Method and device for measuring eye penetration degree based on MRI image

Abstract

The invention relates to the technical field of image processing and eye penetration degree measurement, and provides an MRI image-based eye penetration degree measurement method and device. The method comprises the following steps of: obtaining an orbit MRI (Magnetic Resonance Imaging) transection image, and determining the foremost edge positions of zygomatic processes on two sides, the gravity center position of eyeballs and the front edge position of cornea according to the orbit MRI transection image; and calculating a first distance from the position of the center of gravity of the eyeball to the connecting line of the foremost edges of the zygomatic processes on the two sides and a second distance from the position of the center of gravity of the eyeball to the position of the front edge of the cornea, and taking the sum of the first distance and the second distance as an eye protrusion degree measurement result. And the safety and the accuracy of the measurement of the eye penetration degree can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing and exophthalmos measurement, and in particular to a method and device for measuring exophthalmos based on MRI images. Background Art

[0002] At present, exophthalmos refers to the vertical distance between the apex of the cornea and the outer edge of the bilateral orbits. It is an important indicator for measuring the degree of protrusion of the eyeballs. In the related art, the measurement of exophthalmos is usually carried out using a Hertel exophthalmometer. However, this measurement method has some potential defects and limitations. For example, this measurement method has problems of poor repeatability and consistency. Different operators may obtain different measurement results using the Hertel exophthalmometer, which is related to the operator's skills and experience. In addition, the patient's degree of cooperation during the examination, such as whether he can look straight ahead stably, may also affect the accuracy of the measurement.

[0003] Some literature also provides a method for measuring exophthalmos based on CT (computed tomography) imaging examination. CT scanning uses X-ray technology to penetrate the human body through X-rays at different angles. The detector then receives the signal and generates a cross-sectional image after computer processing. Compared with the Hertel exophthalmosometer measurement method, this method improves the measurement accuracy and has better consistency.

[0004] However, due to the radiation characteristics of CT, its application in pregnant women and children is relatively limited. In addition, CT has low contrast in soft tissue imaging, which may have a certain impact on the accuracy of the measurement results. Magnetic resonance imaging (MRI) is a medical imaging technology that can provide high-definition views of soft tissues and organs. MRI uses strong magnetic fields and radio frequency pulses to resonate hydrogen nuclei (protons) in the body, then detects the attenuation of the resonance signal, and generates images after computer processing. Based on the above principles, MRI does not involve radiation and is relatively safe for the human body. In addition, MRI also has higher contrast and resolution in soft tissue imaging. However, there is currently no solution for measuring exophthalmos based on MRI images. Summary of the invention

[0005] The present invention provides a method and device for measuring exophthalmos based on MRI images, so as to realize the measurement of exophthalmos based on magnetic resonance imaging, thereby further improving the safety and accuracy of exophthalmos measurement.

[0006] In a first aspect, the present invention provides a method for measuring exophthalmos based on MRI images, comprising:

[0007] Obtain transverse MRI images of the orbit;

[0008] The most anterior edge of the zygomatic processes on both sides was determined based on the transverse MRI images of the orbit;

[0009] Determine the center of gravity of the eyeball based on the transverse MRI image of the orbit;

[0010] The position of the anterior corneal edge was determined based on the transverse orbital MRI images;

[0011] Calculate the first distance from the center of gravity of the eyeball to the line connecting the most anterior edges of the zygomatic processes on both sides, and the second distance from the center of gravity of the eyeball to the anterior edge of the cornea, and take the sum of the first distance and the second distance as the exophthalmos measurement result.

[0012] In a second aspect, the present invention provides a device for measuring exophthalmos based on MRI images, comprising:

[0013] An acquisition unit, used for acquiring a transverse MRI image of the orbit;

[0014] The first determination unit is used to determine the most anterior edge position of the zygomatic processes on both sides according to the transverse MRI image of the orbit;

[0015] The second determination unit is used to determine the center of gravity of the eyeball according to the transverse MRI image of the orbit;

[0016] A third determination unit is used to determine the position of the corneal front edge according to the orbital MRI transverse image;

[0017] The calculation unit is used to calculate the first distance from the center of gravity of the eyeball to the line connecting the most anterior edges of the zygomatic processes on both sides, and the second distance from the center of gravity of the eyeball to the anterior edge of the cornea, and use the sum of the first distance and the second distance as the exophthalmos measurement result.

[0018] In addition, the present invention also provides a terminal, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the steps of the method for measuring exophthalmos based on MRI images in the first aspect are implemented.

[0019] The present invention provides a method and device for measuring exophthalmos based on MRI images, by obtaining an orbital MRI transverse image, and determining the most front edge position of the zygomatic processes on both sides, the center of gravity position of the eyeball, and the front edge position of the cornea according to the orbital MRI transverse image; by calculating the first distance from the center of gravity position of the eyeball to the line connecting the most front edge position of the zygomatic processes on both sides, and the second distance from the center of gravity position of the eyeball to the front edge position of the cornea, and finally taking the sum of the first distance and the second distance as the exophthalmos measurement result. It can be seen that the present invention realizes the measurement of exophthalmos based on orbital MRI images. On the one hand, since MRI does not contain radiation, it improves safety compared to the CT method; on the other hand, since MRI has higher contrast and resolution in the imaging of soft tissues, the exophthalmos measurement result obtained based on MRI images is also more accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0021] Figure 1 Schematic diagram of an application scenario of a method for measuring exophthalmos based on MRI images provided in an embodiment of the present invention;

[0022] Figure 2 It is a flow chart of the implementation of the method for measuring exophthalmos based on MRI images provided by an embodiment of the present invention;

[0023] Figure 3 is a schematic diagram of a 2D sequence image and a 3D sequence image of an eyeball provided by an embodiment of the present invention;

[0024] Figure 4 Schematic diagram of the positional relationship between the most anterior edge positions of the zygomatic processes on both sides, the anterior edge positions of the cornea and the center of gravity of the eyeball provided by an embodiment of the present invention;

[0025] Figure 5 It is a schematic structural diagram of an MRI image-based exophthalmos measurement device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0026] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present invention. However, it should be clear to those skilled in the art that the present invention may be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present invention.

