股份有限公司 | Image segmentation method, device, electronic equipment and storage medium |
| CN116071375A (en) * | 2023-03-10 | 2023-05-05 | 福建自贸试验区厦门片区Manteia数据科技有限公司 | Image segmentation method and device, storage medium and electronic equipment |
| CN116071375B (en) * | 2023-03-10 | 2023-09-26 | 福建自贸试验区厦门片区Manteia数据科技有限公司 | Image segmentation method and device, storage medium and electronic equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111127466B (en) | Medical image detection method, device, equipment and storage medium | |
| CN109741346B (en) | Region-of-interest extraction method, device, equipment and storage medium | |
| CN108010021B (en) | Medical image processing system and method | |
| US20200074634A1 (en) | Recist assessment of tumour progression | |
| CN114332132A (en) | Image segmentation method and device and computer equipment | |
| JP2023507109A (en) | Automated tumor identification and segmentation from medical images | |
| CN107563434B (en) | Brain MRI image classification method and device based on three-dimensional convolutional neural network | |
| EP2401719B1 (en) | Methods for segmenting images and detecting specific structures | |
| CN108062749B (en) | Identification method and device for levator ani fissure hole and electronic equipment | |
| CN110246109B (en) | Analysis system, method, device and medium fusing CT image and personalized information | |
| EP2987114B1 (en) | Method and system for determining a phenotype of a neoplasm in a human or animal body | |
| CN111862044A (en) | Ultrasonic image processing method and device, computer equipment and storage medium | |
| US11257210B2 (en) | Method and system of performing medical treatment outcome assessment or medical condition diagnostic | |
| CN109919254B (en) | Breast density classification method, system, readable storage medium and computer device | |
| US20220101034A1 (en) | Method and system for segmenting interventional device in image | |
| CN111968130B (en) | Brain contrast image processing method, device, medium and electronic equipment | |
| Barbosa et al. | Towards automatic quantification of the epicardial fat in non-contrasted CT images | |
| WO2021125950A1 (en) | Image data processing method, method of training a machine learning data processing model and image processing system | |
| Heydarheydari et al. | Auto-segmentation of head and neck tumors in positron emission tomography images using non-local means and morphological frameworks | |
| CN111508004B (en) | Wall motion abnormity ultrasonic processing method, system and equipment based on deep learning | |
| CN114092450A (en) | Real-time image segmentation method, system and device based on gastroscopy video | |
| JP6458166B2 (en) | MEDICAL IMAGE PROCESSING METHOD, DEVICE, SYSTEM, AND PROGRAM | |
| Faisal et al. | Computer assisted diagnostic system in tumor radiography | |
| CN110992312B (en) | Medical image processing method, medical image processing device, storage medium and computer equipment | |
| Delmoral et al. | Segmentation of pathological liver tissue with dilated fully convolutional networks: A preliminary study |
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 |