An image registration based cerebral hemorrhage analysis method and system
By using image registration-based methods and deep neural networks, combined with brain CT and MRI templates, the problem of low detection accuracy in cerebral hemorrhage analysis has been solved, enabling rapid and accurate analysis of the hemorrhage magnitude and brain region, supporting doctors in developing treatment plans.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- SHANDONG UNIV
- Filing Date
- 2022-11-08
- Publication Date
- 2026-07-14
AI Technical Summary
Existing technologies have low detection accuracy in the analysis of cerebral hemorrhage because there are many similarities between different types of cerebral hemorrhage. Relying solely on image analysis makes it difficult to accurately segment and distinguish the hemorrhage, thus delaying the best treatment time.
An image registration-based method was adopted, using normal brain MRI images as registration templates to register patient brain CT images with MRI templates, extracting hematoma features, and combining deep neural networks for automatic segmentation and preprocessing to accurately analyze the hemorrhage magnitude and involved brain regions.
It enables rapid and effective analysis of cerebral hemorrhage, providing information on the magnitude of the hemorrhage and the brain regions involved, helping doctors develop comprehensive treatment plans and improving diagnostic efficiency and accuracy.
Smart Images
Figure CN115937096B_ABST
Abstract
Description
Technical Field
[0001] This disclosure relates to the field of medical image analysis, and in particular to a method and system for analyzing cerebral hemorrhage based on image registration. Background Technology
[0002] The statements in this section are merely background information relating to this disclosure and do not necessarily constitute prior art.
[0003] Cerebral hemorrhage refers to bleeding caused by the rupture of blood vessels within the brain parenchyma without external trauma. It accounts for 20%–30% of all strokes, with an acute-phase mortality rate of 30%–40%. The main causes are related to cerebrovascular diseases, such as hyperlipidemia, diabetes, hypertension, vascular aging, and smoking. Patients with cerebral hemorrhage often experience sudden onset during emotional excitement or strenuous activity. Early mortality is high, and most survivors suffer from varying degrees of motor impairment, cognitive impairment, and speech and swallowing difficulties. In clinical treatment, CT scans are frequently used for rapid examination. Cranial CT scans can clearly show the location and size of the hemorrhage, the morphology of the hematoma, whether it has ruptured into the ventricles, and whether there is a low-density edema zone and mass effect around the hematoma. Generally, when the condition is critical, leading to excessive intracranial pressure and brain herniation, and conservative medical treatment is ineffective, timely surgical intervention should be performed.
[0004] With the development of the times, people's life pressure is constantly increasing, and patients with cerebral hemorrhage are showing a trend of becoming younger. The number of patients is increasing year by year. In addition, the characteristics of rapid onset, speed and severity have increased the burden on hospitals and placed higher and higher demands on doctors' on-site treatment. If the location of the hemorrhage in the patient's brain can be quickly analyzed to obtain the bleeding level and the brain regions involved, it will greatly help doctors to formulate surgical plans and help patients and their families understand the development of the disease.
[0005] Current methods for analyzing cerebral hemorrhage typically involve inputting images of the hemorrhage into a convolutional neural network, then testing the trained model to segment the hemorrhage area and classify the type of hemorrhage. However, because different types of cerebral hemorrhage share many similarities, such as indistinct shapes of the hemorrhage area, training solely based on images does not yield highly accurate results, leading to low detection accuracy. Furthermore, diagnosis requires doctors to combine CT values with brain anatomy to make a diagnosis, which significantly reduces diagnostic efficiency and delays optimal treatment. Summary of the Invention
[0006] To address the technical problems mentioned above, this disclosure provides a method and system for analyzing cerebral hemorrhage based on image registration. It uses a normal brain MRI image as a registration template, which has a corresponding brain region segmentation template. By registering the patient's brain CT scan with the MRI template, the hematoma characteristics are extracted, enabling rapid and effective analysis of the hemorrhage magnitude and the brain regions involved in the hematoma. This provides guidance for doctors and can be used for subsequent visualization modeling, helping them develop more comprehensive treatment plans.