[0027] In order to make the purpose, technical solutions and advantages of the present invention more clear, specific embodiments will be described below in conjunction with the accompanying drawings.

[0028] Figure 1 Schematic diagram of application scenario of the method for measuring exophthalmos based on MRI images provided by an embodiment of the present invention. Figure 1 As shown, the present invention can be applied to a measurement terminal, which performs key point recognition and position positioning on the input orbital MRI transverse image, including positioning the most front edge position of the zygomatic processes on both sides, positioning the center of gravity position of the eyeball, and positioning the front edge position of the cornea, and then calculates a first distance from the center of gravity position of the eyeball to the line connecting the most front edge positions of the zygomatic processes on both sides, and a second distance from the center of gravity position of the eyeball to the front edge position of the cornea, and uses the sum of the first distance and the second distance as the exophthalmos measurement result.

[0029] It should be noted that on the MRI axial image, the vertical distance from the vertex of the anterior corneal edge to the line connecting the vertices of the bony orbital rim is measured as the baseline to obtain the degree of exophthalmos. In actual application, the vertex of the bony orbital rim is actually the anteriormost position of the zygomatic process, and there is one anteriormost position of the zygomatic process on each side. The vertical distance from the vertex of the anterior corneal edge to the line connecting the anteriormost positions of the zygomatic processes on both sides is the degree of exophthalmos.

[0030] The present invention realizes the measurement of exophthalmos based on orbital MRI images. On the one hand, since MRI does not contain radiation, it is safer than CT. On the other hand, since MRI has higher contrast and resolution in soft tissue imaging, the exophthalmos measurement result obtained based on MRI images is also more accurate.

[0031] It should be pointed out that the advantage of CT images is that they can show bones and calcifications, and the contrast of soft tissue imaging is lower than that of MRI. The measurement of exophthalmos is mainly aimed at the orbital area, and more soft tissue related images are involved. Since MRI images have better soft tissue resolution, it is easier to distinguish between the anterior edge of the cornea and the closed eyelids. Therefore, the identification and positioning of key points (the most anterior edge position of the zygomatic processes on both sides, the center of gravity position of the eyeball, and the anterior edge position of the cornea) are more accurate, so it has higher accuracy.

[0032] In addition, it should be pointed out that when performing CT or MRI imaging examinations, the target needs to keep looking straight ahead for a certain period of time. If the target has strabismus, the rotation of the eyeball will cause certain errors. The method for measuring exophthalmos based on MRI images of the present invention can also effectively correct MRI images with strabismus, and obtain more accurate exophthalmos measurement results. The specific principles and processes are detailed in the following embodiments.

[0033] See also Figure 2, which shows a flowchart of the implementation of the method for measuring exophthalmos based on MRI images provided by an embodiment of the present invention, which is described in detail as follows:

[0034] Step 201, obtaining a transverse MRI image of the orbit;

[0035] Step 202, determining the most anterior edge positions of the zygomatic processes on both sides according to the transverse MRI image of the orbit;

[0036] Step 203, determining the center of gravity of the eyeball according to the transverse MRI image of the orbit;

[0037] Step 204, determining the position of the corneal anterior edge according to the orbital MRI transverse image;

[0038] Step 205, calculating a first distance from the center of gravity of the eyeball to a line connecting the most anterior edges of the zygomatic processes on both sides, and a second distance from the center of gravity of the eyeball to the anterior edge of the cornea, and taking the sum of the first distance and the second distance as the exophthalmos measurement result.

[0039] The following describes the specific implementation methods of the above steps one by one:

[0040] First, in step 201, a transverse MRI image of the orbit is acquired.

[0041] The transverse plane, also known as the horizontal plane, refers to a cross-section parallel to the ground, similar to a slice image made in the direction from the top of the head to the soles of the feet.

[0042] In MRI scanning, scanning in different planes is usually performed to obtain comprehensive diagnostic information. Conventional scanning orientations include transverse, sagittal or coronal. The present invention identifies and locates key points (including the most anterior edge positions of the zygomatic processes on both sides, the center of gravity position of the eyeball, and the anterior edge position of the cornea) based on transverse images to achieve the measurement of exophthalmos.

[0043] In the embodiment of the present invention, the MRI transverse image can be a 2D (two-dimensional) sequence image or a 3D (three-dimensional) sequence image, and each has its own advantages and disadvantages.

[0044] 2D sequence images display internal structures through two-dimensional sections. Common 2D sequences include T1-weighted imaging (T1WI), T2-weighted imaging (T2WI), diffusion-weighted imaging (DWI), etc. 2D sequence scanning time is relatively short, suitable for rapid evaluation and preliminary diagnosis, and usually has a high signal-to-noise ratio (SNR), good image quality, suitable for observing subtle structures, but 2D sequences cannot provide complete three-dimensional spatial information, and may have limitations for a comprehensive understanding of complex structures.

[0045] 3D sequence images can provide a more three-dimensional and comprehensive perspective through three-dimensional data acquisition. Common 3D sequences include three-dimensional magnetization intensity prepared gradient echo sequence (3D-MPRAGE), three-dimensional fast spin echo sequence (3D-FSE), etc. 3D sequences can provide higher spatial resolution and more detailed anatomical structure display, especially suitable for the detection of key points of fine structure display. 3D sequence data can be reconstructed in multiple planes (MPR), observe lesions from different angles, and provide more information. The scanning time of 3D sequence is relatively long, and the cooperation requirements of the detector are higher, and the requirements for equipment are also higher. In some cases, 3D sequence can obtain images of multiple sections through a single scan, which can reduce the number of scans.

