66 results about "Curve evolution" patented technology

The invention discloses a method for extracting target closed contour based on shape prior, belonging to the field of computer application technology. From the point of view of simitating the information of human vision, the extraction method inhibiting salient edges of noise and texture is adopted, the shape prior is blended on the basis, and a new target closed contour extraction algorithm can be provided by utilizing curve evolution technology based on differential geometry, so that the important technology is provided for automatic target detection. The experiments on a plurality of imagelibraries prove that when the image background is very disordered and has partial shading, and part of the target boundary is indistinct caused by weaker light, the method can obtain the whole targetclosed contour. Simultaneously, when only a few of templates of a certain category of objects are provided, the method can extract the contour of the category of objects with different postures, namely, the method has invariance property for elastic deformation to a certain extent, so as to be directly applied to intelligent image processing. The method can accurately extract the target closed contour and has wide applicability.

Developable surfaces are generated by interactively evolving curves on a 2D surface embedded in 3D space using an iterative process to produce a model for the construction of a stylized three dimensional sculpture. Each iteration includes tessellating loops formed by the curves on the surface and unfolding the resulting tessellated strips onto a 2D plane. Collisions between the unfolded tessellated strips in 2D and between the loops in 3D are resolved, and the rendered evolved curve in 3D and tessellated strips in 2D are displayed simultaneously during the evolution. A graphical user interface allows the curve network, tessellated strips, and curve evolution parameters to be modified by a user. 2D textures may be mapped to the unfolded tessellated strips, and a set of tool paths may be output for cutting a 2D material.

The invention belongs to the technical field of medical CT (computed tomogram) image segmentation, and relates to a tooth CT image segmentation method based on deep learning. According to the technical scheme provided by the invention, a traditional Level Set algorithm and a U-net network model are combined, and a Level Set algorithm is used for solving the problem of a training set required by the neural network, so that the neural network can be trained by using unmarked tags, at the same time, the neural network model is used to complete the automatic segmentation of the image, the problemof non-convergence of curve evolution is avoided, and a sufficient and accurate segmentation effect can be obtained under the condition that a medical image training set is insufficient.

The invention discloses a narrowband constraint-based local segmentation method for a geometric activity contour model and belongs to the field of image processing. The method comprises the followingsteps of constructing a narrow-band range based on particle swarm threshold segmentation and morphology expansion; constructing an energy function of fusing a global energy item and a local energy item by a self-adaptive coefficient; and solving the energy function by adopting the level set method. According to the invention, the local segmentation calculation region is optimized, and the self-adaptation coefficient of the global energy term and the local energy term is achieved. Therefore, the efficiency and the accuracy of the local segmentation of a grayscale uneven image are improved. Theproblems that the narrow-band control is unstable and the curve evolution precision is insufficient during the local segmentation of the grayscale uneven image are solved. The problem that the energyfunction model fitted by the global energy item and the local energy item cannot be rapidly and accurately segmented is also solved.

Complex labyrinth curves are interactively generated by an iterative process that spatially modulates curve evolution by an image or other function defined on the evolution space. After curves and evolution parameters are initialized [100], the iterative process allows the curve and evolution parameters to be interactively modified by a user [102], resamples the curves [104], and spatially modulates the curves according to the evolution parameters [106]. The evolved labyrinth curves are may be processed [110] for use in various applications including animation, maze creation, intricate artistic patterns, and graphical user interfaces that map linearly ordered data to the evolved curve and allow the data to be navigated using the rendered curve. The evolved curves can also be triangulated and projected to a plane to create patterns for manufacturing developable surfaces.

The invention provides an image segmentation active contour method based on global and local information self-adaptive adjustment. The method comprises the following steps of (1) defining a novel self-adaptive balance function, wherein the weight of each part can be automatically regulated by using the novel self-adaptive balance function according to the self characteristics of an image, and further the curve evolution is driven; (2) in a weight function, adding a Gaussian filter process for regularizing a level set function, and meanwhile, adding a decreasing factor for accelerating the curve evolution speed; and (3) ensuring the precise calculation and the stable evolution of a model through introducing a penalty term. The method has the advantages that good segmentation effects are achieved on the segmentation precision and the processing speed; and the segmentation on a heterogeneous image with nonuniform gray level distribution can be realized.

