Brain development atlas and image processing system based on fetal magnetic resonance imaging

Abstract

The present invention discloses a brain development atlas and image processing system based on fetal magnetic resonance imaging, which belongs to the field of medical imaging technology, and specifically comprises: an image acquisition module, which is used to regularly acquire and preprocess nuclear magnetic resonance image data; an image segmentation module, which is used to segment the fetal brain image in the nuclear magnetic resonance image, obtain brain parameters, and identify brain sub-regions; a three-dimensional reconstruction module, which is used to detect whether the fetal brain image is qualified, use a slice-volume registration algorithm to perform three-dimensional reconstruction of the fetal brain image, segment and extract the three-dimensional brain tissue structure, perform cortical reconstruction on the segmentation result, and generate the three-dimensional structure of the fetal brain tissue; an atlas generation module, which is used to obtain a cross-sectional image of the three-dimensional structure of the fetal brain tissue from a preset angle, and arrange the cross-sectional images in chronological order to generate a fetal brain development atlas; the present invention improves the recognition efficiency and accuracy of fetal development information.

Description

Technical Field

[0001] The present invention relates to the field of medical imaging technology, and in particular to a brain development atlas and image processing system based on fetal magnetic resonance imaging. Background Art

[0002] Fetal brain development has rapid structural, functional and metabolic changes, including the development of neuronal networks, neuronal migration, and the formation of sulci and gyri. In order to fully understand the neurodevelopment of normal fetuses and fetuses with congenital diseases, early detection and treatment of congenital diseases, prenatal maternal and child health examinations and quantitative analysis of the developing human fetal brain are essential. Fetal brain MRI atlas can fill the gap in the quantitative template of normal brain development. However, due to the rapid morphological changes of the fetal brain, constructing a fetal brain atlas is more challenging than that of the adult brain, requiring not only additional processing technology but also a large amount of data support.

[0003] In addition, the processing of raw data of fetal MRI, including quality inspection, segmentation labeling, reconstruction, quantification of developmental indicators and other steps are cumbersome, time-consuming and error-prone, and must be completed with the assistance of professional clinicians. A set of fully automated analysis and processing procedures is lacking. With the development of artificial intelligence methods, there are more high-precision methods in the field of medical image processing to replace traditional image processing methods. For example, deep learning methods based on convolutional neural networks have been widely used in medical image processing. It automatically learns complex representative features directly from the data itself, and more accurate results can be obtained by studying fetal brain MRI using convolutional neural networks. Therefore, the present invention is based on fetal magnetic resonance imaging data to develop a brain development atlas and image processing system based on fetal magnetic resonance imaging. Summary of the invention

[0004] The purpose of the present invention is to provide a brain development map and image processing system based on fetal magnetic resonance imaging to solve the following technical problems:

[0005] At present, for the analysis and research of fetal brain MRI, the processing and analysis of data still requires a lot of manual participation, and there is still no unified standard for using MRI to quantify normal fetal brain development. There is a need for an automatic processing and generation system for fetal brain development maps based on MRI images.

[0006] The purpose of the present invention can be achieved through the following technical solutions:

[0007] Brain development atlas and image processing system based on fetal magnetic resonance imaging, including:

[0008] An image acquisition module, used for regularly acquiring and preprocessing nuclear magnetic resonance imaging data;

[0009] An image segmentation module, used to segment the fetal brain image in the MRI image, scan the fetal brain image, obtain fetal brain parameters, and identify fetal brain sub-regions;

[0010] A three-dimensional reconstruction module is used to detect whether the fetal brain image meets the three-dimensional reconstruction requirements, use a slice-volume registration algorithm to perform three-dimensional reconstruction on the fetal brain image that has passed the detection to obtain a three-dimensional brain tissue structure, segment and extract the three-dimensional brain tissue structure, perform cortical reconstruction on the segmentation result, and generate a complete three-dimensional structure of the fetal brain tissue;

[0011] The atlas generation module is used to obtain the cross-sectional images of the three-dimensional structure of the fetal brain tissue from a preset angle, and arrange the cross-sectional images in chronological order to generate a fetal brain development atlas.

