83 results about "Brain mri" patented technology

Brain Magnetic Resonance Imaging (MRI), Computerized Tomotography (CT), or other diagnostic modalities may employ a three-step procedure during initial (“scout”) cranial pre-scans that corrects for patient positioning (i.e., roll, yaw and pitch) to reduce inter- and intra-patient variation, thereby enhancing the diagnostic and comparative value of subsequent detail scans even across different diagnostic platforms. In MRI, for instance, locating the saggital sinus (SS) and optimizing a line to bisect the brain through this SS may be automated to correct for roll and yaw. By then identifying the contour of the corpus callosum in a lateral saggital scout scan, the Talairach anterior commissure (AC)—posterior commissure (PC) reference line may be found for correcting pitch. Prescription of detailed scans are improved, especially when the three-step correction is repeated periodically identifying the need to repeat a detailed scan or to adjust the coordinates of a subsequent scan.

The invention discloses a fully convolutional network based tumor segmentation method in allusion to a brain MRI image. A task of tumor segmentation is divided into two steps such as coarse segmentation and fine segmentation. The fully convolutional network based brain MRI tumor segmentation method comprises the steps of firstly training a coarse segmentation fully convolutional network model by using a preprocessed sample so as to be used for detecting a tumor region in an original image, and then taking the tumor region as a training sample to train a fine segmentation fully convolutional network so as to be used for performing fine segmentation on the internal structure of the tumor region. The result shows that the method has a good segmentation effect for a brain MRI tumor.

Apparatus for analyzing brain MRI, is disclosed. The apparatus comprises an input for receiving a first and a second MRI scans at the beginning and end of a predetermined time interval post contrast administration; a subtraction map former for forming a subtraction map from said first and said second MRI scans by analyzing said scans to distinguish between two primary populations, a slow population, in which contrast clearance from the tissue is slower than contrast accumulation, and a fast population in which clearance is faster than accumulation; and an output to provide an indication of distribution of said two primary populations, wherein said predetermined time period is at least twenty minutes.

The present invention, referred to as Oasis is Automated Statistical Inference for Segmentation (OASIS), is a fully automated and robust statistical method for cross-sectional MS lesion segmentation. Using intensity information from multiple modalities of MRI, a logistic regression model assigns voxel-level probabilities of lesion presence. The OASIS model produces interpretable results in the form of regression coefficients that can be applied to imaging studies quickly and easily. OASIS uses intensity-normalized brain MRI volumes, enabling the model to be robust to changes in scanner and acquisition sequence. OASIS also adjusts for intensity inhomogeneities that preprocessing bias field correction procedures do not remove, using BLUR volumes. This allows for more accurate segmentation of brain areas that are highly distorted by inhomogeneities, such as the cerebellum. One of the most practical properties of OASIS is that the method is fully transparent, easy to implement, and simple to modify for new data sets.

The invention relates to a method for segmenting a sulus region on the surface of a pallium of a three-dimensional cerebral magnetic resonance image. The technical characteristics lie in that firstly, the 3D cerebral nuclear magnetic resonance image is pretreated and the surface of the pallium is reconstructed, including removing a braincase and non-cerebral tissues; carrying out cerebral tissue segmenting on a cerebral image and reconstructing the surface of the pallium of geometrical accuracy and correct topological structure in the segmented cerebral image; the surface of the pallium is shown by a series of vertexes and triangles; secondly, the maximum principle curvature and the minimum principle curvature of each vertex on the surface of the pallium are estimated; and finally, sulus and gyrus regions are segmented on the surface of the pallium according to the maximum principle curvature and Hidden Markov Random Field expectation maximization framework and each sulus region is marked by connectivity analysis. Compared with other methods, the method has the advantages of simple and effective algorithm and high segmentation accuracy.

Brain Magnetic Resonance Imaging (MRI), Computerized Tomotography (CT), or other diagnostic modalities may employ a three-step procedure during initial (“scout”) cranial pre-scans that corrects for patient positioning (i.e., roll, yaw and pitch) to reduce inter- and intra-patient variation, thereby enhancing the diagnostic and comparative value of subsequent detail scans even across different diagnostic platforms. In MRI, for instance, locating the saggital sinus (SS) and optimizing a line to bisect the brain through this SS may be automated to correct for roll and yaw. By then identifying the contour of the corpus callosum in a lateral saggital scout scan, the Talairach anterior commissure (AC)—posterior commissure (PC) reference line may be found for correcting pitch. Prescription of detailed scans are improved, especially when the three-step correction is repeated periodically identifying the need to repeat a detailed scan or to adjust the coordinates of a subsequent scan.