The invention relates to a demoscopy focus automatic segmentation method in a medical image processing technology. The method includes the following steps: detection and extraction of a focus area; preliminary segmentation of a focus area and a background area in an image and extraction of the shape, position and contour of the focus area; segmentation of a mobile contour model; and construction of an initial curve of the mobile contour model according to the contour of the focus area and establishment of a corresponding energy functional. According to a level set method and a curve evolution theory, a curve capable of meeting minimum of the energy functional is obtained through partial-differential solution and the focus area is segmented automatically from the image. The method performs automatic detection and segmentation on the focus area according to the difference of the demoscopy focus area and the surrounding area; the method is accurate and high in efficiency in focus segmentation, great in practicality, dispensed with manual interruption and capable of assisting dermatology doctors to perform more accurate focus analysis and diagnosis.

The invention relates to a flowage line extracting method based on Canny edge detection and active contour model, and belongs to the technical field of image processing. According to the characteristics of GAC (Grammar Applicative and Cognitive) model and edge detection operator extracting, the Canny edge detection result is integrated into the reconfiguration boundary stop function of GAC model. Compared with the traditional GAC model, the method is more reliable, and even at the weak edge of flowage line and the severe depressed position, the ideal extracting effect can be obtained. As the method inherits the characteristic of high accuracy of Canny edge detection, the extracting accuracy is high. Due to an improved curve evolution speed, the extracting efficiency is higher than that of the traditional GAC model.

The invention discloses a local Gamma fitting-based active contour synthetic aperture radar (SAR) image segmentation method. The SAR image segmentation precision is low in the prior art. The method comprises the following steps of: representing the local structure of each pixel in an original SAR image in a certain neighborhood, regarding a maximum likelihood as a regional separation criterion, and defining a local fitting energy term; then integrating all pixel points in the defined region to acquire an overall optimized energy functional; and finally, describing and solving an energy minimization process by adopting a curve evolution theory and level set method-based geometrical active contour model so as to realize effective segmentation of the SAR image. The method has strong speckle noise inhibiting capability, can realize effective segmentation of the SAR image, and particularly has accurate segmentation effect on complex boundaries such as deep indentation and the like.

The invention relates to an eye fundus image vessel segmentation method based on a local enhancement active contour module. The method comprises: according to a feature vector of a Hessian matrix, vessel enhancement in an eye fundus image is carried out; curvature distribution statistics is carried out on the enhanced image to obtain an eyeball edge and get rid of the eyeball edge; and with a local enhancement active contour module, enhanced image segmentation is carried out by combining global energy information, thereby extracting an eye fundus vessel. According to the invention, on the basis of gray level distribution characteristics of the vessel in the medical eye fundus image, a local energy function is established and a global energy functional item is combined, so that the curve evolution process becomes stable; the speed is accelerated; and the vessel in the eye fundus image can be extracted precisely and effectively.

A computer-implemented method for vertebrae segmentation includes providing an image of a plurality of vertebrae, and determining a seed in each of at least two adjacent vertebrae in the image. The method further includes mapping a unit square to the seeds in the image as corresponding shape constraints on a segmentation, evolving the shape constraints to determine the segmentation of the adjacent vertebrae, wherein evolutions of the shape constraints interact, and outputting a segmented image indicating a location of the vertebra.

The invention belongs to the technical field of mechanical numerical control machining, and relates to a multi-constraint five-shaft machining feeding rate setting method. According to the method, firstly, feeding rate values of sampling points are determined according to chord height difference constraints, cutter shaft angular speed and speed constraints of shafts of a machine tool, and an original feeding rate curve is constructed; by means of a proportional control algorithm, feeding rate values of acceleration or saltus poor points are determined again, so that acceleration and saltus gradually reduce according to the synclastic rule; a curve evolution strategy is adopted, the original feeding rate curve deforms in a one-point constraint or multi-point constraint mode, so that the original feeding rate curve deforms smoothly to a designated feeding rate updating position, and smooth transition of an adjustment area and a non-adjustment area of the feeding rate curve is achieved. Off-line setting of a five-shaft machining adaptability feeding rate is achieved, parallel constraint requirements of five-shaft machining geometric characteristics, process characteristics and machine tool drive characteristics are met, and machining accuracy, quality and efficiency are guaranteed.

The invention discloses a breast ultrasounography image segmentation method and system based on the neutrosophy set theory and level set curve evolution. The method comprises the steps that a breast ultrasounography two-dimensional gray level image is obtained; filtering processing is carried out on the breast ultrasounography two-dimensional gray level image based on the neutrosophy set theory, and the noise effect is removed; fuzzy clustering is carried out on the filtered image in a neutrosophy theory set, and a breast lump candidate area is segmented from the image; the boundary of the breast lump candidate area is used as an initial curve of a level set, and the level set curve evolution is carried out; the internal area of a curve evolution result is used as a suspected breast lump area, the area is assigned to be white, other areas are assigned to be black, and a black and white binary image is used as an output image. The noise effect can be removed from a segmentation result, and accuracy and completeness of lump shapes are kept.