[0012] As a further solution of the present invention: the preprocessing of the image acquisition module includes performing N4 bias field strength correction and motion correction on the nuclear magnetic resonance image.

[0013] As a further solution of the present invention: the process of the image segmentation module segmenting the fetal brain image is:

[0014] The fetal brain is located, and the magnetic resonance imaging data is input into a convolutional neural network to obtain a rough segmentation result of the fetal brain image, which is combined with an uncertain segmentation edge image and manually identified to obtain an accurate fetal brain segmentation edge.

[0015] As a further solution of the present invention: the loss function of the computational segmentation of the convolutional neural network includes:

[0016] Scaling loss function:

[0017] Distance loss function:

[0018] Regional loss function:

[0019] Artificial loss function:

[0020] The overall loss of the convolutional neural network Where x represents pixels, δ ò represents the Dirichlet function, Represents the interaction error.

[0021] As a further solution of the present invention: the graph generation module includes:

[0022] The minimized energy function for graph generation is:

[0023]

[0024] Where Kr represents the variable width Gaussian kernel, t n represents the gestational age of fetus n at the time of scanning, represents the differential embryo of the initial template, represents the diffeomorph after template transformation, Indicates subject I n The diffeomorphic transformation between the template and D represents the structural similarity measure in the topological coordinate system.

[0025] As a further solution of the present invention: the system further includes the data processing module, and the data processing module includes:

[0026] Select slices that meet the requirements, extract cross-sectional features of the fetal brain image through a convolutional neural network, extract the fetal brain parameters and identify the sub-regions.

[0027] As a further solution of the present invention: the fetal brain parameters include biparietal diameter and occipitofrontal diameter indicators, and the sub-regions include gray matter, white matter and cerebrospinal fluid.

[0028] As a further solution of the present invention: the three-dimensional reconstruction module includes:

[0029] The three-dimensional brain tissue structure is segmented based on the convolutional neural network, the gray matter segmentation results are extracted and entered into the cortical reconstruction network.

[0030] Beneficial effects of the present invention:

[0031] The present invention first locates the fetal brain in the MRI image, segments the fetal brain image from the background image based on a neural network and manual assistance, identifies the sub-regions of the fetal brain, obtains various parameters of the fetal brain, and thus determines whether the image meets the requirements of three-dimensional reconstruction. The fetal brain is reconstructed in three dimensions through the SVR algorithm, and the sub-regions in the brain tissue are identified again to generate a fetal brain development map. Based on fetal MRI data, the present invention uses a convolutional neural network and an image registration method to automatically generate a spatiotemporal map of the normal development of the fetus with gestational age, thereby improving the efficiency and accuracy of identifying fetal development information. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] The present invention will be further described below in conjunction with the accompanying drawings.

[0033] Figure 1 It is a module schematic diagram of the present invention. DETAILED DESCRIPTION

[0034] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0035] See also Figure 1 As shown, the present invention is a brain development atlas and image processing system based on fetal magnetic resonance imaging, comprising:

[0036] An image acquisition module, used for regularly acquiring and preprocessing nuclear magnetic resonance imaging data;

[0037] An image segmentation module, used to segment the fetal brain image in the MRI image, scan the fetal brain image, obtain fetal brain parameters, and identify fetal brain sub-regions;

[0038] A three-dimensional reconstruction module is used to detect whether the fetal brain image meets the three-dimensional reconstruction requirements, use a slice-volume registration algorithm to perform three-dimensional reconstruction on the fetal brain image that has passed the detection to obtain a three-dimensional brain tissue structure, segment and extract the three-dimensional brain tissue structure, perform cortical reconstruction on the segmentation result, and generate a complete three-dimensional structure of the fetal brain tissue;

[0039] A map generation module, used to obtain a cross-sectional image of the three-dimensional structure of the fetal brain tissue from a preset angle, and arrange the cross-sectional images in chronological order to generate a fetal brain development map;