The invention provides a multimodal nuclear magnetic resonance image segmentation method for glioblastoma. The segmentation strategy combining the random forest method and the regional growth method is employed, the result of regional growth segmentation of a glioma multimodal magnetic resonance image is replaced with the corresponding random forest segmentation result with low confidence, retraining data is generated to re-train a random forest model, fine segmentation of the glioma multimodal magnetic resonance image is carried out, and a brain MRI image is segmented into a normal brain tissue area, a necrotic area, an active tumor area, a T1 abnormal area and a FLAIR abnormal area. The method is advantaged in that through fine segmentation and positioning of the glioblastoma, doctors are assisted in diagnosis and other treatment tasks, accurate positioning of the glioblastoma and more accurate fine segmentation of different tumor sub areas are carried out, the doctors are facilitated to diagnose the glioblastoma more quickly and accurately, and an accurate treatment scheme is made.

The invention provides a classification method based on a three-dimensional convolution neural network, which is applied to brain MRI images. On the basis of a main network, an auxiliary supervision branch network is designed to supervise and learn a middle layer, and finally the final classification result is acquired by combining the main network with the branch network. The method can makes full use of the three-dimensional convolution neural network to extract the important three-dimensional information of an image, and uses the auxiliary supervision branch network to extract more robust local information of the image to make up for the deficiency of a two-dimensional convolutional neural network in the aspect of three-dimensional feature extraction. The middle layer is supervised andlearnt, so that the network extracts features with a distinguishing ability as early as possible in the learning process. The learning speed is very fast, which has an important influence on the finalclassification result. The auxiliary supervision convolution neural network is added, which can improve the accuracy and robustness of brain MRI image classification, and accelerates the convergenceof the learning process.

The invention discloses an Alzheimer's disease cortex auto-classification method based on multi-scale mesh surface form features. The method includes the steps of determining two sample groups, namely an AD (Alzheimer's disease) group and an NC (normal control) group, and dividing the sample groups into a sample set and a test set under equal proportion; extracting multi-scale mesh surfaces from brain MRI (magnetic resonance imaging) images of samples; calculating LVPD (local vertex point distance) and average curvature for each vertex; with the smoothed LVPD and average curvature, extracting areas having a significant statistical difference and screening two seed points in index sense; extracting a feature row vector for each sample of the training set to form a feature matrix, and training a classifier with the reduced dimensionality feature matrix and corresponding sample classes; testing performance of the classifier with the samples in the test set. By the use of the Alzheimer's disease cortex auto-classification method based on multi-scale mesh surface form features, the defects the prior art is susceptible to cortex segmentation errors and certain-scale difference may be missed are overcome, and the two sample groups can be classified according to the cortex multi-scale form features.

The invention belongs to the technical field of medical image processing, and discloses a hippocampus segmentation method for an automatic brain MRI (Magnetic Resonance Image) on the basis of multipleatlases. The method comprises the following steps that: (1) adopting a non-rigid registration method to carry out registration on atlas set and a brain MRI to be segmented; (2) calculating a similarity between an atlas image and a target image, and constructing and selecting a similar atlas which is most favorable for target image hippocampus segmentation; and (3) obtaining the confidence coefficient weighting probability matrix of the atlas image, establishing a context model based on the similar atlas, and combining the confidence coefficient weighting probability matrix of the atlas imagewith the context model to obtain a hippocampus segmentation result in the target image. By use of the method, image features and image used for segmenting the hippocampus in the atlas image can be mined, an accurate hippocampus segmentation result can be obtained under a situation that time complexity is controlled, and the problem in the prior art that the automatic segmentation accuracy of the hippocampus for the brain MRI image is low is overcome.

The invention provides a multi-modal brain MRI image bidirectional conversion method based on multi-generation and multi-confrontation. The method comprises the steps that an input image (T1 / T2) and acorresponding pathological label are fused through a convolution network and serve as input data of a converter; T1 modal data is input and converted into a T2 modal image through a T2 modal converter; T2 modal data is input and converted into a T1 modal image through a T1 modal converter; confrontation loss between an output image and a real image is constructed; loop verification loss is constructed to achieve the verification of the effectiveness of the converter; content loss between the output image and the real image is constructed, so that the result is closer to the real image; edge loss is introduced to constrain the edges of the real image and output image; the difference between semantic segmentation results of the output image and real image is taken as shape loss to keep theshape consistency. The multi-modal brain MRI image bidirectional conversion method based on the multi-generation and the multi-confrontation can perform bidirectional conversion on multi-modal brain MRI data, and also ensure the invariance of the texture, structure and pathology of the images.