[0046] In an embodiment of the present invention, before identifying and locating key points, the MRI transverse image may be pre-processed, for example, the eyeball may be automatically segmented in the MRI transverse image to achieve more accurate and rapid post-positioning and identification.

[0047] Specifically, the automatic segmentation of the eyeball can include automatic segmentation based on 2D images and automatic segmentation based on 3D images. The specific implementation method is to construct a deep learning network model VB-Net as an automatic eyeball segmentation model, and use the collected data set to train the automatic eyeball segmentation model, so that the model has the ability to automatically segment the eyeball on MRI cross-sectional images.

[0048] For example, for 2D sequence images, 2D sequence orbital MRI transverse images from 109 patients were prospectively collected, axial T2 WIDIXON isophase (voxel 0.5×0.5×2.5mm, dimension 320×320×20), and the eyeball was manually segmented layer by layer using Slicer to obtain a training data set. Then the eyeball segmentation model was trained, the parameters were set to a sampling interval of 0.5×0.5×2.5, the cropping size was 64×64×64, the image was preprocessed by the nearest neighbor interpolation method and adaptive normalization, and then the pre-built deep learning network model VB-Net was trained. Finally, the automatically segmented images were checked one by one, and the volume of the eyeball under manual segmentation and automatic segmentation was compared to evaluate the segmentation performance of the model, and finally a trained automatic segmentation model capable of automatically segmenting the eyeball for MRI transverse images was obtained.

[0049] Exemplarily, for 3D serial images, 3D T2WI serial MRI transverse images (voxel 0.5×0.5×0.5, dimension 320×320×80) were prospectively collected from 109 patients, and semi-automatic segmentation was performed using 3DSlicer. The parameters were set to a sampling interval of 0.5×0.5×2.5, and a cropping size of 64×64×64. Image preprocessing was performed using nearest neighbor interpolation and adaptive normalization, and the pre-built deep learning network model VB-Net was trained. Finally, the automatically segmented images were checked one by one, and the volumes of the eyeballs under manual segmentation and automatic segmentation were compared to evaluate the segmentation efficiency of the model. Finally, a trained automatic segmentation model capable of automatically segmenting the eyeballs for MRI transverse images was obtained.

[0050] In an embodiment of the present invention, an automatic eyeball segmentation model is used to automatically segment the orbital MRI transverse image to obtain an eyeball image. Later, key point recognition and positioning can be achieved based on the segmented eyeball image, thereby greatly reducing the complexity of image processing and improving recognition efficiency.

[0051] In step 202, the most anterior edge positions of the zygomatic processes on both sides are determined based on the orbital MRI transverse image.

[0052] In this embodiment, the anterior edge positions of the zygomatic processes on both sides are identified and located based on the orbital MRI transverse image. First, the orbital MRI transverse image is automatically segmented by the above-mentioned automatic eye segmentation model to obtain the eye image. Then, the trained zygomatic process anterior edge position recognition model is used to identify the anterior edge positions of the zygomatic processes on both sides on the eye image.

[0053] Figure 3 Schematic diagram of 2D sequence images and 3D sequence images (actually 3D, the plane diagram cannot truly show the 3D effect) of the eyeball, as shown in Figure 3 As shown, the orbital MRI transverse image includes multiple layer images.

[0054] Specifically, for 2D sequence images, the largest layer image of the eyeball with the largest cross-sectional area can be extracted based on the cross-sectional area of ​​each layer of the eyeball image in the orbital MRI transverse image; the largest layer image of the eyeball is input into the trained zygomatic process anterior edge position recognition model to obtain the largest layer image of the eyeball with the anterior edge positions of the zygomatic processes on both sides marked.

[0055] Specifically, for 3D sequence images, each eyeball layer image in the orbital MRI transverse image can be input into the trained zygomatic process anterior edge position recognition model to obtain multiple anterior edge positions corresponding to each eyeball layer image; the anterior edge positions of the zygomatic processes on both sides are selected from the multiple anterior edge positions corresponding to each eyeball layer image.

[0056] It should be pointed out that for 2D sequence images, the most anterior edge position of the zygomatic processes on both sides can be considered to be located on the largest layer image of the eyeball (i.e., the layer image with the largest eyeball area on each layer image). The most anterior edge position of the zygomatic processes on both sides can be obtained by directly selecting the largest layer image of the eyeball and using the trained zygomatic process most anterior edge position recognition model for recognition. For 3D sequence images, since one more dimension is considered, the most anterior edge position of the zygomatic processes on both sides on the largest layer image of the eyeball may not be the most anterior edge position. Therefore, it is necessary to use the trained zygomatic process most anterior edge position recognition model to recognize each layer image to obtain multiple anterior edge positions, and then select the most anterior edge position of the zygomatic processes on both sides from the multiple anterior edge positions. Among them, the distinction between the anterior edge and the most anterior edge can be obtained based on the same three-dimensional rectangular coordinate system according to the comparison of three-dimensional coordinates.

[0057] For example, a 2D sequence of orbital MRI transverse images includes 20 layers, and the largest layer image of the eyeball with the largest eyeball cross-section needs to be extracted from the 20 layers of images. The anterior edge position of the zygomatic process on both sides can be obtained by performing a single recognition using the anterior edge position recognition model of the zygomatic process.