[0040] In a preferred embodiment of the present invention, the following steps are included:

[0041] S1. Fetal MRI data acquisition and preprocessing;

[0042] S2. Perform fetal brain localization and convolutional neural network rough segmentation;

[0043] S3. The rough segmentation result is combined with the uncertain edge image and provided to the user for adjustment to obtain accurate segmentation;

[0044] S4. Generate quantitative indicators such as biparietal diameter and occipital-frontal diameter for the segmented two-dimensional image, as well as further segmentation results such as gray matter, white matter, and cerebrospinal fluid;

[0045] S5. Check the data quality to see whether it can successfully perform 3D super-resolution reconstruction;

[0046] S6. Reconstruct the quality-checked data using the slice-volume registration algorithm (SVR);

[0047] S7. Segmentation and extraction of 3D brain tissue structures using super-resolution reconstructions;

[0048] S8. Perform cortical reconstruction based on the 3D tissue segmentation results for subsequent analysis.

[0049] In another preferred embodiment of the present invention, the preprocessing of the image acquisition module includes performing N4 bias field strength correction and motion correction on the nuclear magnetic resonance image.

[0050] In another preferred embodiment of the present invention, the process of the image segmentation module segmenting the fetal brain image is:

[0051] Positioning the fetal brain, inputting the nuclear magnetic resonance imaging data into a convolutional neural network to obtain a rough segmentation result of the fetal brain image, combining the uncertain segmentation edge image, and manually identifying to obtain an accurate fetal brain segmentation edge;

[0052] The convolutional neural network can extract feature information of each layer and reference volume using two-dimensional convolutional neural network (2D CNN) and three-dimensional convolutional neural network (3DCNN) respectively. These two convolution-based neural networks use standard convolutional neural network modules, each of which contains a convolutional layer and corresponding batch normalization layer, activation layer and pooling layer.

[0053] In a preferred case of this embodiment, the loss function of the convolutional neural network for computing segmentation includes:

[0054] Scaling loss function:

[0055] Distance loss function:

[0056] Regional loss function:

[0057] Artificial loss function:

[0058] The overall loss of the convolutional neural network Where x represents pixels, δ ò represents the Dirichlet function, Represents the interaction error.

[0059] In medical imaging, there are often only one or two targets in an image, and the pixel ratio of the target is relatively small, making network training more difficult. Therefore, it is necessary to select a suitable loss function and optimize the network reasonably to focus on smaller targets. The loss function is mainly an important indicator for evaluating the matching degree between the model's predicted value and the true label.

[0060] Convolutional neural networks are used to extract spatial information of input images. For each input layer, there is a corresponding two-dimensional convolutional neural network, which can generate feature maps of the input layer and extract higher-order features to input into the bidirectional recurrent neural network. For the reference volume, there is a corresponding three-dimensional convolutional neural network to extract feature information of the volume. The structures of two-dimensional and three-dimensional convolutional neural networks respectively contain multiple convolutional layers with different sizes of convolution kernels and different channels to extract high-order features of the input. There are corresponding batch normalization layers, activation layers and pooling layers after each convolution layer. The batch normalization layer is used to normalize the feature map to prevent the data of the input activation layer from being offset or enlarged; the activation layer is used to increase the nonlinear expression ability of the network; the pooling layer is used to compress the amount of data and parameters, speed up the training speed and avoid overfitting.

[0061] In another preferred embodiment of the present invention, the graph generation module includes:

[0062]

[0063] Where Kr represents the variable width Gaussian kernel, t n represents the gestational age of fetus n at the time of scanning, represents the differential embryo of the initial template, represents the diffeomorph after template transformation, Indicates subject I n The diffeomorphic transformation between the template and D represents the structural similarity measure in the topological coordinate system.