The invention discloses a full convolutional neural network epilepsy focus segmentation method based on a multi-mode image, and mainly solves the problem that the focus in the epilepsy image is difficult to segment in the prior art. The implementation method comprises the following steps: adjusting an original brain MRI image and a PET image to the same resolution space, carrying out edge cutting;dividing the clipped MRI / PET image into a training set and a test set; establishing full convolutional neural network Y-Net; icarrying out input of training sets to Y- Training in the Net network, and carrying out the trained Y-net network; storing the convolution kernel parameters of the convolution layer in the Net network; loading stored convolution kernel parameters to constructed Y-Net network, and inputting the test set to obtain an automatic segmentation result of the epileptic focus. The method has the advantages of easiness in segmentation and high segmentation precision, and can beused for segmenting epileptic focus areas in brain nuclear magnetic resonance imaging (MRI) and positron emission tomography (PET) images.

Embodiments of the present disclosure provide a method for constructing brain templates. First, thousands of brain MRI images are collected. Then, the collected brain MRI images are preprocessed and normalized to construct brain templates. Further, the constructed brain templates include brain templates corresponding to different age and gender groups and brain tissue probability maps corresponding to different age and gender groups. Therefore, embodiments of the present disclosure can construct brain templates by using the large sample size of brain MRI data, and support clinical applications and brain researches.

The invention discloses a brain template generating method. The method includes the steps of firstly, collecting thousand-level-quantity Chinese brain MRI images; secondly, screening and normalizing on the basis of the collected Chinese brain MRI images to generate standard brain templates, wherein the generated standard brain templates include brain templates corresponding to different ages and genders and brain tissue probability graphs corresponding to the brain templates. By the method, the standard brain templates for Chinese can be built, and clinical application and brain science researches are supported.

The invention provides an MRI (Magnetic Resonance Imaging) segmentation method for integrating an attention mechanism aiming at brain lesions. The method comprises the following steps: S1, collecting a brain MRI image with a segmentation image result, and establishing a training set; s2, preprocessing the original brain MRI image to be segmented in the training set; s3, establishing a convolutional neural network with an attention mechanism, and training a model of the convolutional neural network by using the training set; s4, after model training is completed, using trained model parameters to predict the verification set image, and generating a brain MRI tissue and lesion segmentation map; s5, establishing an evaluation file, and evaluating a segmentation result; according to the method, more critical and important information can be extracted, and meanwhile, the training effect of training on a small data set is improved by means of transfer learning.

The invention provides a depression recognition system fusing the cerebral cortex thickness and the edge system morphological features. According to the present invention, the cortex thickness features of a brain is fused with the morphological features of hippocampus and almond kernels of an edge system to carry out the specific feature selection, and a deep neural network (DNN) is used to fuse the features and classify, so that the accuracy of utilizing an MRI image to recognize the depression is remarkably improved. The system comprises (a) an MRI image acquisition module used for acquiringthe brain MRI data of a testee; (b) a feature selection module used for acquiring the cerebral cortex thickness features of the brain and the morphological features of hippocampus and almond kernelsof the edge system according to the acquired MRI data, and further selecting the specific features for recognizing the depression; and (c) a classification and recognition module which uses the deep neural network (DNN) to fuse the cerebral cortex thickness features of the brain and morphological features of the edge system, classify, and recognize the depressive patients and the normal testee.

2-Chloro-2′-deoxyadenosine, hereinafter referred to as cladribine, or a pharmaceutically acceptable salt thereof may be used in the treatment or amelioration of neuromyelitis optica, hereinafter referred to as NMO e.g. in patients known to have NMO-IgG seropositivity or in patients optic neuritis, myelitis and at least two of MRI evidence of contiguous spinal cord lesion 3 or more segments in length, onset brain MRI nondiagnostic for multiple sclerosis or NMO-IgG seropositivity.

