317 results about "Vessel segmentation" patented technology

The invention discloses a fundus image retinal vessel segmentation method and a fundus image retinal vessel segmentation system based on deep learning. The fundus image retinal vessel segmentation method comprises the steps of performing data amplification on a training set, enhancing an image, training a convolutional neural network by using the training set, segmenting the image by using a convolutional neural network segmentation model to obtain a segmentation result, training a random forest classifier by using features of the convolutional neural network, extracting a last layer of convolutional layer output from the convolutional neural network, using the convolutional layer output as input of the random forest classifier for pixel classification to obtain another segmentation result, and fusing the two segmentation results to obtain a final segmentation image. Compared with the traditional vessel segmentation method, the fundus image retinal vessel segmentation method uses the deep convolutional neural network for feature extraction, the extracted features are more sufficient, and the segmentation precision and efficiency are higher.

A method of analyzing structure of a network of vessels in a medical image, comprising:receiving the medical image depicting the network of vessels;obtaining a mask of the network of vessels in the image; andgenerating a non-forest graph mapping a plurality of paths of vessels in the network to directed paths in the graph, with each edge in the graph either directed to indicate a known direction of flow in the corresponding vessel, or undirected to indicate a lack of knowledge of direction of flow in the corresponding vessel, and with all directed edges in a path directed in a same direction as the path.

Method and system for processing an object within a diagnostic image comprises segmenting a three dimensional (3D) object within a diagnostic image. A contour of the object is fitted with a 3D mesh comprising splines in at least first and second directions. The splines provide a plurality of editable control points, and the splines in the first direction intersect with the splines in the second direction at intersection points. A position of at least one control point on the 3D mesh is adjusted based on a user input.

The invention relates to an attention mechanism U-shaped densely connected retinal (a novel retinal blood vessel segmentation method fusing DenseNet and Attention U-net network) blood vessel segmentation method, comprising the steps of retinal blood vessel image preprocessing and constructing a retinal blood vessel segmentation model. The method of the invention can effectively solve the problemsthat adjacent blood vessels are easy to connect, the micro blood vessels are too wide, the small blood vessels are easy to break, the blood vessel intersection is insufficient in segmentation, the image noise is too sensitive, the gray value of the object and the background is crossed, the optic disc and the lesion focus are missegmented, and the like. The method of the invention integrates a plurality of network models under the condition of low complexity, the excellent segmentation results are obtained on DRIVE dataset, the accuracy and sensitivity are 96.95% and 85.94%, respectively, whichare about 0.59% and 7.92% higher than those published in the latest literature.

The invention discloses a method for automatically detecting a coronary artery calcified plaque of a human heart. The method comprises the following steps of S1.adopting a deep learning neural networkto segment an original graph of a coronary artery CTA sequence in order obtain a coronary artery extraction graph of the human heart; S2.processing the coronary artery extraction graph of the human heart to generate a straightening picture of each branch vessel; S3.carrying out blood vessel segmentation on each straightening picture to obtain a straightening blood vessel graph of each branch vessel; S4.adjusting a window width and a window level, calculating a pixel value of the whole picture of each straightening blood vessel graph, if a pixel point whose pixel value is greater than 220 exists, determining that a calcified plaque exists, and screening out the straightening blood vessel graph with the calcified plaque; S5.converting the straightening blood vessel graph with the calcifiedplaque into a grey scale graph, filling the pixel point whose gray value is larger than 220 with the color, and obtaining a calcified plaque extraction result; and S6.calculating a rate of stenosis ofthe blood vessel and obtaining a quantization value. The method is effective for detection of most calcified plaques, automatic detection can be realized, and the efficiency is greatly improved.

The present invention provides a dynamic model directed angiography three-dimensional reconstruction method, pertaining to an intersectional field of digital image processing and medicine imaging, for satisfying special requests of auxiliary detection and surgery guidance for cardiovascular disease in clinical medicine. The invention includes steps of angiography image preprocessing, vascellum segmentation, vascellum skeleton and radii extraction, model directed vascellum base element recognition, vascellum matching and vascellum three-dimensional reconstruction. The invention also provides a cardiovascular dynamic model construction method including steps of cardiovascular slice data extraction, cordis static and dynamic model building, and static and dynamic model building. According to the invention, good angiography three-dimensional reconstruction result may be obtained, which effectively assists detection and surgery guidance for cardiovascular disease, thereby satisfying clinical requests.

