40 results about "Brain ct" patented technology

The invention relates to an automatic detection method for a stroke focus based on computer vision, and belongs to the technical field of artificial intelligence. The method comprises the following steps: acquiring brain CT images, and constructing a training data set; self-adaptive adjustment and standardization of the window width and the window level of the brain CT image are carried out through a self-adaptive window width and window level adjustment strategy; processing the standardized brain CT image through a multi-scale non-uniform region-of-interest sampling strategy; constructing a stroke focus automatic detection model; carrying out double-flow convolution feature extraction and fusion by adopting a double-flow encoder network; processing the fusion feature map through a density perception space attention module; calculating a focus boundary perception loss function, adopting a gradient weighting strategy to carry out gradient adaptive optimization, and updating model parameters through an optimizer to obtain a trained model; and inputting a region-of-interest image block obtained by processing a newly collected brain CT image into the trained model to obtain a final detection result. The accuracy of stroke focus detection can be improved.

The invention discloses a DSA image coronary artery interlayer automatic identification system based on U-KAN and anatomy prior. The DSA image coronary artery interlayer automatic identification system comprises a data preprocessing module, an anatomy prior information extraction module, a U-KAN model construction module, a model training module, an interlayer area automatic identification module and a result processing and evaluation module. According to the method, the KAN layer is introduced, so that the expression capability and the recognition precision of the U-Net model on complex medical images (such as DSA images) are remarkably enhanced. According to the method, anatomical prior information (such as geometric morphology of coronary artery, vascular branches and the like) is creatively fused into the deep learning model, so that the sensitivity of the model to a lesion area is improved. According to the invention, through the optimized U-KAN network, the identification precision is ensured, and the calculation efficiency is high. The U-KAN model provided by the invention has strong expansibility, can be applied to automatic identification of coronary artery interlayer, and can also be applied to other types of medical image analysis tasks, brain CT and automatic detection and segmentation of tumor MRI images.

The application provides a child skull fracture automatic detection method and device based on generated data augmentation. The method comprises the following steps: acquiring a first child brain CT image set, and performing pixel labeling and preprocessing to obtain fracture data, suture data and normal data; obtaining a trained fracture lesion generation model and a suture generation model according to the fracture data, the suture data and the normal data; randomly sampling Gaussian noise for multiple times, inputting the Gaussian noise into the fracture lesion generation model and the suture generation model to obtain generated fracture CT patch data and generated suture CT patch data; training a candidate fracture detection model according to the fracture data, the suture data, the normal data, the generated fracture CT patch data and the generated suture CT patch data, and outputting a fracture detection model; and acquiring a second child brain CT image, and detecting the second child brain CT image according to the fracture detection model. The fracture lesion data and the suture data are augmented, so that the resolution of fractures and sutures can be improved.

The invention discloses a generative artificial intelligence cerebral artery lesion detection method based on plain-scan brain CT. The method comprises the steps that 1, a plain-scan brain CT image to be processed is acquired; 2, converting the plain scanning brain CT image into a CTA image through an adaptive noise elimination network; and 3, carrying out multi-modal lesion detection on the basis of the CTA image generated in the step 2. According to the method, a CTA image is generated through an adaptive noise elimination network (ANE-NET), and a real-time lesion feature analysis engine (RTAL-FE) is embedded in the generation process, so that real-time classification prediction of lesion types is realized. And furthermore, a detection result is output through a dynamic weight decision model (DWD-M), so that the generation quality and the detection efficiency are remarkably improved. The problems that a traditional method is low in generation quality and lags behind detection are solved, and the method is particularly suitable for low-dose and non-invasive cerebrovascular disease screening scenes and has important clinical application value.

