14 results about "Manual segmentation" patented technology

The invention provides methods and systems for identifying shape-based atypical segmentations. In one example, a method includes receiving a segmentation of a region of interest (ROI) of a medical image, the segmentation output by a segmentation model; computing a confidence measure for the segmentation, the confidence measure indicating how well a shape of the segmentation can be encoded by one or more dominant shape variation patterns of a set of predetermined segmentations of the ROI; and in response to the confidence measure satisfying a predetermined condition with respect to a threshold, displaying the segmentation, storing the segmentation, and / or using the segmentation for one or more downstream processes; otherwise, prompting a user to perform a manual segmentation.

The invention discloses a visible light region-of-interest and 3D point cloud mapping method, device, equipment and medium, and relates to the technical field of three-dimensional vision and biological recognition crossing. According to the method, a face ROI can be quickly recognized through a visible light image deep learning model (such as YOLO / UNet), manual segmentation is replaced, and the processing speed is increased by 80% or above; a pre-calibrated projection matrix is used to realize sub-pixel-level two-dimensional-three-dimensional mapping, and error accumulation caused by a complex registration process is effectively avoided.

The application discloses a three-dimensional body construction method and device based on CT images, a storage medium and a terminal. A CT sequence image of a target object is acquired. In response to a user's manual contour segmentation operation on a first target CT image in the CT sequence image, the first target CT image is subjected to first contour segmentation, and a first contour image of the target object is obtained. Based on each first contour image, a second target CT image in the CT sequence image, except for the first target CT image, is subjected to second contour segmentation, and a second contour image of the target object is obtained. A three-dimensional body of the target object is constructed according to each first contour image and each second contour image. Since the second contour segmentation is active processing and calculation on the manually segmented first contour image, when the second contour image is obtained, manual operation on a large number of images can be avoided, human-computer interaction is reduced, the three-dimensional body construction efficiency is improved, and the accuracy of the three-dimensional body of the target object can be ensured by the manually segmented first contour image.

This invention relates to the field of medical image processing technology, specifically to a method for constructing an intelligent segmentation model for magnetic resonance imaging (MRI) images. The method includes: acquiring LGE-CMR images; constructing a cardiac region localization network, which is used to locate the cardiac region in the LGE-CMR image; and constructing a region-of-interest (ROI) intelligent enhancement module, which enhances the RIO based on the output of the cardiac region localization network. This invention achieves automatic cardiac region localization by constructing a cardiac region localization network, and combines this with the RIO intelligent enhancement module to selectively adjust contrast and sharpness to highlight lesion areas. Furthermore, it incorporates a feature fusion module from a multi-scale, multi-class segmentation network to enhance the feature expression of small regions. This allows for automated multi-class segmentation of myocardial scars and microvascular occlusions, solving the problems of time-consuming manual segmentation and inter-observer errors. It also avoids the difficulty of existing automatic segmentation models in simultaneously and accurately identifying three types of regions and the possibility of missed or false detections of small regions.

The application is suitable for the technical field of medical image analysis and artificial intelligence assisted diagnosis, and provides a breast cancer neoadjuvant chemotherapy efficacy prediction method based on DCE-4DNeRF, first, a DCE-4DNeRF model is used, non-uniformly sampled original DCE-MRI sequences are reconstructed into uniformly sampled sequences which are continuous in time and space through spherical harmonic functions and projection mechanism; second, a tumor perception prediction network is constructed, the network introduces biological position coding based on anatomical prior, and combines a differentiable sampling mechanism, so that the network can automatically focus on the key tumor area without manual segmentation; finally, the dynamic spatio-temporal features of the two stages before and after chemotherapy are fused to predict the efficacy. The application effectively overcomes the data time inconsistency, realizes the end-to-end and label-free accurate prediction, significantly improves the prediction performance, and provides a reliable basis for clinical individualized treatment decision.