In the context of the early detection and monitoring of eye diseases, such as glaucoma and diabetic retinopathy, the use of optical coherence tomography presents the difficulty, with respect to blood vessel segmentation, of weak visibility of vessel pattern in the OCT fundus image. To address this problem, a boosting learning approach uses three-dimensional (3D) information to effect automated segmentation of retinal blood vessels. The automated blood vessel segmentation technique described herein is based on 3D spectral domain OCT and provides accurate vessel pattern for clinical analysis, for retinal image registration, and for early diagnosis and monitoring of the progression of glaucoma and other retinal diseases. The technique employs a machine learning algorithm to identify blood vessel automatically in 3D OCT image, in a manner that does not rely on retinal layer segmentation.

Method and system for processing an object within a diagnostic image comprises segmenting a three dimensional (3D) object within a diagnostic image. A contour of the object is fitted with a 3D mesh comprising splines in at least first and second directions. The splines provide a plurality of editable control points, and the splines in the first direction intersect with the splines in the second direction at intersection points. A position of at least one control point on the 3D mesh is adjusted based on a user input.

The invention discloses a retinal vessel segmentation method based on combination of deep learning and a traditional method and relates to the fields of computer vision and mode recognition. According to the method, two grayscale images are both used as training samples of a network, corresponding data amplification, including elastic deformation, smooth filtering, etc., is done against the problem of less retinal image data, and wide applicability of the method is improved. According to the method, an FCN-HNED retinal vessel segmentation deep network is constructed, an autonomous learning process is realized to a great extent through the network, convolutional features of a whole image can be shared, feature redundancy can be reduced, the category of multiple pixels can be recovered from the abstract features, a CLAHE graph and a gauss matched filtering graph of the retinal vessel image are input into the network, an obtained vessel segmentation graph is subjected to weighted average, and therefore a better and more intact retinal vessel segmentation probability graph is obtained. Through the processing mode, the robustness and accuracy of vessel segmentation are improved to a great extent.

The invention discloses a retinal blood vessel segmentation method based on the deep-learning adaptive weight, comprising: performing sample expansion on retinal blood vessel images and grouping the samples; constructing a full-convolution neural network for blood vessel segmentation; pre-training the network through the training samples; performing the global adaptive weight segmentation on the retinal blood vessel images; obtaining the initial model parameters for retinal blood vessel segmentation; adding a conditional random field layer to the last network layer and optimizing the network; using the rotation test method to input the testing samples to the network; and obtaining the retinal blood vessel segmentation result. The blood vessel segmentation full-convolution neural network structure and the adaptive weight method proposed by the invention can realize the human-eye level image segmentation and can test on two internationally published retinal image databases DRIVE and CHASE_DB1 with the average accuracy of 96.00 % and 95.17% respectively, both higher than the latest algorithm.

The invention discloses a blood vessel segmentation method for a liver CTA sequence image. Firstly contrast enhancement and noise smoothing preprocessing are performed on an inputted three-dimensional liver sequence image; then liver blood vessels and the boundary thereof are enhanced and blood vessel centers are thinned by adopting OOF and OFA algorithms; seed points of the blood vessel center lines are automatically searched according of the geometrical structure of the blood vessels, and the center lines of the liver blood vessels are extracted so as to construct a liver blood vessel tree; and finally the liver blood vessels are preliminarily segmented through combination of a fast marching method and corresponding blood vessel and background gray scale histograms are calculated, and accurate segmentation of the liver blood vessels is realized by adopting an image segmentation algorithm. The liver blood vessels can be effectively and accurately segmented by fully utilizing the geometrical shape and gray scale information of the blood vessels for aiming at the CTA sequence image which is low in contrast, high in noise and fuzzy in boundary. The blood vessel segmentation method for the liver CTA sequence image can be popularized to other three-dimensional blood vessel segmentation.

The invention provides a retinal fundus vessel segmentation method based on a deep multi-scale attention convolutional neural network. An internationally disclosed retinal fundus vessel data set DRIVEis adopted to perform validity verification: firstly, dividing the retinal fundus vessel data set DRIVE into a training set and a test set, and adjusting the picture size to 512*512 pixels; then, enabling the training set to be subjected to four random preprocessing links to achieve a data enhancement effect; designing a model structure of the deep multi-scale attention convolutional neural network, and inputting the processed training set into the model for training; and finally, inputting the test set into the trained network, and testing the model performance. The main innovation point ofthe method is that a double attention module is designed, so that the whole model pays more attention to segmentation of small blood vessels; and a multi-scale feature fusion module is designed, so that the global feature extraction capability of the whole model on the segmented image is stronger. The segmentation accuracy of the model on a DRIVE data set is 96.87%, the sensitivity is 79.45%, thespecificity is 98.57, and the method is superior to classical UNet and an existing most advanced segmentation method.