The invention discloses a brain magnetic resonance image supervoxel generating method based on iterative spatial fuzzy clustering. The method comprises the steps of firstly, obtaining a set of seed templates from a brain MRI template based on the population since human brains have the same topological structure, secondly, providing an iterative space fuzzy clustering algorithm in order to eliminate the influence of partial volume effects, distributing the voxels to each seed to generate supervoxels. The method can be well applied to brain magnetic resonance images and generate effective brainmagnetic resonance image supervoxels.

Aiming at the influence of the gray unevenness of the brain MRI image and the medical artifact on the segmentation in the prior art, the invention adopts the combination of the region information aggregation and the threshold segmentation as the design idea, introduces the fuzzy multi-threshold technology, and puts forward a brain MRI image segmentation method based on the fuzzy multi-threshold and the region information. Based on fuzzy multi-threshold segmentation, By constructing fuzzy membership function and fuzzy membership aggregation based on local information, To further improve the quality of image segmentation; the invention adopts the fuzzy theory and the information aggregation technology, suppresses the influence of the gray level non-uniformity and the medical artifact on thesegmentation result, can retain more original image information, effectively avoids the false segmentation caused by the artifact, and improves the effect of brain MRI image segmentation. The invention adopts an improved quantum particle swarm optimization algorithm and introduces an exponential descent type shrinkage expansion coefficient. the search performance of the algorithm is improved. At the same time, the convergence rate of the algorithm is improved.

The invention discloses a brain tumor detection method based on attention mechanism and MRI (Magnetic Resonance Imaging) multi-modal fusion, and the method comprises the steps: replacing a common convolution block of an encoder with a mixed cavity convolution block based on a Multi-Unet model; a multi-branch output convolution block MB-OutConv for short is autonomously designed by referring to a multi-branch encoder structure of an Inception model; designing an attention module CB-Attention based on channels, capturing pixel point association among the channels of the original segmented image, and performing attention weighting on the channels; a neural network is properly improved, a new attention module is independently designed to further perfect a segmentation result, and the attention module is based on an image channel and completes attention weighting at a pixel point level. And finally, segmenting the tumor and other lesion areas in the brain MRI image. On the basis of a multi-modal convolutional neural network Multi-Unet, part of encoder branches are improved, and an attention module is added behind the multi-modal convolutional neural network Multi-Unet to jointly improve the segmentation effect of the brain tumor.

The invention discloses an MRI brain tumor image segmentation method based on a DBN neural network. Firstly, a plurality of images are picked out from an existing patient brain MRI sequence image library as training samples, and the training samples are preprocessed and the significance map is calculated. Then the down-sampling is sent to DBN neural network for unsupervised and supervised training. Aiming at the extreme imbalance of training samples, the down-sampling is carried out in the non-tumor area, which improves the detection rate of positive samples. After the training, the test images to be segmented can be sent to the network for segmentation, and visual attention model is introduced to enhance the segmentation accuracy of the network for the more difficult segmentation area, and the segmentation results are output at last.

The invention discloses a multi-modal neural image data automatic information fusion system, which comprises: an original data management module used for storing and managing neural image data collected for a subject and research records; a data conversion module that is used for carrying out desensitization processing and format conversion on the original data; a multi-modal data management module that is used for storing and managing the data converted by the data conversion module; a data preprocessing and feature extraction module that is used for carrying out preprocessing and feature extraction on the data stored in the multi-modal database; an atlas and template data management module that is used for storing and managing atlas and templates; a multi-modal feature data management module that is used for storing and managing feature data; and a grouping or tracking comparison evaluation module that is used for grouping or tracking comparison evaluation of the testees. The systemcan be applied to data analysis and information fusion of multi-modal brain MRI signals, and can realize automatic analysis of data of patients with brain tumors, spinal cord and spinal column injuries and cerebral apoplexy.

A method and apparatus for detecting multiple anatomical landmarks in a 3D volume. A first anatomical landmark is detected in a 3D volume using marginal space learning (MSL). Locations of remaining anatomical landmarks are estimated in the 3D volume based on the detected first anatomical landmark using a learned geometric model relating the anatomical landmarks. Each of the remaining anatomical landmarks is then detected using MSL in a portion of the 3D volume constrained based on the estimated location of each remaining landmark. This method can be used to detect the anatomical landmarks of the crista galli (CG), tip of the occipital bone (OB), anterior of the corpus callosum (ACC), and posterior of the corpus callosum (PCC) in a brain magnetic resonance imaging (MRI) volume.