[0058] For example, the 3D sequence of orbital MRI transverse images includes 80 layers, and the zygomatic process anterior edge position recognition model needs to be used to recognize the 80 layers of images respectively, to obtain 80 bilateral zygomatic process anterior edge positions, and then select one bilateral zygomatic process anterior edge position from the 80 bilateral zygomatic process anterior edge positions. Of course, in order to improve the recognition efficiency, several layers including the eyeball area can also be selected from the 80 layers of images, and then the bilateral zygomatic process anterior edge positions are only recognized for the several layers including the eyeball area.

[0059] In an embodiment of the present invention, the method of using a trained machine learning model to automatically identify and determine the anterior edge position of the zygomatic processes on both sides can be specifically as follows: input the image of the largest plane of the eyeball into the trained zygomatic process anterior edge position recognition model to obtain the image of the largest plane of the eyeball with the anterior edge position of the zygomatic processes on both sides marked.

[0060] Among them, the model for identifying the anterior edge position of the zygomatic process is constructed based on a hybrid model, which includes a deep residual network for feature extraction and a U-Net convolutional neural network for image segmentation; the training data source of the hybrid model is a certain number of orbital MRI transverse images with the anterior edge positions of the zygomatic processes on both sides marked.

[0061] In an embodiment of the present invention, whether in a training process or an actual measurement process, before the image of the largest aspect of the eyeball is input into the anterior edge position recognition model of the zygomatic process, the image of the largest aspect of the eyeball can be preprocessed first. Specifically, the image of the largest aspect of the eyeball is traversed from front to back to obtain the first target coordinate whose value is greater than a preset value; the image of the largest aspect of the eyeball is cropped based on the target coordinate to reduce the image of the largest aspect of the eyeball to a preset size.

[0062] In this embodiment, since the front edge of the zygomatic process and the characteristic part for identifying the front edge of the zygomatic process are located at a relatively forward position of the image, the image can be traversed from front to back and the first large numerical coordinate can be returned. The preset value can be 1000 for example. The image can be cropped based on this coordinate. After cropping, the size of the data image can be changed from 320×320 to 120×320. Through the above image preprocessing steps, the target range can be narrowed, the accuracy of eye segmentation can be improved, and the computing power burden of the model can be reduced, the model operation efficiency can be improved, and it is also convenient for the model to locate the characteristic parts.

[0063] In this embodiment, the training process of the model requires data annotation, that is, the position of the zygomatic process is annotated in the MRI image. However, since the position of the most front edge of the zygomatic process required for data annotation is the coordinate of a pixel point, this target is too small for the naked eye, which brings difficulties to the process of data annotation, so a method of annotating with different markers is adopted. Specifically, the method of annotating the most front edge of the zygomatic process in the image of the largest level of the eyeball can be regional annotation; correspondingly, the step of obtaining the most front edge position of the zygomatic process from the image of the largest level of the eyeball with regional annotation can include: taking the center of gravity of the two marked areas respectively, and obtaining two pixel coordinates indicating the most front edge position of the zygomatic process. Specifically, K-means can be used to cluster the marked data to obtain two marked areas, and then the center of gravity of the two marked areas is taken respectively, and finally the two required marking points are obtained.

[0064] In a specific embodiment, the deep residual network used for feature extraction includes ResNet34; the number of input channels of the first convolution layer of ResNet34 is 1.

[0065] The model architecture proposed in this example is a hybrid of ResNet34 and U-Net, designed specifically for medical image segmentation tasks. The model makes full use of PyTorch's deep learning capabilities and is particularly suitable for processing magnetic resonance imaging. The architecture consists of a ResNet34-based encoder (for feature extraction) and a customized decoder (for segmentation tasks) as well as a dedicated dataset class for processing MRI data.

[0066] In this embodiment, ResNet34 is selected as the encoder. ResNet34 was originally known for its performance in image classification, which brings powerful feature extraction capabilities. However, MRI images are usually single-channel images, which is different from the three-channel RGB images that ResNet34 is accustomed to. To solve this problem, the first layer of ResNet34 is modified so that the number of input channels of the first layer convolution of ResNet34 is 1 to accept single-channel input, which is a key adaptability for processing MRI data. In addition, the parameters of the last two layers of the network (layers 3 and 4) and the fully connected layer are also trainable. This decision ensures that the neural network can be fine-tuned for practical tasks, thereby enhancing the ability to capture complex medical images.

[0067] In this embodiment, the decoder can be composed of multiple custom "decoder block" modules, each of which performs a series of upsampling and convolution operations. These modules can use the jump connection of the corresponding layer of the encoder. This method helps to regain the spatial information lost during the encoder downsampling process, which is a key factor for accurate segmentation. The final convolutional layer of the decoder maps the rich upsampled feature map to the desired location, and finally provides the output of the desired zygomatic protuberance location.

[0068] In addition, the NMRDataset class is an important part of the architecture of this embodiment. This class is designed to process the standard format ".nii.gz" files in medical images, which ensures that the model can effectively process real-world nuclear magnetic resonance data. It can not only load and process MRI images, but also load and process the corresponding labels, thereby simplifying the training process. This class also contains a conversion pipeline, which includes important preprocessing steps such as resizing and normalization to ensure the consistency and quality of the input data.

[0069] The model in this example is well integrated with the PyTorch ecosystem, with data loaders for efficient batch processing and CUDA for GPU acceleration to ensure efficient model training. The training loop is custom designed, using BCEWithLogitsLoss as the loss function and using the adam optimizer to apply different learning rates to different parts of the model. This nuanced training approach allows for finer control and optimization of the learning process.

[0070] In this embodiment, the data set can be divided into training set, validation set and test set in a ratio of 7:2:1, and 200 epochs are run. One epoch means that the model performs a complete forward and backward propagation on the entire training data set. In other words, one epoch means that all training samples are used to update the weights of the model once. The loss function converges well to a position close to zero on both the training set and the test set.