The invention relates to a CT image lung lobe segment segmentation method, device and system, a storage medium and equipment. The method comprises the steps that an output lung contour in a CT image is detected, wherein the lung contour comprises an intra-lung region and an extra-lung region; in the lung contour, a machine segmentation method is used to screen out the intra-lung region, and the intra-lung region is used as a candidate region; in the 3D level of the candidate region, blood vessel segmentation and lung fissure segmentation are carried out on lung segments and lung lobes simultaneously; according to the result of blood vessel segmentation, a blood vessel tree is constructed to acquire the three-dimensional blood vessel distribution of the lungs; and the blood vessel tree andthe result of lung fissure segmentation are combined, and lung lobe segment segmentation is carried out to acquire the final segmentation result of the candidate region. According to the CT image lunglobe segment segmentation method, device and system, the storage medium and equipment, errors are effectively reduced; the diagnosis rate and accuracy rate are improved; and the limit of individual lung shape differences is eliminated.

A method for segmenting tubular structures in digital medical images includes extracting a subregion from a 3-dimensional (3D) digital image volume containing a vessel of interest, identifying potential vessel centerpoints for each voxel in the subregion by attaching to each voxel a tip of a 3D cone that is oriented in the direction of the voxel's image gradient and having each voxel within the cone vote for those voxels most likely to belong to a vessel centerline, selecting candidates for a second vote image that are both popular according to a first vote image, as well as being consistently voted upon by a radius image, reconfiguring the subregion as a graph where each voxel is represented by a node that is connected to 26 nearest neighbors by n-link edges, and applying a min-cut algorithm to segment the vessel within the subregion.

The invention relates to segmenting blood vessels in perfusion MR images, more particularly, the invention relates to automated vessel removal in identified tumor regions prior to tumor grading. Pixels from a perfusion related map from dynamic contrast enhancement (DCE) MR images are clustered into arterial pixels and venous pixels by e.g. a k-means class cluster analysis. The analysis applies parameters representing the degree to which the tissue entirely consists of blood (such as relative blood volume (rBV), peak enhancement (ΔR2max) and / or post first-pass enhancement level (ΔR2p)) and parameters representing contrast arrival time at the tissue (such as first moment of the area (fmAUC) and / or contrast arrival time (T0)). The artery and venous pixels are used to form a vessel mask. The invention also relates to a computer aided diagnostic (CAD) system for tumor grading, comprising automated tumor segmentation, vessel segmentation, and tumor grading.

The present invention relates to an FCN retina image blood vessel segmentation through combination of depth separable convolution and channel weighing. The method comprises the steps of: 1) performingCLAHE and Gamma correction to enhance a contrast ratio of a green channel of an eye bottom image; 2) in order to adapt network training, performing partitioning of the enhanced image to expand data;and 3) replacing a standard convolution mode with depth separable convolution to increase the width of the network, and introducing a channel weighing module to explicitly perform modeling of a dependency relation of a feature channel so as to improve the separability of the features. The depth separable convolution and the channel weighing are combined to be applied into the FCN network, an expert manual identification result is taken as supervision to perform experiment in a DRIVE database. The result shows that: the method can accurately perform segmentation of the retina image blood vessels and has high robustness.

The invention provides a CNN based egg embryo classification method. The method comprises that egg embryo images in nine days are collected and divided into two types of samples according to the survival state of an embryo; the embryo images are preprocessed, ROI are extracted from the images, and the sizes of the images are normalized; a channel weighting and combined monitoring combined CNN network structure (SJ-CNN) is used to train a target set; and a trained model is used to classify the images to be detected. Compared with a traditional scheme, a complex image processing process is avoided and is not influenced by incomplete image vessel segmentation or violent noise, the method is highly adaptive to a new sample, and high-precision classification can be realized.

The invention discloses a method and a system for processing and analyzing a blood vessel image, and provides a quantitative method for correlational researches such as researches about geometric and three-dimensional network characteristics and laws of a blood vessel network. According to the technical scheme, the method includes the steps of firstly, inputting a three-dimensional blood vessel image to the system, then obtaining relevant various characteristic parameters through isotropic interpolation, area-of-interest extraction, appropriate threshold-value blood vessel segmentation, connected region extraction, small-region elimination, morphological processing, blood vessel network central line extraction and three-dimensional image reconstruction and visualization operation, and sending and displaying a result to a user, and the characteristic parameters comprise blood vessel areas, blood vessel sizes, blood vessel lengths, the blood vessel segment number, blood vessel diameters, blood vessel density distribution, the loop blood vessel number in the blood vessel network, loop relevant segment ratios, weighted average series and the like.