The invention discloses a brain tissue segmentation method based on regularized graph segmentation . Firstly, a new calculation method of similarity between voxels is designed based on intensity distance and spatial similarity, so as to cluster voxels, and the brain MRI image is divided into a series of supervoxels which are uniform and well fitting to the image edges. By incorporating a priori probabilities of different brain tissues into the graph cutting framework, a formula of energy calculation is designed to calculate the energy value of each supervoxel when each supervoxel is assigned different labels. Then the supervoxel is segmented by graph cutting method and the Magnetic Resonance Imaging (MRI) image is segmented into different tissues. The invention can segment three kinds of brain tissues from the initial brain MRI, and the boundary fitting degree between each tissue in the segmentation result is high. Compared with the prior MRI image segmentation method, the invention has better segmentation effect, higher boundary fitting degree, higher efficiency and faster processing speed, and can better suppress the influence of noise.

A method used in 3D cerebral magnetic resonance data processing for quantitatively analyzing the complexity of cortex and finding out the exception of cerebral cortex shape caused by psychosis and neuropathy includes such steps as acquiring 3D nuclear magnetic resonance data, pre-processing, extracting the body elements at the interface between all or local cerebral cortex gray matter and cerebrospinal fluid, forming the discrete external surface of cerebral gray matter, and calculating the dimensions of parting information about the discrete data in 3D space to obtain the quantitative description of said shape complexity.

Apparatus for analyzing brain MRI, is disclosed. The apparatus comprises an input for receiving a first and a second MRI scans at the beginning and end of a predetermined time interval post contrast administration; a subtraction map former for forming a subtraction map from said first and said second MRI scans by analyzing said scans to distinguish between two primary populations, a slow population, in which contrast clearance from the tissue is slower than contrast accumulation, and a fast population in which clearance is faster than accumulation; and an output to provide an indication of distribution of said two primary populations, wherein said predetermined time period is at least twenty minutes.

The invention discloses a neural network-based optic nerve automatic segmentation and compression degree measurement and calculation method, relates to the technical field of image data processing, and solves the technical problems of time waste and low efficiency of nuclear magnetic tomography analysis of an existing saddle region. The method comprises the following steps: obtaining an MRI image file of a tumor patient in the saddle region; establishing a visual intersection segmentation model, selecting a U-Net segmentation network, adding a space and channel attention mechanism, and performing visual intersection region segmentation on each cut MRI image; measuring and calculating according to the cross compression degree; training and testing the model; according to the method, a plurality of brain MRI images of a tumor patient in the saddle region can be rapidly analyzed in an extremely short time, a large amount of repetitive work of doctors is reduced, and the practical application purposes of assisting the doctors in compression degree judgment and prognosis prediction and providing reference for operations clinically are achieved.

The invention belongs to the field of image processing, and particularly relates to a brain tumor MRI image three-dimensional segmentation method based on an RAPNet network, and the method comprises the steps: constructing the RAPNet network, and carrying out the training of the RAPNet network; the brain MRI image is input into a trained RAPNet network for image recognition and segmentation, a segmented brain tumor MRI image and a substructure region thereof are obtained, the RAPNet network comprises a backbone network, a feature pyramid and auxiliary prediction, and the backbone network is composed of a cavity convolution and a plurality of ISE-R2CU units and used for extracting shallow features and deep features of the input image; according to the method, the feature pyramid formed by the 3D cavity convolution and the cross-model attention mechanism is combined with the trunk to learn the effective features of the whole tumor and the substructure thereof, so that the method has the advantage of fitting various tissue boundaries in the tumor.

The invention relates to an MRI image fusion method based on Laplace pyramid transformation and MRI equipment. The MRI image fusion method comprises the following steps: S1, for the features of a brain MRI image, decomposing a plurality of source images by using Laplace pyramid decomposition to obtain different frequency layers, and adopting different fusion rules in the different frequency layersso as to reserve feature information of each source image in the different frequency layers in a fused image; S2, respectively calculating the regional mean value of the top layer and the point sharpness of other layers to serve as fusion scales; S3, performing normalization processing on the regional mean value and the point definition; S4, comparing the normalized region mean value and point definition value of each layer of different source images, and obtaining a fusion result of each layer of images by adopting different fusion strategies; and S5, for each layer of the Laplace pyramid ofthe obtained fusion image, performing recursion downwards layer by layer from the top layer, and finally obtaining the fusion image. By adopting the MRI image fusion method, the multi-focus fusion image with low noise and clear edge can be obtained.