[0071] After actual testing, the model results of this embodiment can well present the position of the most anterior edge of the zygomatic process. In the nuclear magnetic resonance images with artifacts, the recognition ability exceeds that of non-professionals who have undergone short-term training. If the distance between the result and the mark is less than 1 pixel, the accuracy of the model on the test set reaches 100%. When marking, the difference of one pixel is difficult to distinguish and difficult to accurately locate, so this error is considered to be systematic and can be improved later through precise marking and increased data volume.

[0072] In step 203, the center of gravity of the eyeball is determined based on the orbital MRI transverse image.

[0073] The steps of this embodiment are used to determine the position of the center of gravity of the eyeball, which can be calculated by some mathematical methods. These mathematical methods require the use of the center of mass moment and mass. However, in this embodiment, since the eyeball is homogeneous, both the center of mass moment and the mass can be canceled out in the end. Therefore, the coordinates of the center of gravity of the eyeball can be obtained by calculating the mean of the coordinates of the relevant points of the eyeball.

[0074] Specifically, for a 2D sequence of orbital MRI transverse images, the center of gravity of the eyeball must be located on the image of the largest plane of the eyeball, so the center of gravity of the eyeball can be obtained by calculating the average of the positions of all points corresponding to the eyeball image on the image of the largest plane of the eyeball. Exemplarily, the horizontal axis coordinates of each point on the image of the largest plane of the eyeball are averaged to obtain the horizontal axis coordinate of the center of gravity, and the vertical axis coordinates of each point are averaged to obtain the vertical axis coordinate of the center of gravity.

[0075] Specifically, for a 3D sequence of orbital MRI transverse images, it is necessary to calculate the position of the eyeball's center of gravity based on all points on the three-dimensional eyeball surface. First, obtain the coordinates of all points on the eyeball surface in the orbital MRI transverse image in a three-dimensional rectangular coordinate system. Second, calculate the horizontal axis mean, vertical axis mean, and vertical axis mean of all points on the eyeball surface, and determine the position in the three-dimensional rectangular coordinate system whose coordinates are the horizontal axis mean, vertical axis mean, and vertical axis mean as the position of the eyeball's center of gravity.

[0076] In step 204, the position of the corneal anterior edge is determined based on the orbital MRI transverse image.

[0077] The steps of this embodiment are used to determine the position of the anterior edge of the cornea. In the embodiment of the present invention, a more important point is how to determine the position of the anterior edge of the cornea on the image of the largest plane of the eyeball, that is, the position of the apex of the anterior edge of the cornea, which is a key factor to ensure accurate measurement of exophthalmos. Some measurement methods in related technologies often determine the anterior surface of the eyeball, rather than the position of the anterior edge of the cornea. This results in that in the case of non-emmetropia, the anterior surface position of the eyeball located is not the anterior edge position of the cornea, and the exophthalmos should be the vertical distance from the anterior edge position of the cornea to the line connecting the most anterior edge position of the zygomatic processes on both sides, and should not be the vertical distance from the anterior surface position of the eyeball to the line connecting the most anterior edge position of the zygomatic processes on both sides. If the anterior surface position of the eyeball is located, it will cause measurement errors. At present, the measurements mentioned in most related technologies are all the anterior surface position of the eyeball, such as CT measurement methods, which is mainly because CT has a slightly poor resolution for soft tissues and is not easy to identify the anterior edge position of the cornea.

[0078] In this embodiment, for the 2D sequence of orbital MRI transverse images, since the layer thickness of the 2D transverse images is relatively large, about 2.5 mm, it is necessary to assume that the position of the corneal anterior edge is located on the largest layer image of the eyeball (not affected by the layer thickness of the 2D sequence images). Therefore, the front end position of the longest axis of the eyeball area on the largest layer image of the eyeball can be determined as the position of the corneal anterior edge. Figure 4 FIG. 1 shows a schematic diagram of the positional relationship between the most anterior edge positions of the zygomatic processes on both sides, the anterior edge positions of the cornea, and the center of gravity of the eyeball provided by an embodiment of the present invention, as shown in FIG. Figure 4 As shown, the eyeball is similar to a sphere, and the cornea is equivalent to an arc convex on the sphere, where the front edge of the cornea is (x3 ， y3), passing through the anterior edge of the cornea and the center of gravity of the eyeball (x3 ， a) is the longest axis of the eyeball area, so the front end position of the longest axis of the eyeball area on the largest layer image of the eyeball is the front edge of the cornea.

[0079] It should be noted that due to the large layer thickness of the 2D transverse image, about 2.5mm, the height of the eyeball in the vertical direction will be overestimated when the eyeball is volume rendered, so that the stereoscopic recognition of the longest axis of the eyeball is disturbed by the vertical axis. For example, the distance from the farthest point on the surface of the eyeball to the center of gravity may not be the distance from the anterior edge of the cornea to the center of gravity, but the distance from the upper and lower vertices of the eyeball to the center of gravity (overestimation of the height of the eyeball in the vertical direction causes erroneous recognition). Therefore, in the 2D sequence, it is necessary to identify the front end site based on the longest axis only for the plane where the center of gravity is located, rather than the entire sphere, that is, the position of the anterior edge of the cornea.

[0080] For the 3D sequence of orbital MRI transverse images, due to the small layer thickness, it usually does not affect the identification of the corneal front edge position. All points on the surface of the eyeball can be traversed to calculate the distance from each point on the surface of the eyeball to the center of gravity of the eyeball, and the position of the point farthest from the center of gravity of the eyeball is used as the corneal front edge position. In addition, since the MRI detection process takes a long time, it may be difficult for the tester to maintain a stable forward-looking state for a long time. If strabismus appears in the image, that is, the position of the corneal front edge deviates from the front direction of the eye, the corneal front edge position may not be located on the largest layer image of the eyeball, and the corneal front edge position detection method corresponding to the above 2D sequence may produce some errors. The corneal front edge position identification method of traversing all points on the surface of the eyeball, calculating the distance from each point on the surface of the eyeball to the center of gravity of the eyeball, and using the position of the point farthest from the center of gravity of the eyeball as the corneal front edge position in this embodiment will be more accurate.