The invention discloses an SAR (Synthetic Aperture Radar) image change detection method based on difference image fusion and an improved level set, which comprises the following steps: firstly, respectively carrying out logarithmic ratio operation and mean ratio operation on SAR images at two moments to obtain a logarithmic ratio image and a mean ratio image, and then fusing the logarithmic ratioimage and the mean ratio image to obtain a fusion difference image; Carrying out pre-classification on the fusion difference image by utilizing a KI threshold value to obtain an initial mark field anda mean value and a variance of the two types of data sets; According to the initial mark field and the mean value and variance of the two types of data sets, calculating the membership degree of eachpixel point on the fusion difference graph belonging to a certain type, and then constructing a self-adaptive kernel function; Controlling curve evolution of a level set function by combining a global energy item and a local energy item according to the kernel function and gradient information of the image; And finally, dividing a change region and an unchanged region according to a curve duringconvergence. The SAR image change detection method not only protects detail information such as edges, but also effectively inhibits speckle noise, thereby improving SAR image change detection precision.

The invention provides a two-dimensional sequence medical image segmentation method based on prior knowledge geodesic geometric flow. Firstly, a layer of two-dimensional image is initially segmented by applying Watershed Algorithm and the rough segmentation result of the initial layer is taken as an initial boundary; subsequently, a resegmentation is performed to adjacent layers of images at the upper lower sides respectively from the initial layer layer by layer in a level set method; wherein, the resegmentation result of each layer is taken as the initial boundary of the next layer and the gradient prior knowledge information is provided to the next layer in the form of adjacent layer gradient reference term for segmentation layer by layer till all of the layers are segmented; finally, the segmentation results of all the layers are combined. By introducing the prior knowledge of the adjacent layer in an edge detection function to improve the stop condition of curve evolution and introducing geodesic geometric flow by taking interlayer gradient similarity as the prior knowledge, the method improves the phenomena of edge leakage when a geometric active contour model is opposite to the discontinuous edge or weak edge in the layers and enhances the precision and stability of three-dimensional medical image segmentation.

The invention belongs to the technical field of digital image processing, and particularly relates to a self-adaptive noise-containing SAR image variational segmentation method. According to the invention, a self-adaptive edge detection operator is introduced to control the diffusion of all-variation rule items, and a noise-containing SAR image segmentation variation model is built according to the reconstructed data items of a multiplicative noise distribution function. The model has the characteristics of non-linear, non-convex and non-smoothness performances, and is difficult to solve. According to a curve evolution theory and a operator splitting method, the minimization energy functional problem is formalized as the minimum value problem with the constraint. Meanwhile, a fast numerical approximation iterative solution method is designed to carry out SAR image segmentation. According to the invention, the self-adaptive noise-containing SAR image full-variation segmentation method is good in robustness for the multiplicative noise of SAR images, and the edge details can be well kept. The noise-containing SAR image segmentation is realized. Meanwhile, the method lays a foundationfor the interpretation analysis and other subsequent applications on SAR images. The method is friendly in application environment and wide in market prospect.

The invention provides a method for using incremental variation level set to segment medical images fast; the method comprises the following steps of: firstly, selecting an initial boundary; adopting fast algorithms such as a narrow band method, etc. to solve the curve evolution process of the level set according to a subregion and the average grey level calculated by the initial boundary; extracting a zero level set, namely a new boundary; judging whether the stop condition is met or not, if so, segmentation results are obtained; if not, the movement of the boundary is used for leading to the change of regions; calculating the average grey level of a new region in the range of narrow band according to the increment; then carrying out the process of using fast algorithms such as a narrow band method, etc. to solve the curve evolution of the level set; finally obtaining the zero level set, namely the segmentation result; the invention adopts the incremental method to solve the average grey level in an iterative mode according to the dynamic change of pixel in the region and the region, and changes an analytical formula thereof into a progressive iterative formula, thus being capable of adopting the fast algorithms such as a narrow band method, etc., improving the segmentation efficiency largely and leading the model to have more practical significance.

The invention belongs to the field of image processing technology and discloses a multi-shape-prior level set independent component analysis method and an image partitioning system. The method comprises the steps that a to-be-partitioned image and shape priors are input; curve initialization is performed; shape prior alignment is performed; the aligned shape priors are coded through a level set function; a shape prior matrix is formed; independent component analysis is used to perform dimension reduction; the current level set function is projected to a low-dimension space; probability distribution of the shape priors is estimated, shape driving energy items are constructed and combined with data driving energy items, and an energy function is formed; and the energy function is minimized,curve evolution is driven, and a partitioning result is obtained. Through the method and the system, high-dimension redundant features of the shape priors can be eliminated, therefore, distribution ofthe shape priors can be subjected to more accurate statistical analysis, more effective shape constraint can be formed, and finally an accurate partitioning result can be obtained.