The present invention relates to the field of image processing technology, and in particular to a multi-plane cranial CT reconstruction method for patients with cerebral venous thrombosis, comprising: obtaining a closed area of ​​each plane of a CT image of a patient's CT scan; obtaining an area to be analyzed on each plane according to the shape distribution and grayscale distribution of each closed area of ​​each CT image; dividing the area to be analyzed on each plane according to the angle distribution of the area extension lines between the area to be analyzed on each plane and the area to be analyzed on the adjacent plane, and obtaining suspected lesion areas on each plane; performing edge matching on the suspected lesion areas on each plane and the adjacent plane according to the distance distribution and relative position distribution between the edge pixels of the suspected lesion areas on each plane and the edge pixels of the suspected lesion areas on the adjacent plane; and performing three-dimensional reconstruction of the patient's cranial CT according to the matching results. The present invention improves the effect of three-dimensional reconstruction of cranial CT.

A method for deriving dementia-related information using volume predicted from brain CT includes: a step of an analysis apparatus receiving a brain CT (Computed Tomography) image of a subject; a step of the analysis apparatus inputting the brain CT image into a pre-trained segmentation model to extract regions of interest; a step of the analysis apparatus inputting pixel information of the regions of interest into a pre-trained first learning model to predict the volume of at least one region among the regions of interest; and a step of the analysis apparatus inputting the volume of the at least one region into a pre-trained second learning model to derive dementia-related information of the subject.

The application provides a neurosurgical brain puncture operation robot and electronic equipment, in the neurosurgical brain puncture operation robot, a control host is used for matching a brain CT image to be planned with an intracranial hematoma puncture path database constructed to determine a hematoma puncture planning path, and the mechanical arm is controlled according to the hematoma puncture planning path to drive the surgical instrument to reach the hematoma position of the brain.In the embodiment of the application, only the CT image scanned in the diagnosis process is used as the brain CT image to be planned, and no other additional image is needed, and the intracranial hematoma puncture path database constructed is matched, so that the planning of the intracranial hematoma puncture path is completed only using the brain tissue in a short time, the planning efficiency is improved, and moreover, the planning process does not need additional manual operation, and the implementation difficulty is smaller for young doctors and primary doctors with less clinical experience, and on this basis, the treatment efficiency of the patient with cerebral hemorrhage is effectively improved.

The embodiment of the invention provides a simulation method and device for craniocerebral growth and a storage medium, and belongs to the technical field of images. The method comprises the following steps: constructing three-dimensional brain models in one-to-one correspondence with standard brain CT sample images corresponding to different growth time periods; taking the three-dimensional brain model corresponding to an initial time period in different growth time periods as an initial simulation model for segmentation to obtain a plurality of brain model components; performing finite element solution of growth parameters on a gridding three-dimensional craniocerebral model obtained by gridding each craniocerebral model component to obtain a predicted craniocerebral model of the next growth time period; obtaining a first target growth parameter corresponding to the current growth time period according to the predicted craniocerebral model and the three-dimensional craniocerebral model in the same growth time period; and taking the predicted craniocerebral model as an initial simulation model to obtain a first target growth parameter of the next growth time period. According to the embodiment of the invention, the simulation efficiency and simulation precision of skull simulation can be considered.

The disclosure provides a kind of cerebral hemorrhage analysis method and system based on image registration, comprising: automatically segmenting cerebral hemorrhage area in brain CT image;The cerebral CT image and the segmented cerebral hemorrhage image are pretreated;Cerebral CT image is registered according to MRI template, and the deformation field is obtained;The cerebral hemorrhage image is transformed using the deformation field;Analysis patient bleeding level and the brain area involved.Use normal brain MRI image as registration template, register patient brain CT with MRI template, extract patient hematoma feature, can quickly and effectively analyze bleeding level and the brain area involved by hematoma, provide guidance for doctors, and can be used for subsequent visualization modeling, help doctors to formulate more comprehensive treatment plan.

