116 results about "Graph cut algorithm" patented technology

Methods and apparatus for blending regions from multiple images to produce a blended image. An image blending module may obtain multiple digital images of a scene. A base image and a source image are selected, and a stroke is applied to the source image to indicate a desired region which is to be blended with the base image. A region in the source image is identified from the stroke using a segmentation technique such as a graph cut algorithm, and the identified region is blended with the corresponding region of the base image, for example using alpha blending. Additional strokes may be applied to the source image to select other regions to be blended with the base image. A different image may be selected as a source image, and a region from the different image may be selected for blending with the base image.

A method and related system for automatically and efficiently isolating the heart in Computer Tomography (CT) or Magnetic Resonance Imaging cardiac scans is disclosed. The method involves segmenting a heart within a set of volumetric data. In accordance with one aspect of the present invention, the set of volumetric data is processed to determine the minimum value of an energy function having a first term, a second term and a third term. The heart is segmented based on the processing of the set of volumetric data.

A system and method of determining a common neighborhood of users sharing a common activity from a plurality of users is provided. The system and method may be used to predict, for a user in the common neighborhood of users, a potential activity from the activities of at least one other user in the common neighborhood of users.

The invention discloses a medical image Graph Cut segmentation method based on a statistical shape model. The medical image Graph Cut segmentation method mainly solves the problem that low-contrast organs cannot be effectively segmented in a medical image in the prior art, and is implemented by the steps of: (1) establishing the statistical shape model of the low-contrast organs, and acquiring grey scale information; (2) pre-segmenting the low-contrast organs; (3) initialization Graph; (4) and segmenting the low-contrast organs. The medical image Graph Cut segmentation method based on the statistical shape model adds priori knowledge of shapes of organs on the basis of rapid segmentation by adopting a Graph Cut algorithm, reduces the possibilities of over-segmentation and under-segmentation, determines initial positions of the low-contrast organs according to relative relationship between animal organs and animal in-vitro contours, improves the segmentation efficiency, and is a rapid and effective organ segmentation method.

The invention discloses a segmentation method and system for an abdomen soft tissue nuclear magnetism image. The segmentation method comprises the steps that pre-segmentation is conducted on an area to be segmented through an area growing algorithm, then a morphological operator is adopted to conduct expansion and corrosion operations to carry out further processing on the pre-segmentation result, so that the pre-segmentation result forms an original segmentation outline. After rectification is conducted between a shape template set and the original segmentation outline, kernel principal component analysis is conducted, and prior shape information is obtained through a statistics model. The prior shape information is combined with data items of an energy function of a nuclear magnetism image segmentation model, and an energy function is built; a kernel graph cuts algorithm is used for carrying out segmentation on the original segmentation outline and an objective outline is obtained. The segmentation method and system can achieve semi-automatic segmentation, the system is simple, the robustness of the nuclear magnetism image segmentation algorithm is effectively improved so as to enable the segmentation result to be more accurate, and the segmentation method and system for the abdomen soft tissue nuclear magnetism image can be applied to nuclear magnetism image segmentation.

The invention belongs to the field of biomedical image processing and specifically relates to a PET-CT lung tumor segmentation method combining a three dimensional graph cut algorithm with a random walk algorithm. The method comprises the following steps of: performing linear up-sampling on an original PET image and performing affine registration on PET and CT images; calibrating tumor seed points and non-tumor seed point; performing random walk algorithm segmentation on the PET image in combination with the tumor seed points; acquiring a foreground target area Ro completely including a target lung tumor area, using the areas, except the Ro, as a background area Rb of a non-lung tumor area; establishing gauss mixture models for the foreground area Ro and the background area Rb separately; computing energy items according to the gauss mixture models of the foreground area Ro and the background area Rb and obtaining a final segmentation result by using an graph cut algorithm. The method fully utilizes the function information and the PET image and the structure information of the CT image, enables complements between the random walk algorithm and the graph cut algorithm, and achieves an accurate lung tumor segmentation result.