The invention discloses a multi-feature-based gray uneven image fast segmentation method. Similarity theory fast estimation bias field information is introduced, which simplifies a local information model. The running speed is greatly improved. The sensitivity to initialization contour information is reduced. Compared with a classical algorithm, the method has the advantages that a Heaviside function similar to a step function is constructed; a segmentation curve is more smooth; a dual termination condition is introduced; and according to different self-adaptive end curve evolution processes of an image content, the speed of a segmentation algorithm is improved.

The invention discloses a level set SAR (Synthetic Aperture Radar) image segmentation method based on a self-adaptive finite element, which is mainly used for solving the problem that a conventional variational level set model based on statistical distribution is imprecise in the non-homogeneous SAR image segmentation. The method comprises the concrete implementation steps of: (1) optimizing an image partitioning energy term on the basis of minimum cutset criterion of image partitioning; (2) defining the weighted energy functional through combining with a level set rule term and a length bound term; (3) carrying out variation and minimization on the energy functional to obtain a curve evolution control equation; (4) carrying out discretization on a finite element mesh to obtain a semi-implicit discrete scheme of the curve evolution control equation; and (5) adjusting strategy by adopting the self-adaptive finite element mesh based on posteriori error estimate, realizing the level set evolution based on a triangular mesh and obtaining a segmentation result of the SAR image. According to the invention, the energy functional is defined by utilizing pairing similarity so that the limitation of the conventional statistical model is overcome; in the meantime, the numerical computation strategy based on the self-adaptive finite element is adopted so that the effective balance of segmentation quality and computing efficiency is realized.

The invention relates to a cardiac nuclear magnetic resonance image segmentation method which comprises the following steps of: 1, carrying out Gaussian filtering pretreatment on an image; 2, calculating an external force field of an edge for keeping a normal gradient vector flow; 3, defining an initialization outline position of an inner membrane of the cardiac ventriculus sinister; 4, adding a circular energy constraint in a curve evolution process, and segmenting to obtain the inner membrane; 5, defining a final segmentation outline result of the inner membrane as an initialization outlineposition of an outer membrane; 6, setting the edge strength of a zone enclosed by the inner membrane outline in an original edge graph as 0, and re-calculating the external force field; and 7, addinga circular energy constraint in a curve evolution process, and segmenting to obtain the outer membrane. The invention has the advantages of large capturing range, strong noise proof capacity, better robustness to weak edge leakage, and capability of accurately segmenting the inner membrane and the outer membrane of the ventriculus sinister wall.

The invention belongs to the image processing technology field and discloses an image segmentation method based on a PCA reconstruction error level set. A to-be-segmented image is inputted; image characteristics are extracted; an evolution curve is initialized; base vectors of inner and outer regions of the image are obtained utilizing the PCA; each pixel of the image is reconstructed according tothe basis vectors; a reconstruction error of each pixel of the image is calculated; the reconstruction errors of the pixels are accumulated, and a data-driven energy item is constructed; a novel energy function is minimized, and curve evolution is driven to obtain the segmentation result. The method is advantaged in that the used PCA technology assumed image information satisfies Gaussian distribution with respect to image segmentation of a segmented constant model level set, a non-homogeneous image can be well segmented, and noise robustness is further realized; Gaussian model calculation ismore time-consuming compared with image segmentation of a Gaussian model level set, the operation speed of the used PCA technology is faster.

The invention discloses a multiresolution and multiregion variational level set image segmentation method which comprises the following steps: setting the order of resolution and the number of segmented regions, and carrying out continuous downsampling on an original image in each dimension according to a spatial resolution so as to generate an image with a resolution of 2L; carrying out curve evolution on the image by using a variational level set minimized energy model so as to generate N-1 zero level set evolutionary curve equations; constructing an initialized evolutionary curve by takingthe evolutionary curve (obtained by taking 2i as coefficient) as the next resolution, then carrying out curve evolution on the initialized evolutionary curve by using a multiresolution level set method so as to obtain N-1 zero level set curve evolution equations in current resolution; and finally, repeating the evolution process until the original-resolution image is achieved, and then obtaining the segmentation results. The method provided by the invention has the advantages of avoiding the superposition and missing of the segmented regions, reducing the noise interference, and reducing the search space.