[0081] In step 205, a first distance from the center of gravity of the eyeball to the line connecting the most anterior edges of the zygomatic processes on both sides and a second distance from the center of gravity of the eyeball to the anterior edge of the cornea are calculated, and the sum of the first distance and the second distance is used as the exophthalmos measurement result.

[0082] In the embodiment of the present invention, by calculating the distance from the center of gravity of the eyeball to the line connecting the most anterior edges of the zygomatic processes on both sides, and adding the distance from the center of gravity of the eyeball to the anterior edge of the cornea, it is equivalent to obtaining the distance from the anterior edge of the cornea to the line connecting the most anterior edges of the zygomatic processes on both sides when the anterior edge of the cornea is at the front and the anterior edges of the zygomatic processes on both sides and the anterior edge of the cornea are located in the same horizontal plane, thereby obtaining the equivalent exophthalmos value.

[0083] Specifically, the measurement terminal may obtain the exophthalmos measurement result based on a preset exophthalmos calculation formula, wherein the exophthalmos calculation formula includes:

[0084]

[0085] Wherein, d represents the measurement result of the exophthalmos, (x1, y1) represents the position coordinates of the most anterior edge of the zygomatic process on one side, (x2, y2) represents the position coordinates of the most anterior edge of the zygomatic process on the other side, and (x3, y3) represents the position coordinates of the anterior edge of the cornea.

[0086] Figure 4 This is a schematic diagram of the relationship between the most anterior edge of the zygomatic processes on both sides, the anterior edge of the cornea, and the center of gravity of the eyeball. Figure 4 As shown in the figure, the coordinates of the most front edge of the zygomatic process on both sides are (x1, y1) and (x2, y2) respectively. The front edge of the cornea is (x3, y3) when looking straight ahead, (b, c) when strabismus occurs, and the coordinates of the center of gravity of the eyeball are (x3, a)

[0087] In one embodiment, if the target in the image has strabismus, the exophthalmos degree should be (b， c) to (x3 ， a) plus (x3 ， a) The vertical distance to the two connecting lines. (b ， c) to (x3 ， The distance of a) is (b-x3) 2 +(ca) 2 .

[0088] In fact, assuming that the change in the sagittal position of the center of gravity during eye rotation is negligible and the shape of the eye does not change, then (b-x3) 2 +(ca) 2 It can be equivalent to the distance from (x3, y3) to (x3, a), that is, (y3-a) 2 That is, by rotating the eyeball so that the anterior edge of the cornea is at the front, the anterior edge of the zygomatic processes on both sides and the anterior edge of the cornea are at the same horizontal plane, and then the equivalent exophthalmos degree is calculated. That is, directly calculate (x3 ， The vertical distance from the zygomatic process to the front edge of the zygomatic process on both sides can be used to calculate the degree of exophthalmos.

[0089] Related technology points out that the center of gravity may be offset by a few tenths of a millimeter, but according to Figure 3 As shown, it can be considered that when the offset is projected onto the axis of the exophthalmos measurement, the offset can be ignored. And the calculation result based on the embodiment of the present invention is also close to the manual measurement value (less than 0.5 mm deviation).

[0090] In the embodiment of the present invention, for a strabismus scene, the actual corneal front edge coordinate (b ， c) Convert the corneal anterior edge coordinates (x3) equivalent to emmetropia ， y3), thereby achieving correction of exophthalmos in strabismus scenes.

[0091] It should be mentioned that in some other related technologies, since the position of the corneal anterior edge cannot be accurately located, the position of the anterior surface of the eyeball (i.e. Figure 4 Midpoint(x3 ， y3) The intersection point with the sphere below, so there is an error, which affects the measurement accuracy.

[0092] The present invention provides a method for measuring exophthalmos based on MRI images, by obtaining an orbital MRI transverse image, and determining the most front edge position of the zygomatic processes on both sides, the center of gravity position of the eyeball and the front edge position of the cornea according to the orbital MRI transverse image; by calculating the first distance from the center of gravity position of the eyeball to the line connecting the most front edge position of the zygomatic processes on both sides, and the second distance from the center of gravity position of the eyeball to the front edge position of the cornea, and finally taking the sum of the first distance and the second distance as the exophthalmos measurement result. It can be seen that the present invention realizes the exophthalmos measurement based on the orbital MRI image. On the one hand, since MRI does not contain radiation, it improves safety compared to the CT method; on the other hand, since MRI has higher contrast and resolution in the imaging of soft tissues, the exophthalmos measurement result obtained based on the MRI image is also more accurate.

[0093] It should be understood that the order of execution of the steps in the above embodiment does not necessarily mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present invention.

[0094] The following is an embodiment of the device of the present invention. For details not described in detail, reference may be made to the corresponding method embodiments described above.

[0095] Figure 5 The schematic diagram of the structure of the device for identifying pathogenic microorganism information provided by the embodiment of the present invention is shown. For the convenience of explanation, only the part related to the embodiment of the present invention is shown, which is described in detail as follows:

[0096] like Figure 5 As shown, the MRI image-based exophthalmos measurement device 5 includes: an acquisition unit 51 , a first determination unit 52 , a second determination unit 53 , a third determination unit 54 and a calculation unit 55 .