The invention provides a geological core image crack and pore identification method and device, electronic equipment and a storage medium, and the method comprises the steps: collecting core image data in the same drilling well, carrying out the initialization processing, and inputting the processed images into a model A and a model B; after the model A receives the image, performing preliminary identification on core cracks in the image, outputting a mask for each core crack, segmenting each core crack in the image by using the model B, and enhancing the mask generated by the model A; and adjusting the size of the segmented image to be adaptive to the format and size of the model A, inputting the segmented image into the model A and the model B again for iterative training until the core fracture recognition result in the trained image tends to be stable, and outputting core fracture recognition result data after training. According to the method, the model A and the model B are combined, rapid and accurate identification of the crack holes in the rock core image is achieved, and time and energy in the manual segmentation and calculation process are saved.

The application belongs to the technical field of medical image processing and computer-aided diagnosis, and discloses an orbital lymphoma and inflammatory pseudotumor classification method, system, device and terminal, in the obtained orbital DCE-MRI image, the tumor region and the eye cone triangular region in the original image are manually segmented and pretreated; the neural network is used for extracting the features of the tumor region, and the features are subjected to clustering statistical analysis as the tumor region features; the eye cone triangular region is subjected to feature extraction as a similarity judgment standard, and the self-classification and self-recovery network is used for extracting the eye cone region features; the original image is used as the input network and combined with the features of the tumor region and the eye cone triangular region to train the network model; the multi-modal orbital data is input into the classification model for processing, so that the orbital lymphoma and inflammatory pseudotumor are classified. The application comprehensively considers the texture features of the tumor and the depth features of the eye cone region, effectively improves the prediction accuracy, and has the characteristics of high accuracy.

A medical imaging method is disclosed herein. The method includes receiving an image processing neural network (122) configured to output a segmentation. The method further includes receiving a training three-dimensional medical image (123). The method comprises repeatedly: drawing a cross-sectional view (124) of the training three-dimensional medical image; receiving split edit data (129) from a split entry tool (128); constructing a manual partition in the cross-sectional view using the partition edit data (132); and collecting per voxel tool usage metadata that describes the use of the split entry tool. The method further includes determining, at least in part, a per-voxel confidence score representing a quality of annotation of a region suitable for training using the metadata using a per-voxel tool (138); constructing training data (140); and training the image processing neural network using the training data. The training of the image processing neural network is adjusted using the per voxel confidence score, thereby modifying the training of the image processing network in response to the annotation quality.

The invention relates to the technical field of part machining and manufacturing, in particular to a model segmentation setting method and system based on a bottle body contour and a computer program product, and the method comprises the steps: if a to-be-machined bottle body model does not have a segmentation template file, constructing a to-be-machined bottle body segmentation model in a visual interface in a manual segmentation mode; and if the to-be-processed bottle body model has the segmented template file, all segments of the segmented template file are mapped to the to-be-processed bottle body model in an equal proportion, and a to-be-processed bottle body segmented model is obtained. Visual segmentation is directly carried out on the CAD model of the bottle body to be processed, so that an operator does not need to master a complex programming technology, a template segmentation method is further introduced on the basis of manual segmentation, the segmented bottle body model is stored as a template, equal-proportion mapping is carried out on the bottle body models to be processed with similar shapes, and therefore the processing efficiency is improved. Standardized node positioning is realized, the production efficiency is effectively improved, and the production cost is reduced.

Improving accuracy of a predicted segmentation mask, comprising: extracting a ground truth red green blue (RGB) image buffer and a binary contour image buffer from a ground truth RGB image container used for segmentation training; generating the predicted segmentation mask from the ground truth RGB image buffer; generating a second binary contour from the predicted segmentation mask using a specific algorithm; calculating a segmentation loss between a manually segmented mask of the ground truth RGB image buffer and the predicted segmentation mask; calculating a contour accuracy loss between a contour of the binary contour image buffer and a binary contour of the predicted segmentation mask; calculating a total loss as a weighted average of the segmentation loss and the contour accuracy loss; and generating an improved binary contour by compensating the contour of the binary contour image buffer with the calculated total loss, wherein the improved binary contour is used to improve accuracy of the predicted segmentation mask.