[0007] To achieve the above objectives, the present disclosure adopts the following technical solution:
[0008] The first aspect of this disclosure provides a method for analyzing cerebral hemorrhage based on image registration.
[0009] A method for analyzing cerebral hemorrhage based on image registration includes the following steps:
[0010] Segmenting brain hemorrhage images from brain CT images;
[0011] Preprocessing of brain CT images and segmented brain hemorrhage images;
[0012] The preprocessed brain CT images were registered with the brain MRI template to obtain the deformation field;
[0013] The deformation field generated by registration is applied to the preprocessed brain hemorrhage image;
[0014] Information was extracted from the brain hemorrhage images before and after preprocessing.
[0015] As some optional implementations, segmenting brain hemorrhage images from brain CT images includes: training a deep neural network and using the deep neural network to automatically segment the brain hemorrhage region.
[0016] As some optional implementations, preprocessing of brain CT images includes: skull removal, resampling, and normalization; preprocessing of segmented brain hemorrhage images includes: resampling and normalization.
[0017] As some optional implementation methods, the process of registering the preprocessed brain CT image with the brain MRI template to obtain the deformation field is specifically as follows: the brain CT image and the brain MRI template are registered by first rigidly registering and aligning the image center, and then by non-rigid registration, so as to obtain the deformation field of the brain CT image during the registration process.
[0018] As some optional implementations, applying the deformation field to the preprocessed brain hemorrhage image specifically involves inputting the deformation field and the preprocessed brain hemorrhage image into a graphics transformation processor to transform the image.
[0019] As some optional implementation methods, the information extraction of the brain hemorrhage images before and after preprocessing is as follows: extracting the brain hemorrhage magnitude information from the brain hemorrhage images segmented from the brain CT images; and extracting the brain regions involved in the brain hemorrhage from the preprocessed brain hemorrhage images based on the MRI template.
[0020] As some optional implementation methods, in the segmented brain hemorrhage image, the number of voxel values with a gray value of 1 is counted as the order of hemorrhage; in the preprocessed brain hemorrhage image, the position of voxel values with a gray value of 1 is counted, and the brain structure represented by the same position is found on the MRI template to obtain the brain region involved in the brain hemorrhage.
[0021] The second aspect of this disclosure provides a brain hemorrhage analysis system based on image registration.
[0022] A brain hemorrhage analysis system based on image registration, comprising:
[0023] The image segmentation module is configured to segment brain hemorrhage images from brain CT images;
[0024] The preprocessing module is configured to preprocess brain CT images and segmented brain hemorrhage images;
[0025] The registration module is configured to register the preprocessed brain CT image with the brain MRI template to obtain the deformation field;
[0026] The application module is configured to apply the deformation field to the preprocessed brain hemorrhage image;
[0027] Information extraction is configured to extract information from brain hemorrhage images before and after preprocessing.
[0028] A third aspect of this disclosure provides a computer-readable storage medium having a program stored thereon that, when executed by a processor, implements the steps of an image registration-based brain hemorrhage analysis method as described in the first aspect of this disclosure.
[0029] The fourth aspect of this disclosure provides an electronic device including a memory, a processor, and a program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of an image registration-based brain hemorrhage analysis method as described in the first aspect of this disclosure.
[0030] Compared with the prior art, the beneficial effects of this disclosure are:
[0031] This disclosure first trains a deep learning network, improves the model structure and parameters based on sample characteristics, and accurately and quickly segments the cerebral hemorrhage area from brain CT scans. Necessary preprocessing operations are performed on the brain CT scans and corresponding hemorrhage images, followed by matching and transformation with a brain MRI template to extract hematoma features and summarize disease information. This disclosure can quickly extract cerebral hemorrhage features from brain CT scans and indicate the hemorrhage magnitude and location in the brain, allowing doctors to more intuitively understand the patient's cerebral hemorrhage condition and formulate a more comprehensive treatment plan. Furthermore, it helps patients and their families understand the progression of the disease. Attached Figure Description
[0032] The accompanying drawings, which form part of this disclosure, are used to provide a further understanding of this disclosure. The illustrative embodiments of this disclosure and their descriptions are used to explain this disclosure and do not constitute an undue limitation of this disclosure.