[0097] An acquisition unit 51 is used to acquire a transverse MRI image of the orbit;

[0098] A first determining unit 52 is used to determine the most anterior edge positions of the zygomatic processes on both sides according to the transverse MRI image of the orbit;

[0099] A second determining unit 53 is used to determine the center of gravity of the eyeball according to the transverse MRI image of the orbit;

[0100] A third determining unit 54 is used to determine the position of the corneal anterior edge according to the orbital MRI transverse image;

[0101] The calculation unit 55 is used to calculate a first distance from the center of gravity of the eyeball to the line connecting the most anterior edges of the zygomatic processes on both sides, and a second distance from the center of gravity of the eyeball to the anterior edge of the cornea, and take the sum of the first distance and the second distance as the exophthalmos measurement result.

[0102] In one embodiment, the orbital MRI transverse image is a 2D sequence image, and the first determination unit 52 is specifically used to extract the largest layer image of the eyeball with the largest cross-sectional area according to the cross-sectional area of ​​each layer of the eyeball image in the orbital MRI transverse image;

[0103] The image of the largest plane of the eyeball is input into the trained model for recognizing the most anterior edge position of the zygomatic process, so as to obtain the image of the largest plane of the eyeball with the most anterior edge positions of the zygomatic processes on both sides marked.

[0104] The second determination unit 53 is specifically used to calculate the average position of all points corresponding to the eyeball image on the eyeball maximum layer image to obtain the center of gravity position of the eyeball.

[0105] The third determination unit 54 is specifically used to determine the front end position of the longest axis of the eyeball area on the largest layer image of the eyeball as the corneal front edge position.

[0106] In one embodiment, the orbital MRI transverse image is a 3D sequence image, and the first determination unit 52 is specifically used to input each eyeball layer image in the orbital MRI transverse image into a trained zygomatic process anterior edge position recognition model to obtain multiple anterior edge positions corresponding to each eyeball layer image; and select the anterior edge positions of the zygomatic processes on both sides from the multiple anterior edge positions corresponding to each eyeball layer image.

[0107] The second determination unit 53 is specifically used to obtain the coordinates of all points on the surface of the eyeball in the transverse section of the orbital MRI image in the three-dimensional rectangular coordinate system; calculate the horizontal axis mean, vertical axis mean and vertical axis mean of all points on the surface of the eyeball, and determine the position with the coordinates of the horizontal axis mean, vertical axis mean and vertical axis mean in the three-dimensional rectangular coordinate system as the center of gravity position of the eyeball.

[0108] The third determination unit 54 is specifically used to traverse all points on the eyeball surface, calculate the distance from each point on the eyeball surface to the eyeball center of gravity, and take the position of the point farthest from the eyeball center of gravity as the corneal front edge position.

[0109] In one embodiment, a model for identifying the anterior edge position of the zygomatic process is constructed by a hybrid model, which includes a deep residual network for feature extraction and a U-Net convolutional neural network for image segmentation; the training data source of the model for identifying the anterior edge position of the zygomatic process is a certain number of layer images of orbital MRI transverse images with the anterior edge positions of the zygomatic processes on both sides marked; the deep residual network used for feature extraction includes ResNet34; the number of input channels of the first layer convolution of ResNet34 is 1.

[0110] In one embodiment, the calculation unit 55 is specifically used to obtain the exophthalmos measurement result based on a preset exophthalmos calculation formula, and the exophthalmos calculation formula includes:

[0111]

[0112] Where d represents the measurement result of exophthalmos, (x1 ， y1) represents the position coordinate of the most anterior edge of the zygomatic process on one side, (x2 ， y2) represents the position coordinate of the most anterior edge of the zygomatic process on the other side, (x3 ， y3) represents the coordinates of the corneal anterior edge.

[0113] As can be seen from the above, the present invention provides a device for measuring exophthalmos based on MRI images, by obtaining an orbital MRI transverse image, and determining the most anterior edge position of the zygomatic processes on both sides, the center of gravity position of the eyeball, and the anterior edge position of the cornea according to the orbital MRI transverse image; by calculating the first distance from the center of gravity position of the eyeball to the line connecting the most anterior edge position of the zygomatic processes on both sides, and the second distance from the center of gravity position of the eyeball to the anterior edge position of the cornea, and finally taking the sum of the first distance and the second distance as the exophthalmos measurement result. It can be seen that the present invention realizes the measurement of exophthalmos based on orbital MRI images. On the one hand, since MRI does not contain radiation, it improves safety compared to the CT method; on the other hand, since MRI has higher contrast and resolution in soft tissue imaging, the exophthalmos measurement result obtained based on MRI images is also more accurate.

[0114] The embodiment of the present invention further provides a measurement terminal, which may include: a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps in the above-mentioned embodiments of the method for measuring exophthalmos based on MRI images are implemented, for example Figure 2 Alternatively, when the processor executes the computer program, the functions of each unit in the above-mentioned device embodiments are implemented, for example Figure 4 The functions of the units 41 to 45 are shown.

[0115] Exemplarily, the computer program may be divided into one or more units, which are stored in the memory and executed by the processor to implement the present invention. The one or more units may be a series of computer program instruction segments capable of implementing specific functions, which are used to describe the execution process of the computer program in the measurement terminal.

[0116] The measurement terminal may be a computing device such as a desktop computer, a notebook, a PDA, or a cloud server. The measurement terminal may include, but is not limited to, a processor and a memory. Those skilled in the art will appreciate that the measurement terminal may include more or fewer components than shown in the figure, or may combine certain components, or different components. For example, the measurement terminal may also include input and output devices, network access devices, buses, etc.

[0117] The processor may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc.

[0118] The memory may be an internal storage unit of the measurement terminal, such as a hard disk or memory of the measurement terminal. The memory may also be an external storage device of the measurement terminal, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the measurement terminal. Furthermore, the memory may include both an internal storage unit and an external storage device of the measurement terminal. The memory is used to store the computer program and other programs and data required by the terminal. The memory may also be used to temporarily store data that has been output or is to be output.