The application discloses an abnormality recognition method based on brain CT images, and relates to the technical field of image processing.The method comprises the following steps: acquiring a brain CT image sequence and extracting an imageomics feature, and constructing an initial imageomics feature set; analyzing the dynamic evolution mode of the feature in the image sequence, and obtaining an imageomics trajectory dynamic coefficient by calculating the entropy value and standard deviation of the feature change rate sequence; constructing a brain structure-function heterogeneous graph, inputting the heterogeneous graph into an abnormality recognition model constructed based on a graph neural network, and outputting a graph-level semantic embedding vector; performing similar case matching, abnormality type determination and key brain area positioning based on the vector, and generating a structured abnormality recognition report.The application solves the problems of insufficient utilization of dynamic information of brain CT images, lack of fusion of anatomical structure prior knowledge and poor result interpretability in the prior art, and improves the accuracy, interpretability and clinical applicability of brain abnormality recognition.

The invention relates to the technical field of medical image processing, in particular to a patient brain CT image intelligent calibration system and method based on big data. A density gradient analysis unit is subjected to self-adaptive partitioning according to brain anatomical characteristics, and through three-dimensional direction gradient detection and Laplacian differential operator dual-scale fusion, a brain CT image is obtained; calculating the density gradient value of each region; the displacement compensation weight unit depends on an association rule model trained by big data, a continuous weight field is generated through two-stage decision, and a specific compensation coefficient is loaded to a key area; the image calibration execution unit adopts a differentiation algorithm to carry out sub-pixel-level non-rigid correction on a high-density mutation region and carry out rigid correction on a homogeneous region; and a partition correction result is integrated, so that the influence of head micromotion is effectively counteracted, tissue boundary blur is eliminated, the brain CT image definition and structural accuracy are improved, and reliable support is provided for clinical diagnosis.

The application discloses a middle cerebral artery high-density sign identification method, device, equipment and storage medium. In the identification method, first, the middle cerebral artery region of a brain CT plain scan image is extracted to obtain a region extraction image; then, according to a blood vessel enhanced image corresponding to the region extraction image, position information of a middle cerebral artery high-density sign candidate box is obtained; then, texture features are obtained according to the region extraction image and the position information, and shape features are obtained according to the blood vessel enhanced image and the position information; finally, a middle cerebral artery high-density sign identification result is obtained according to the texture features and the shape features. The application can realize accurate and reliable identification of HMCAS in a complex brain lesion environment, and has strong adaptability to scanning equipment and individual imaging differences.

The invention discloses a system, a method, equipment and a medium for generating diffusion from craniocerebral CT to craniocerebral CTA under self-adaptive characterization guidance, and belongs to the technical field of medical image processing and artificial intelligence. The system comprises an adaptive representation extraction module used for extracting a depth feature vector representing a blood vessel and an anatomical structure from an input non-enhanced CT image; the conditional potential diffusion model is used for executing a de-noising diffusion process in a potential space by fusing U-Net of a cross attention mechanism under the condition of the feature vector to generate target feature representation; and the image reconstruction module is used for decoding and reconstructing the potential features into a synthetic CTA image. According to the method, the diffusion generation process is guided through the self-adaptively extracted anatomical priori depth, high-quality and high-anatomical-consistency three-dimensional craniocerebral CTA image synthesis without an iodine contrast agent is realized, and the defects of dependence on the contrast agent, inaccurate blood vessel reduction and poor three-dimensional continuity in the prior art are effectively overcome.

The application relates to a method for obtaining a perfusion parameter map and a lesion area and a computer readable storage medium, and the method comprises the following steps: obtaining three-dimensional dynamic CT perfusion images based on brain CT perfusion images at different time points; performing filtering processing on each voxel in the three-dimensional dynamic CT perfusion images to obtain filtered three-dimensional dynamic CT perfusion images, wherein the filtering processing comprises the following steps: at different time points, for each current voxel, filtering is performed by using spatial similarity and time intensity similarity of adjacent voxels of the current voxel; and obtaining a perfusion parameter map and a lesion area according to the filtered three-dimensional dynamic CT perfusion images. The application filters by using spatial similarity and time intensity similarity, considers not only spatial similarity information but also time information of continuous scanning CT perfusion images, improves the signal-to-noise ratio of the images and the calculation accuracy, and the obtained perfusion parameter map and lesion area in the subsequent process are more reliable.