[0033] Figure 1 This is a flowchart illustrating the image registration-based brain hemorrhage analysis method provided in Embodiment 1 of this disclosure.
[0034] Figure 2 This is a schematic diagram of a brain CT image provided in Embodiment 1 of this disclosure.
[0035] Figure 3 This is a schematic diagram of a preprocessed brain CT image provided in Embodiment 1 of this disclosure.
[0036] Figure 4 This is a schematic diagram of a brain MRI template provided in Embodiment 1 of this disclosure.
[0037] Figure 5 This is a schematic diagram of an MRI template representing 35 brain structures provided in Embodiment 1 of this disclosure.
[0038] Figure 6 This is a schematic diagram of a registered brain CT scan provided in Embodiment 1 of this disclosure. Detailed Implementation
[0039] The present disclosure will be further described below with reference to the accompanying drawings and embodiments.
[0040] It should be noted that the following detailed descriptions are exemplary and intended to provide further illustration of this disclosure. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure pertains.
[0041] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the exemplary embodiments according to this disclosure. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms “comprising” and / or “including” are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.
[0042] Where there is no conflict, the embodiments and features described herein can be combined with each other.
[0043] Example 1:
[0044] Embodiment 1 of this disclosure provides a method for analyzing cerebral hemorrhage based on image registration.
[0045] A method for analyzing cerebral hemorrhage based on image registration includes the following steps:
[0046] Segmenting brain hemorrhage images from brain CT images;
[0047] Preprocessing of brain CT images and segmented brain hemorrhage images;
[0048] The preprocessed brain CT images were registered with the brain MRI template to obtain the deformation field;
[0049] Apply the deformation field to the preprocessed brain hemorrhage image;
[0050] Information was extracted from the brain hemorrhage images before and after preprocessing.
[0051] As some optional implementations, segmenting brain hemorrhage images from brain CT images includes: training a deep neural network and using the deep neural network to automatically segment the brain hemorrhage region.
[0052] As some optional implementations, preprocessing of brain CT images includes: skull removal, resampling, and normalization; preprocessing of segmented brain hemorrhage images includes: resampling and normalization.
[0053] As some optional implementation methods, the process of registering the preprocessed brain CT image with the brain MRI template to obtain the deformation field is specifically as follows: the brain CT image and the brain MRI template are registered by first rigidly registering and aligning the image center, and then by non-rigid registration, so as to obtain the deformation field of the brain CT image during the registration process.
[0054] As some optional implementations, applying the deformation field to the preprocessed brain hemorrhage image specifically involves inputting the deformation field and the preprocessed brain hemorrhage image into a graphics transformation processor to transform the image.
[0055] As some optional implementation methods, the information extraction of the brain hemorrhage images before and after preprocessing is as follows: extracting the brain hemorrhage magnitude information from the brain hemorrhage image segmented from the brain CT image; and extracting the brain region information involved in the brain hemorrhage from the brain hemorrhage image after preprocessing, based on the MRI template.
[0056] As some optional implementation methods, in the segmented brain hemorrhage image, the number of voxel values with a gray value of 1 is counted as the order of hemorrhage; in the preprocessed brain hemorrhage image, the position of voxel values with a gray value of 1 is counted, and the brain structure represented by the same position is found on the MRI template to obtain the brain region involved in the brain hemorrhage.
[0057] (1) Segmenting brain hemorrhage images from brain CT images
[0058] First, a deep neural network is trained to automatically segment the brain hemorrhage area. The process of segmenting the brain hemorrhage area from brain CT images using a trained deep neural network can be broadly divided into three parts: feature extraction, enhanced feature extraction, and classification output.