[0064] In another preferred embodiment of the present invention, the brain development atlas and image processing system based on fetal magnetic resonance imaging further includes the data processing module comprising:

[0065] Select slices that meet the requirements, extract cross-sectional features of the fetal brain image through a convolutional neural network, extract the fetal brain parameters and identify the sub-regions.

[0066] In another preferred embodiment of the present invention, the fetal brain parameters include biparietal diameter and occipitofrontal diameter indicators, and the sub-regions include gray matter, white matter and cerebrospinal fluid.

[0067] In another preferred embodiment of the present invention, the three-dimensional reconstruction module includes:

[0068] Segment the 3D brain tissue structure based on the convolutional neural network, extract the gray matter segmentation results, and enter the cortical reconstruction network;

[0069] SVR is a two-dimensional to three-dimensional registration process. Each two-dimensional layer in the layer set needs to be registered to the three-dimensional reference volume to obtain the corresponding motion parameters. A high-resolution three-dimensional volume can be reconstructed based on the obtained motion parameters.

[0070] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A brain development atlas and image processing system based on fetal magnetic resonance imaging, characterized in that: include: An image acquisition module, used for regularly acquiring and preprocessing nuclear magnetic resonance imaging data; An image segmentation module, used to segment the fetal brain image in the MRI image, scan the fetal brain image, obtain fetal brain parameters, and identify fetal brain sub-regions; A three-dimensional reconstruction module is used to detect whether the fetal brain image meets the three-dimensional reconstruction requirements, use a slice-volume registration algorithm to perform three-dimensional reconstruction on the fetal brain image that has passed the detection to obtain a three-dimensional brain tissue structure, segment and extract the three-dimensional brain tissue structure, perform cortical reconstruction on the segmentation result, and generate a complete three-dimensional structure of the fetal brain tissue; A map generation module, used to obtain a cross-sectional image of the three-dimensional structure of the fetal brain tissue from a preset angle, and arrange the cross-sectional images in chronological order to generate a fetal brain development map; The process of segmenting the fetal brain image by the image segmentation module is as follows: Positioning the fetal brain, inputting the nuclear magnetic resonance imaging data into a convolutional neural network to obtain a rough segmentation result of the fetal brain image, combining the uncertain segmentation edge image, and manually identifying to obtain an accurate fetal brain segmentation edge; The loss function of the computational segmentation of the convolutional neural network includes: Scaling loss function: Distance loss function: Regional loss function: Artificial loss function: The overall loss of the convolutional neural network Where x represents pixels, δ ò represents the Dirichlet function, Represents the interaction error.

2. The brain development atlas and image processing system based on fetal magnetic resonance imaging according to claim 1, characterized in that: The preprocessing of the image acquisition module includes performing N4 bias field strength correction and motion correction on the nuclear magnetic resonance image.

3. The brain development atlas and image processing system based on fetal magnetic resonance imaging according to claim 1, characterized in that: The graph generation module comprises: The minimized energy function for graph generation is: Where Kr represents the variable width Gaussian kernel, t n represents the gestational age of fetus n at the time of scanning, represents the differential embryo of the initial template, represents the diffeomorph after template transformation, Indicates subject I n The diffeomorphic transformation between the template and D represents the structural similarity measure in the topological coordinate system.

4. The brain development atlas and image processing system based on fetal magnetic resonance imaging according to claim 1, characterized in that: It also includes a data processing module, which includes: Select slices that meet the requirements, extract cross-sectional features of the fetal brain image through a convolutional neural network, extract the fetal brain parameters and identify the sub-regions.

5. The brain development atlas and image processing system based on fetal magnetic resonance imaging according to claim 1, characterized in that: The fetal brain parameters include biparietal diameter and occipitofrontal diameter indicators, and the sub-regions include gray matter, white matter and cerebrospinal fluid.

6. The brain development atlas and image processing system based on fetal magnetic resonance imaging according to claim 1, characterized in that: The three-dimensional reconstruction module comprises: The three-dimensional brain tissue structure is segmented based on the convolutional neural network, the gray matter segmentation results are extracted and entered into the cortical reconstruction network.

Images