The application relates to a kind of brain CT perfusion image full-automatic post-processing methods, comprising: reading CT perfusion image;The CT perfusion image is preprocessed, and the skull perfusion image with skull is obtained, and the brain tissue image containing lateral ventricle after removing skull;Obtain lateral ventricle segmentation result binary graph, comprising: using the skull perfusion image with skull, obtain the optimal image layer of the largest connected domain area in skull;In the brain tissue image, from the optimal image layer, find the first distance towards the direction of skull top, find the second distance towards the direction of skull bottom, obtain the image layer containing lateral ventricle;The image layer containing lateral ventricle is converted into feature image, and lateral ventricle segmentation is carried out on each layer feature image, and lateral ventricle segmentation result binary graph is obtained;Using the brain tissue image, arterial input function and perfusion parameter map are sequentially obtained;The perfusion parameter map is binarized, and the lesion area is obtained in combination with the lateral ventricle segmentation result binary graph.

The application relates to the technical field of medical image processing, and discloses an intracranial aneurysm recognition method and system based on a brain CT. The method obtains a brain CT image sequence, adopts an anisotropic diffusion filtering algorithm for pretreatment to suppress noise and retain blood vessel edge details. Subsequently, a region growing algorithm is used to segment an intracranial blood vessel region, and a skeletonization algorithm is used to extract a blood vessel center line to calculate curvature and diameter changes. Finally, based on curvature anomaly and diameter ratio analysis, combined with a multi-scale sliding window and a blood vessel topological structure verification, automatic recognition of intracranial aneurysms is realized, and the accuracy and reliability of diagnosis are improved.

The utility model relates to the technical field of examination beds, in particular to a CT examination bed which comprises a bed bottom plate, a head supporting mechanism is arranged at one end of the bed bottom plate, and a lifting and moving mechanism is arranged at the bottom end of the bed bottom plate. The head supporting mechanism comprises a connecting column, one end of the connecting column is fixedly connected with one end of the bed bottom plate, and the other end of the connecting column is fixedly connected to the head supporting frame. Through a head supporting frame, a bottom supporting plate and a side position plate which are arranged on the head supporting mechanism, the head of a patient is placed on the bottom supporting plate, a first adjusting screw rod is driven to rotate by rotating a first handle, and therefore the bottom supporting plate is driven to move up and down, and flexible adjustment can be conveniently conducted according to the heights of different head positions. A second handle is rotated to drive a second adjusting screw rod to rotate, so that a side plate is driven to move front and back, the side fixing effect of the head of the patient is achieved, the situation that the head shakes to affect the image quality and cause diagnosis errors during brain CT scanning of the patient is avoided, and the practical effect of the examining table is improved.

The invention discloses a three-dimensional brain CT medical visual question-answering method and system based on an anatomical memory matrix, and aims to solve the problem that an existing medical visual question-answering model is difficult to fully process association between three-dimensional brain CT image space continuity and an anatomical structure. According to the method, through explicit modeling of three-dimensional anatomical priori and cross-slice semantic association, the spatial understanding ability of the model for brain structures and pathological features is enhanced, and the accuracy and reliability of medical visual questions and answers in a three-dimensional brain CT scene are improved.

The application provides an acute cerebral stroke early intelligent screening early warning method and system based on deep learning, relates to the technical field of medical image processing, and comprises the following steps: acquiring patient brain CT images, historical medical records and cerebral vascular data, extracting density abnormality features and symmetry change features to judge lesions, constructing a multi-scale feature pyramid and recursively refining lesion boundaries, combining three-dimensional lesion feature maps and illness change time sequence features, adopting a self-attention mechanism and a variational autoencoder to analyze cerebral vascular tree structures, and finally generating a cerebral vascular function impairment probability early warning map, so that early accurate screening and early warning of cerebral stroke can be realized, and the timeliness of treatment is improved.