[0059] For our data, we first divided the training and validation sets in a 7:3 ratio. Then, we trained the model and improved its parameters, using manually segmented samples from doctors. The validation set provided model evaluation, keeping the error within acceptable limits. Afterward, when a brain CT scan was input, the network could automatically segment the brain hemorrhage area. Figure 2 As shown, the brain hemorrhage area is highlighted after a sample is segmented.
[0060] (2) Preprocessing brain CT images and segmented brain hemorrhage images
[0061] Preprocessing of brain CT images includes: skull removal, resampling, and normalization; preprocessing of segmented brain hemorrhage images includes: resampling and normalization.
[0062] Before registration, the brain CT scan needs to undergo preprocessing including skull removal, resampling, and normalization. This is because the brain MRI template used during registration only contains images of the internal brain tissue, lacking images of the skull and surrounding soft tissues such as the skin. To improve registration accuracy, the skull needs to be removed from the brain CT images, retaining only the internal brain tissue images. The principle behind this is primarily based on threshold segmentation. CT can effectively distinguish tissues with density differences and has high spatial resolution; therefore, the CT values of bone images in CT data differ significantly from those of the surrounding soft tissues. Thus, a threshold segmentation algorithm is chosen to detect the skull from CT images.
[0063] The specific implementation method involves using the Python interface of the 3DSlicer software to write a Python script suitable for 3DSlicer to automatically remove the skull and lateral soft tissue from brain CT images, improving registration accuracy. The process is as follows: First, a segmentation module is created to store the results. The HU value range of the skull is determined, and the skull is segmented. Next, the maximum connected component method is used to remove error points caused by noise. Finally, the remaining mask is extracted and stored as output to disk for subsequent operations. By flexibly calling modules within 3DSlicer, the skull and its external soft tissue in brain CT images are removed, leaving only the internal brain tissue image for further processing.
[0064] Since the brain MRI template image used in the registration process has dimensions of x:160, y:192, z:224, and the brain CT template has dimensions of x:512, y:512, z:165, the CT image needs to be resampled before registration to make the dimensions consistent. This is achieved by using the resampling function of the SimpleITK software package.
[0065] Before registration, a normalization operation is required, which reduces the grayscale values of the brain CT image to between 0 and 1 by a certain ratio, which is beneficial for subsequent calculations. The preprocessed brain CT image is shown in Figure 3.
[0066] (3) Register the preprocessed brain CT images with the brain MRI template to obtain the deformation field.
[0067] A rigid registration method was used to align the image centers first, followed by non-rigid registration, to register brain CT images with brain MRI templates, thereby obtaining the deformation field of the brain CT images during the registration process.
[0068] Medical image registration technology was used to match CT images and brain MRI images. The brain MRI template refers to using a normal person's brain MRI image as the registration template. Image registration is the task of matching two or more images, such as two images taken at different times, angles, and locations. Medical image registration involves finding a set transformation method to make the spatial points of two medical images identical. In simple terms, image registration is an operation to align and match two different images; only in this way can the comparison of two images be meaningful. Based on the characteristics of our data, we chose to apply rigid registration first, followed by non-rigid registration, to transform the images. The specific implementation tool was the SimpleElastix medical registration library. SimpleElastix is a toolkit integrated into SimpleITK, which includes the Elastix C++ registration library methods, providing a series of methods such as registration, iteration, and transformation. We compiled and generated it according to the manual and used it in Python to assist our work.