The embodiment of the invention discloses a blood vessel and eye fundus image segmentation method, device and equipment and a readable storage medium, and relates to the computer vision technology ofartificial intelligence. Specifically, the method comprises steps of acquiring a blood vessel image to be segmented, such as a fundus image; performing feature extraction on the blood vessel image such as the fundus image to obtain high-level feature information; performing dictionary learning on the high-level feature information based on a preset dictionary to obtain dictionary representation corresponding to the high-level feature information; selecting a plurality of channels of the high-level feature information according to the dictionary representation to obtain target feature information; fusing the target feature information with the high-level feature information to obtain channel attention feature information; and segmenting blood vessels in the blood vessel image, such as the fundus image, according to the channel attention feature information to obtain a blood vessel segmentation result. According to the scheme, global information loss of the characteristic blood vessel image such as the fundus image can be avoided, and the segmentation accuracy of the blood vessel image such as the fundus image is greatly improved.

The invention provides a retinal blood vessel morphology quantization method based on a connected region. The method obtains a retinal blood vessel segmentation image after the fundus image is preprocessed, and then performs post-processing on the blood vessel segmentation image. On this basis, the vascular network is thinned and boundary treated, and the vascular centerline network and vascular boundary map are obtained. Corner detection is then performed and removed from the vascular centerline network so that the vascular segments of the vascular network form separate communication areas. Traversing is performed on the blood vessel segment, approximate the centerline of the blood vessel segments, and the blood vessel segment is changed into a broken line to calculate the direction of the blood vessel. At last, that initial diameter value is calculated, the center of the circle is selected by sliding on the centerline of the blood vessel segment, a semicircle window is created according to the direction of the circle cardiovascular and the diameter value measured in the early stage, and the distance between the window and the two intersection points of the blood vessel boundary is taken as a new diameter value. From this iteration, a set of vessel diameter values are measured, and the median value is the vessel diameter of the vessel segment. The invention is applicable to the quantification of large-scale retinal blood vessel morphology, and has high reliability.

The invention provides a three-dimensional lung vessel image segmentation method based on a geometric deformation model. The method comprises the following steps: (1) determining vessel segmentation computing regions according to the physiological structure characteristics of a human body, wherein region selection completely covers targets to be segmented and the shape characteristics of the regions are stable, thereby avoiding computing a global region and improving segmentation speed; (2) computing the mean value of the vessel regions and positioning internal and external homogeneous regions of the targets; (3) computing vessel edge energy and evolving a curved surface along second derivatives in an image gradient direction so that the curved surface is accurately converged to a target edge; (4) correspondingly establishing a three-dimensional vessel segmentation curved surface evolution model and effectively combining the mean value and edge energy of the internal and external regions of the lung vessels; and (5) adopting optimized level set evolution for obtaining solution according to the established deformation model and impliedly solving a curved surface motion according to the level set function curved surface evolution. A large quantity of lung CT image experiments proof that the method provided by the invention has the advantages of rapid and accurate lung vessel segmentation and strong robustness.

The invention provides an ocular fundus image registration method estimated based on distance transformation parameter and rigid transformation parameter, comprising the following main steps: (1) ocular fundus images are loaded; (2) the optic disk center is extracted to estimate image translation parameters; (3) gradient vectors of pixel points in the adjacent zone of the optic disk are calculated, vessel segmentation is carried out, vessel distribution probability characteristics are calculated, and the estimation of image rotation parameters are obtained by minimizing two probability distribution relative entropies (Kullback-Leibler Divergence); (4) the Euclidean distance transformation of vessels segmented in step 3 is calculated; (5) accurate registration of images is carried out. The invention is a quick, precise, robust and automatic ocular fundus image registration algorithm, and has great application value on the aspect of ocular fundus image registration.

The invention discloses an iterative eye fundus image blood vessel segmentation method based on the distance modulation loss, which comprises the steps of (0) collecting a color eye fundus image to form an original image; (1) performing normalization processing on the original image; (2) iteratively training a dense convolution neural network based on the distance modulation loss; and (3) carryingout iterative blood vessel segmentation by using the trained dense convolution neural network. The method can process color eye fundus images under different collection conditions, can provide interactive blood vessel segmentation experience for ophthalmologists, is more robust for blood vessel detection and provides a reliable guarantee for subsequent auxiliary diagnosis.