The invention discloses a baseline brain CT image processing method and platform and a storage medium. The platform comprises an image segmentation module used for obtaining a baseline brain CT image of a user and inputting the baseline brain CT image into a pre-trained segmentation model to obtain a brain parenchyma region bleeding image and a bleeding volume; the image vector construction module is used for inputting the brain parenchyma area bleeding image into a pre-trained radiomics model to obtain a radiomics score; constructing an image vector based on the radiomics score and the bleeding volume; the text vector construction module is used for extracting blood biochemical marker information and medical record information of the user and constructing a text vector; and the output module is used for inputting the image vector and the text vector into a pre-trained risk judgment model and outputting a risk score of the baseline brain CT image. According to the method, the baseline brain CT image can be automatically subjected to quantitative evaluation based on the image processing platform, and the cerebral hemorrhage expansion risk score is output for reference of the user.

This invention discloses an interactive hematoma segmentation and analysis method and system based on brain CT images. The method includes: acquiring brain medical images and setting window width and window level for the medical images; determining a first hematoma region based on the human-computer interaction selection area; performing three-dimensional median filtering; performing three-dimensional Otsu threshold segmentation; performing three-dimensional opening operation; selecting the largest three-dimensional connected component of the binary segmented image to obtain the segmentation result of the second hematoma region; obtaining the hematoma centroid and hematoma volume; and calculating the long axis direction of the hematoma using PCA. This method has a small memory footprint, is easy to integrate into software and deploy on various devices, and provides relatively complete hematoma information, effectively improving the calculation speed and hematoma segmentation accuracy. It can assist doctors or surgical robots in brain hematoma localization analysis and puncture path planning, and has the characteristics of lightweight, easy integration, ease of operation, information diversity, accuracy, and speed.

The application discloses a stroke CT image prototype confidence calibration classification method and system, and relates to the fields of medical image intelligent analysis, computer-aided diagnosis and artificial intelligence image recognition. The method extracts global image features after preprocessing of brain CT images; a learnable category prototype is constructed and dynamically optimized; a prototype comparison constraint is established based on the cosine similarity of the global image features and the category prototype; prototype enhanced features are obtained according to the cosine similarity, and multi-scale fine processing is performed; the distance between the fine features and the category prototypes is calculated to obtain initial classification probability and initial confidence; the initial confidence is calibrated in combination with the nearest distance; a perturbation sample is constructed and an uncertainty perception consistency constraint is established; the trainable parameters of the classification network are optimized based on a total loss function to obtain a classification model, and category prediction is realized. The method can improve the category feature separability in stroke CT image classification and reduce the risk of misjudgment.

The invention relates to an intelligent medical instrument, in particular to an electrode anatomy positioning system before intractable epilepsy lesion resection operation, which is used for solving the problems that before resection operation, intracranial electrode contact anatomy attribute judgment is lack of standards, classification depends on artificial experience and is difficult to repeat, and multi-modal images are lack of fusion and quantitative analysis. According to the scheme, brain segmentation is carried out based on a T1w image of a patient before electrode implantation, and each brain region segmentation image is made into a mask image; rigid body registration is carried out based on the brain CT with the implanted electrode and the T1w image, an electrode entry point and a target point are marked, an electrode contact is reconstructed, then the position of the contact is judged according to the coordinates of the electrode contact and the mask image, electrode classification and positioning are achieved, a doctor can conveniently distinguish the source of an intracranial electroencephalogram signal, and then the position of a focus is determined. In addition, the position relation and the metabolism condition of the electrode contacts and the cortex layers of the different brain partitions can be clearly checked by fusing multi-mode images.

The invention provides an intelligent CT vehicle using method based on multi-modal data fusion, a vehicle main body is included, a stretcher, brain CT equipment, an operation panel and a display terminal are arranged in a compartment of the vehicle main body, and a plurality of video acquisition devices and audio acquisition devices are further arranged in the compartment of the vehicle main body. The problem that the medical record file is generated on the CT ambulance immediately and accurately is solved, the medical record file can be updated in real time according to a CT scanning result and uploaded to a hospital immediately, and more rescue time is won for patients.