[0069] During the registration process, we selected the CT image as the moving image and the MRI image as the fixed image. The moving image was iterated continuously to better match the fixed image. Since our CT and MRI images have different origins and orientations, we first chose a rigid transformation method to translate and rotate the original image without changing its size or internal curves. Specifically, we called the function `sitk.GetDefaultParameterMap()`, inputting the parameter 'translation' to obtain a parameter map, which was then input into our `sitk.ElastixImageFilter()` image processor to complete the rigid transformation, resulting in an image aligned with the reference image. Subsequently, we performed a non-rigid transformation to completely match our CT image onto the reference brain MRI image. The non-rigid transformation was implemented using `sitk.GetDefaultParameterMap()`, inputting the parameter 'bspline', and then inputting it into the image processor. To achieve better local registration results rather than just focusing on the general outline alignment, we used a multi-resolution strategy to create a resolution pyramid during the registration process. Each resolution level underwent the maximum number of iterations to find the best result. We use mutual information as the evaluation metric for each iteration, and finally obtain the registration image with the highest processor score as our output image. In addition, to allow the final graphics processor to focus more on finding and optimizing local information, we added a parametric affine between these two steps to roughly adjust the curve of the aligned image after rigidity changes. Experiments show that this makes the mutual information of the final output more correlated and the effect more obvious.
[0070] (4) Applying the deformation field to the preprocessed brain hemorrhage image
[0071] First, the brain hemorrhage images segmented by the deep network need to undergo resampling and normalization preprocessing to maintain image dimensions consistent with MRI, similar to the preprocessing of brain CT images. Then, the brain hemorrhage images are input into the image processor in the program. This image processor calls the function sitk.TransformixImageFilter() to obtain the image deformation field obtained during registration. This field includes the contour and internal curve changes of the brain CT image before and after registration. Because the brain hemorrhage image and the brain CT image have a one-to-one correspondence, the same transformation is performed on the brain hemorrhage image based on the parameters of the deformation field. The two transformed images also maintain a one-to-one correspondence. Therefore, the brain hemorrhage image corresponds to the MRI template image, allowing analysis methods to be implemented on their superimposed images.
[0072] (5) Information extraction was performed on the brain hemorrhage images before and after preprocessing.
[0073] From the segmented brain hemorrhage images in brain CT images, the magnitude of brain hemorrhage is extracted. From the preprocessed brain hemorrhage images, the brain regions involved in the brain hemorrhage are extracted based on the MRI template.
[0074] In the segmented brain hemorrhage image, the number of voxel values with a gray value of 1 is counted as the magnitude of the hemorrhage; in the preprocessed brain hemorrhage image, the location of voxel values with a gray value of 1 is counted, and the brain structures represented by the same location are found on the MRI template to obtain the brain regions involved in the brain hemorrhage.
[0075] like Figure 5 As shown, the magnitude of cerebral hemorrhage is first extracted. To reduce errors, the segmented cerebral hemorrhage image is used instead of the preprocessed cerebral hemorrhage image. The number of voxel points with a gray value of 1 in the cerebral hemorrhage image is counted as the magnitude of the patient's cerebral hemorrhage for doctors' reference.
[0076] It is necessary to analyze the brain regions involved in the cerebral hemorrhage, as hemorrhages in important brain structures may be accompanied by more serious complications. The method involves overlaying preprocessed cerebral hemorrhage images with MRI images representing 35 brain structural regions to extract information about the brain regions involved in the hemorrhage. Since the preprocessed cerebral hemorrhage images and the brain MRI template images correspond, the positions of voxels with a grayscale value of 1 in the cerebral hemorrhage image are counted. At the same position, the grayscale value of the brain MRI template image is examined; different grayscale values correspond to 35 different brain structures. Statistical methods are used to obtain information about the patient's condition.
[0077] Example 2:
[0078] Embodiment 2 of this disclosure provides a brain hemorrhage analysis system based on image registration.
[0079] A brain hemorrhage analysis system based on image registration, comprising:
[0080] The image segmentation module is configured to segment brain hemorrhage images from brain CT images;
[0081] The preprocessing module is configured to preprocess brain CT images and segmented brain hemorrhage images;
[0082] The registration module is configured to register the preprocessed brain CT image with the brain MRI template to obtain the deformation field;
[0083] The application module is configured to apply the deformation field to the preprocessed brain hemorrhage image;
[0084] Information extraction is configured to extract information from brain hemorrhage images before and after preprocessing.