[0119] The technicians in the relevant field can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In practical applications, the above-mentioned function allocation can be completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated in a processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, which will not be repeated here.

[0120] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0121] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0122] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0123] The embodiments described above are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. A method for measuring exophthalmos based on MRI images, characterized in that: include: Obtain transverse MRI images of the orbit; The most anterior edge of the zygomatic processes on both sides was determined based on the transverse MRI images of the orbit; Determine the center of gravity of the eyeball based on the transverse MRI image of the orbit; The position of the anterior corneal edge was determined based on the transverse orbital MRI images; Calculate the first distance from the center of gravity of the eyeball to the line connecting the most anterior edges of the zygomatic processes on both sides, and the second distance from the center of gravity of the eyeball to the anterior edge of the cornea, and take the sum of the first distance and the second distance as the exophthalmos measurement result.

2. The method for measuring exophthalmos based on MRI images according to claim 1, characterized in that: The orbital MRI transverse images are 2D sequence images; Determining the position of the most anterior edge of the zygomatic processes on both sides according to the transverse image of the orbital MRI includes: According to the cross-sectional area of ​​each layer of the eyeball image in the orbital MRI transverse image, the largest layer image of the eyeball with the largest cross-sectional area is extracted; The image of the largest plane of the eyeball is input into the trained model for recognizing the most anterior edge position of the zygomatic process, so as to obtain the image of the largest plane of the eyeball with the most anterior edge positions of the zygomatic processes on both sides marked.

3. The method for measuring exophthalmos based on MRI images according to claim 2, characterized in that: Determining the center of gravity of the eyeball according to the orbital MRI transverse image comprises: The average position of all points corresponding to the eyeball image on the eyeball maximum layer image is calculated to obtain the center of gravity position of the eyeball.

4. The method for measuring exophthalmos based on MRI images according to claim 2, characterized in that: Determining the position of the anterior corneal edge according to the orbital MRI transverse image comprises: The front end position of the longest axis of the eyeball area on the largest slice image of the eyeball is determined as the front edge of the cornea.

5. The method for measuring exophthalmos based on MRI images according to claim 1, characterized in that: The orbital MRI transverse images are 3D sequence images; Determining the position of the most anterior edge of the zygomatic processes on both sides according to the transverse image of the orbital MRI includes: Input each eyeball layer image in the orbital MRI transverse image into the trained zygomatic process anterior edge position recognition model to obtain multiple anterior edge positions corresponding to each eyeball layer image; The most anterior edge positions of the zygomatic processes on both sides are selected from a plurality of anterior edge positions corresponding to each eyeball layer image.

6. The method for measuring exophthalmos based on MRI images according to claim 5, characterized in that: Determining the center of gravity of the eyeball according to the orbital MRI transverse image comprises: Obtain the coordinates of all points on the surface of the eyeball in the lateral MRI image of the orbit in a three-dimensional rectangular coordinate system; The horizontal axis mean, vertical axis mean and vertical axis mean of all points on the eyeball surface are calculated, and the position with the coordinates of the horizontal axis mean, vertical axis mean and vertical axis mean in the three-dimensional rectangular coordinate system is determined as the center of gravity of the eyeball.

7. The method for measuring exophthalmos based on MRI images according to claim 6, characterized in that: Determining the position of the anterior corneal edge according to the orbital MRI transverse image comprises: Traverse all points on the eyeball surface and calculate the distance from each point on the eyeball surface to the center of gravity of the eyeball; The position of the point farthest from the center of gravity of the eyeball was defined as the corneal anterior edge position.

8. The method for measuring exophthalmos based on MRI images according to claim 2 or 5, characterized in that: The zygomatic process front edge position recognition model is constructed by a hybrid model, which includes a deep residual network for feature extraction and a U-Net convolutional neural network for image segmentation; the training data source of the zygomatic process front edge position recognition model is a certain number of orbital MRI transverse image layer images containing the anterior edge positions of the zygomatic processes on both sides marked; The deep residual network for feature extraction includes ResNet34; the number of input channels of the first convolution layer of ResNet34 is 1.

9. The method for measuring exophthalmos based on MRI images according to any one of claims 1 to 7, characterized in that: The step of calculating a first distance from the center of gravity of the eyeball to a line connecting the most anterior edges of the zygomatic processes on both sides, and a second distance from the center of gravity of the eyeball to the anterior edge of the cornea, and taking the sum of the first distance and the second distance as the exophthalmos measurement result comprises: The exophthalmos degree measurement result is obtained based on a preset exophthalmos degree calculation formula, wherein the exophthalmos degree calculation formula includes: Where d represents the measurement result of exophthalmos, (x1 ， y1) represents the position coordinate of the most anterior edge of the zygomatic process on one side, (x2 ， y2) represents the position coordinate of the most anterior edge of the zygomatic process on the other side, (x3 ， y3) represents the coordinates of the corneal anterior edge.

10. A device for measuring exophthalmos based on MRI images, characterized in that: include: An acquisition unit, used for acquiring a transverse MRI image of the orbit; The first determination unit is used to determine the most anterior edge position of the zygomatic processes on both sides according to the transverse MRI image of the orbit; The second determination unit is used to determine the center of gravity of the eyeball according to the transverse MRI image of the orbit; A third determination unit is used to determine the position of the corneal front edge according to the orbital MRI transverse image; The calculation unit is used to calculate the first distance from the center of gravity of the eyeball to the line connecting the most anterior edges of the zygomatic processes on both sides, and the second distance from the center of gravity of the eyeball to the anterior edge of the cornea, and use the sum of the first distance and the second distance as the exophthalmos measurement result.

Images