The invention discloses a cerebral apoplexy medical image segmentation method and device based on deep learning, and the method specifically comprises the steps: collecting brain CT and MRI images, and carrying out the preprocessing and registration; then training a semantic segmentation model which is based on a Transform structure and is optimized through a symmetric mask attention mechanism; and finally, according to the output of the model, post-processing and visualizing the segmentation result of the focus of cerebral apoplexy, displaying the anatomical features of the brain medical image, and providing a basis for medical parameter evaluation.

The invention discloses a cerebrovascular lesion feature generation method based on multi-instance learning and anatomical structure positioning, and the method comprises the steps: obtaining a brain image which comprises a brain CT image and an MRI image; the brain image is preprocessed; positioning and extracting the preprocessed brain image through an anatomical structure positioning network to obtain a cerebrovascular anatomical structure image; wherein the anatomical structure positioning network adopts a deep learning model; performing deep semantic coding on the cerebrovascular anatomical structure image to obtain a high-dimensional feature vector, performing multi-instance weighted fusion on the high-dimensional feature vector through an attention mechanism to obtain a fusion feature, converting the fusion feature to obtain an intermediate feature, and performing label classification and severity regression calculation according to the intermediate feature to obtain a target feature; according to the method, the accuracy and the stability of cerebrovascular lesion feature recognition in CT and MRI images can be remarkably improved.

The invention relates to the technical field of internet big data, in particular to a hemorrhagic side and non-hemorrhagic side cerebrospinal fluid volume quantification method based on deep learning, which comprises the following steps: S1, acquiring N brain CT images to be processed; s2, inputting N brain CT images into the trained end-to-end multi-task collaborative architecture network, and respectively predicting and outputting a corresponding cerebrospinal fluid segmentation result, a brain midline prediction result and a cerebral hematoma segmentation result; s3, partitioning the left hemisphere and the right hemisphere of the brain based on the brain midline prediction results of all the brain CT images; s4, the number of hematoma pixel points of the left hemisphere and the right hemisphere of the brain is calculated based on the partitioning result of the left hemisphere and the right hemisphere of the brain and in combination with the hematoma segmentation results of all the brain CT images, and the hemorrhagic side and the non-hemorrhagic side of the brain are determined; and S5, based on the cerebrospinal fluid segmentation results of all the brain CT images, respectively calculating the hemispherical cerebrospinal fluid volumes of the hemorrhagic side and the non-hemorrhagic side of the brain and the volume difference. According to the method, the accuracy and objectivity of cerebrospinal fluid volume quantification can be improved.

The invention discloses a medical image processing method based on deep learning, and the method comprises the following steps: S1, collecting medical images of three modes of brain CT, MRI and PET, and carrying out the multi-mode adaptive preprocessing; s2, registering an image to be registered in the preprocessed multi-modal image data to a standard space through a deep learning registration network to obtain registered multi-modal image data; s3, inputting the registered multi-modal image data into a Transform model optimized based on an adaptive asymmetric attention mechanism for training and reasoning, and outputting an initial semantic segmentation hotspot map; s4, post-processing is conducted on the initial semantic segmentation hotspot map, and a segmentation prediction result is exported. According to the method, an adaptive asymmetric attention mechanism is adopted, for the recognized unilateral focus area, the model can automatically reduce the attention weight of the symmetric area of the unilateral focus area, and therefore it is effectively avoided that symmetric false positive focuses are generated in the unhealthy side brain area.

This invention discloses an automatic generation method for brain CT medical reports based on co-occurrence relation hierarchical attention. The method preprocesses the brain CT dataset and establishes a vocabulary; constructs a feature extractor for brain CT images to extract visual features; and builds a co-occurrence relation semantic attention module to extract semantic attention features of common medical terms in brain CT images, which includes a word embedding layer and a semantic attention mechanism. A topic vector-guided visual attention module is also constructed, where topic vectors fuse semantic information from common and rare medical terms to fully express sentence-level medical terminology topics. These topics then guide the visual attention mechanism to capture important lesion region features. This method combines the co-occurrence relationships between common medical terms to infer missing semantic information, thereby extracting richer semantic attention features. This hierarchical collaboration improves the accuracy and diversity of the generated brain CT medical reports.