[0085] Example 3:
[0086] Embodiment 3 of this disclosure provides a computer-readable storage medium having a program stored thereon that, when executed by a processor, implements the steps of a brain hemorrhage analysis method based on image registration as described in the first aspect of this disclosure.
[0087] The detailed steps are the same as the system operation method provided in Example 1, and will not be repeated here.
[0088] Example 4:
[0089] Embodiment 4 of this disclosure provides an electronic device, including a memory, a processor, and a program stored in the memory and executable on the processor. When the processor executes the program, it implements the steps in the image registration-based brain hemorrhage analysis method described in the first aspect of this disclosure.
[0090] The detailed steps are the same as the system operation method provided in Example 1, and will not be repeated here.
[0091] Those skilled in the art will understand that embodiments of this disclosure can be provided as methods, systems, or computer program products. Therefore, this disclosure can take the form of hardware embodiments, software embodiments, or embodiments combining software and hardware aspects. Furthermore, this disclosure can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage and optical storage) containing computer-usable program code.
[0092] This disclosure is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this disclosure. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can 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 a machine for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.
[0093] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.
[0094] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.
[0095] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.
[0096] The above description is merely a preferred embodiment of this disclosure and is not intended to limit this disclosure. Various modifications and variations can be made to this disclosure by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.
Claims
1. A method for analyzing cerebral hemorrhage based on image registration, characterized in that, include: Segmenting brain hemorrhage images from brain CT images includes: training a deep neural network and using the deep neural network to automatically segment the brain hemorrhage region; Preprocessing of brain CT images and segmented brain hemorrhage images; The preprocessed brain CT images were registered with the brain MRI template to obtain the deformation field; Apply the deformation field to the preprocessed brain hemorrhage image; Information was extracted from the brain hemorrhage images before and after preprocessing. The magnitude of brain hemorrhage was extracted from the brain hemorrhage images segmented from brain CT images. The brain regions involved in the brain hemorrhage were extracted from the brain hemorrhage images after preprocessing based on MRI templates. In the segmented brain hemorrhage image, the number of voxel values with a gray value of 1 is counted as the magnitude of the hemorrhage; in the preprocessed brain hemorrhage image, the location of voxel values with a gray value of 1 is counted, and the brain structures represented by the same location are found on the MRI template to obtain the brain regions involved in the brain hemorrhage.
2. The method for analyzing cerebral hemorrhage based on image registration as described in claim 1, characterized in that, Preprocessing of brain CT images includes: skull removal, resampling, and normalization; preprocessing of segmented brain hemorrhage images includes: resampling and normalization.
3. The method for analyzing cerebral hemorrhage based on image registration as described in claim 1, characterized in that, The process of registering the preprocessed brain CT image with the brain MRI template to obtain the deformation field involves: first using rigid registration to align the image center, and then using non-rigid registration to register the brain CT image with the brain MRI template, thereby obtaining the deformation field of the brain CT image during the registration process.
4. The method for analyzing cerebral hemorrhage based on image registration as described in claim 1, characterized in that, The process of applying the deformation field to the preprocessed brain hemorrhage image specifically involves inputting the deformation field and the preprocessed brain hemorrhage image into a graphics transformation processor to transform the image.
5. A brain hemorrhage analysis system based on image registration, employing a brain hemorrhage analysis method based on image registration as described in any one of claims 1-4, characterized in that, The image segmentation module is configured to segment brain hemorrhage images from brain CT images; The preprocessing module is configured to preprocess brain CT images and segmented brain hemorrhage images; The registration module is configured to register the preprocessed brain CT image with the brain MRI template to obtain the deformation field; The application module is configured to apply the deformation field to the preprocessed brain hemorrhage image; Information extraction is configured to extract information from brain hemorrhage images before and after preprocessing.
6. A computer-readable storage medium having a program stored thereon, characterized in that, When executed by a processor, the program implements the steps in the image registration-based brain hemorrhage analysis method as described in any one of claims 1-4.
7. An electronic device comprising a memory, a processor, and a program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps in the image registration-based brain hemorrhage analysis method as described in any one of claims 1-4.