The invention discloses a non-alcoholic fatty liver image grading method based on a diffusion neural network, and belongs to the field of medical image processing. The method comprises the following steps: collecting an abdominal ultrasound image file of a patient, screening 10 layers of complete liver and spleen continuous slices / frames, preprocessing, manually segmenting and marking, and constructing a special training data set; constructing a diffusion neural network model containing diffusion feature extraction, attention fusion and segmentation output modules, dividing a data set, and carrying out 20 rounds of iterative training until convergence; and finally, carrying out 16 * 16 gridding cutting on the segmented liver and spleen mask pattern, randomly selecting 8 liver and 6 spleen sampling areas, carrying out weighted fusion on a gray mean and an ultrasonic texture feature value to obtain a liver and spleen ultrasonic feature value, and realizing grading by calculating the ratio of the two. According to the method, the ultrasonic image characteristics are optimally designed, the segmentation precision is high, the sample adaptability is high, the grading result fits the clinical pathological characteristics, and reliable technical support is provided for non-invasive and accurate diagnosis of the non-alcoholic fatty liver disease.

The application relates to a grinding area planning method and device, electronic equipment and a storage medium, wherein the method comprises the following steps: obtaining medical image data of a hip joint part of a target object; inputting the medical image data into a pre-trained deep learning model to obtain bone tissue data and position information of a hip socket center point output by the deep learning model; generating a pelvis model to be ground based on the bone tissue data; determining a grinding area of the pelvis model according to the pelvis model, the position information and a preset acetabular cup model, and visually displaying the pelvis model and the grinding area. Through the application, the problem that manual segmentation based on pelvis tissue depends on the experience and proficiency of operators in the related art, resulting in low segmentation efficiency and low segmentation accuracy, is solved, automatic segmentation of medical image data based on a pre-trained deep learning model is realized, the segmentation accuracy and efficiency can be improved, and the pelvis model and the grinding area can be visually displayed.

The application provides a page processing method and device, wherein the page processing method comprises the following steps: determining an element set corresponding to a target page according to a layout sketch of the target page; drawing an element block corresponding to each element in the element set according to attribute information of each element, and generating an element block distribution diagram according to the element block corresponding to each element; performing iterative segmentation processing on the element block distribution diagram to determine a target segmentation region generated by a segmentation processing result of each iterative segmentation period; and constructing a page layout tree structure of the target page according to a generation order of each target segmentation region. The element block distribution diagram corresponding to the target page is drawn according to the layout sketch of the target page, and then the element block distribution diagram is subjected to iterative segmentation processing to obtain the target segmentation region corresponding to each iterative segmentation period, so that the page layout tree structure of the target page is constructed according to the generation order of each target segmentation region, without relying on manual segmentation of the target page, thereby improving the segmentation efficiency.

Modern retrofit construction practices use 3D point cloud data of the building envelope to obtain the as-built dimensions. However, manual segmentation by a trained professional is required to identify and measure window openings, door openings, and other architectural features, making the use of 3D point clouds labor-intensive. Automatic point Cloud Building Envelope Segmentation (Auto-CuBES) algorithms can significantly reduce the time spent during point cloud segmentation. The Auto-CuBES algorithm inputs a 3D point cloud generated by commonly available surveying equipment and outputs a dimensioned wire-frame model of the building envelope. By leveraging unsupervised machine learning methods in the Auto-CuBES methods facades, windows, and doors can be identified while keeping the number of calibration parameters low. Additionally, some embodiments of Auto-CuBES can generate a heat map of each facade indicating nonplanar characteristics that are valuable for optimization of connections used in overclad envelope retrofits.