The invention discloses a blood vessel seed point selecting method and a blood vessel extracting method in angiography. The method comprises the following steps of providing a tomoscan image of a human body part and obtaining a plurality of candidate blood vessels; establishing an identification function of the candidate blood vessels according to the center points of the candidate blood vessels; calculating the value of the identification function and selecting one or more blood vessel seed points according to the value of the identification function. According to the blood vessel seed point selecting method in the angiography, the blood vessel seed points can be selected without the need of a standard model, and the method has the characteristics of high identification reliability and fast identification speed, and blood vessels can be fast extracted by finally combining a blood vessel fast dividing technology.

The invention provides a segmentation method and device for blood vessels in a fundus image, and a storage medium. The method comprises the steps that blood vessel segmentation based on Hessian matrixenhancement is conducted, and a threshold value is adopted for segmentation of the blood vessels; a multi-directional linear structure element is used for open operation processing on a G channel, morphological reconstruction is conducted for further enhancement, the multi-directional open operation is adopted and the minimum response is taken to obtain a background without a linear structure, and two images are subtracted to obtain a main vascular network; secondly, multi-directional Gaussian smoothing filtering and multi-directional Gaussian-Laplace filtering are conducted successively, andmulti-directional morphology and morphological reconstruction are adopted to enhance and preserve the filtered vascular network; finally an adaptive threshold value is determined to segment the bloodvessels; the segmentation results of the above two stages are combined to merge and make the final repair to obtain a final blood vessel binary image. The blood vessel segmentation method can improvethe sensitivity of blood vessel recognition and has better calculation efficiency without the need to train in advance.

The invention discloses a diabetic retinopathy detection system based on serial structure segmentation. wherein the fundus image acquisition device is used for acquiring a retina fundus image; the data processing device is used for analyzing and processing the acquired fundus image; A data processing apparatus includes: a data processor; Preprocessing function module, Blood vessel segmentation function module, Visual disc segmentation function module, Centrally recessed determination function module, Exudation segmentation function module, and the statistical calculation function module and the doctor diagnosis function module. The data processing device is used for counting the exudation area and calculating the probability of diabetic macular edema lesions in the input fundus image, andfinally, a final diagnosis and treatment scheme is given by combining a statistical calculation result and the fundus doctor according to the divided exudation area and disease probability and combining with the specialty of the fundus doctor. Various related physiological structures of the fundus are systematically considered, a lesion area is segmented, then a diagnosis report is given by a fundus doctor, detection is efficient, lesion detection is more accurate, the workload of the doctor can be greatly reduced, and the working efficiency is improved.

The invention relates to an improved region growing method which is applied to vessel segmentation and extraction in a coronary artery angiography image. The improved region growing method comprises the following steps of: preprocessing the image to obtain an original image capable of directly performing region growth; making a regulation and randomly generating a group of seed points; setting a stack data structure, enabling a newly grown pixel point to enter a stack, and taking out the point previously entering the stack to serve as a current point to be subjected to growth when the current point completes the growth; sequentially performing growth on each seed point, wherein a seed point gray value serves as an average value at a growing initial stage, and calculating a new average gray value when a new pixel point is grown every time along with the growth of the seed points; and completing the growth when no pixel point meeting growth standards exists and no seed point exists. The improved region growing method has the advantages that the seed points are automatically generated, no manual intervention is needed, the local average values around each pixel point serve as growth parameters in a growing process, the coronary artery angiography image with uneven brightness can be segmented, and the efficiency and the accuracy of the image segmentation are improved.

The invention belongs to the technical field of medical image processing and discloses a blood vessel image segmentation method based on centerline extraction and a nuclear magnetic resonance imagingsystem. The method comprises steps of: preprocessing cerebral blood vessel data based on the vesselness filtering of a Hessian matrix; extracting a blood vessel centerline by using a topology refinement method; extracting the features of training samples and test samples by using centerline points as positive samples and non-blood vessel points as negative sample points; by using the feature of the training samples and a corresponding tag training SVM model, using the features of the test samples as the inputs of a trained SVM model and using an output tag as the segmentation method of the blood vessel. The blood vessel image segmentation method reduces workload, improves computational efficiency, requires no manual target calibration or background, fully automatically segments the blood vessel, and greatly improves segmentation efficiency. The blood vessel image segmentation method segments cerebral blood vessels accurately and rapidly without human intervention and can achieve a truepositive rate and a true negative rate as high as 0.85.