The invention discloses an intracranial hemorrhage area segmentation method based on three-dimensional super voxel and a system thereof. The intracranial hemorrhage area segmentation method is characterized in that a CT image pre-processing phase and an intracranial hemorrhage area segmentation phase based on the three-dimensional super voxel can be provided; according to the CT image pre-processing phase, the format conversion of the two-dimensional CT image sequence can be carried out, the skull structure can be extracted, and the intracranial area can be found; according to the intracranial hemorrhage area segmentation phase, the two-dimensional local CT image can be reconstructed on the three-dimensional space, and the three-dimensional image can be divided into the super voxels having the similar sizes by adopting the super voxel algorithm, and the super voxels can be divided into the foreground part and the background part by adopting the graph cut algorithm. The intracranial structure can be extracted by adopting the pre-processing, and the segmentation can be refined step by step, and the super voxels can be used for the calculation by replacing the pixels, and then the hemorrhage area detection accuracy can be effectively improved. The method and the system provided by the invention are advantageous in that the hemorrhage areas having different reasons, different positions, and different sizes can be effectively detected, and the important function can be provided for the computer-aided medical application in the clinic.

The invention provides a motion object real time extraction method of a Vibe improvement algorithm based on combination of graph cut. The motion object real time extraction method comprises steps of performing retreatment on video orders through a Vibe background modeling algorism, dividing the pixels in the image into a foreground, a background and an unknown type and establishing an incompletion ternary diagram, obtaining an incompletion ternary diagram according to the Vibe algorithm, establishing mapping from an image to the diagram, constructing a goal energy function by combining with a membership degree of the foreground and the background and the similarity degree characteristics of the adjacent pixels, obtaining the gather of the sides and the points which can minimize the energy function through the max-flow min-cut theory and performing image division based on the min-cut, performing a post-processing based on the division result of the graph cut algorithm, marking the motion object in the original image by an outer-connected rectangle, and further detecting the motion object through the modeling match. The invention improves the robustness and accuracy of extracting the motion object.

The invention discloses a PET and CT image lung tumor segmenting method based on graph cut. The method includes the steps that first, PET image data acquisition and CT image data acquisition are conducted, a PET image is sampled, affine alignment is carried out on the PET image and a CT image, and accordingly pixels on the PET image and pixels on the CT image are made to be in one-to-one correspondence; seed point calibration is conducted on tumor locations and non-tumor locations of the images; a tumor golden standard of tumors is obtained with the help and supervision of clinical oncologists; through PET information extraction and CT information extraction, a lung tumor is segmented and tested by confluence analysis of the extracted information of the PET image and the CT image through a graph cut algorithm, and consequently a final testing result can be obtained.

The invention discloses a light field foreground segmentation method and device based on K-means clustering. The method includes the steps: extracting a refocusing image, a polar line plane image anda full clear image from a light field image to be processed; processing the polar line plane image by a structure tensor method to acquire polar line plane depth information; processing the refocusingimage by a discrete cosine response method to acquire refocusing information; acquiring area color features, area geometric features, area corresponding point features and area refocusing features according to each of a plurality of areas divided from the full clear image by a super-pixel segmentation technology; calculating similarity of the areas by K-means clustering; marking a foreground anda background by a graph cut algorithm based on the similarity to acquire foreground segmentation results of the light field image. The foreground segmentation results processed by the method are moreaccurate than those in the prior art.

The invention discloses a robustness auto-partitioning method for an abdomen computed tomography (CT) sequence image of a liver. The robustness auto-partitioning method comprises a data inputting step : in which a CT sequence to be partitioned is input and an initial slice is designated; a model building step in which a liver brightness model and an appearance model are built according to data characteristics of the input sequence,a complex background is suppressed and a liver region is highlighted; and an automatic partitioning step in which the initial slice is rapidly and automatically partitioned through combining the brightness model and the appearance model by a graph cut algorithm, and all slices in the liver CT sequence are iteratively partitioned upwards and downwards by taking the initial partition slice as a starting point according to spatial correlations between adjacent slices. According to the method, the corresponding brightness and appearance models are built with regards to the particular CT sequence, and thus, the liver with a low partitioning contrast ratio, boundary fuzziness and shape irregularity can be effectively and automatically partitioned. Moreover, the auto-partitioning method for the abdomen CT sequence image of the liver can be promoted to automatic partitioning of other abdominal organs, such as partitioning of the abdomen CT sequence image of a spleen and a kidney.

The invention discloses a template-based Poisson fusion image splicing method, system and device, and a medium, and the method comprises the steps: obtaining two to-be-spliced images, and carrying outthe feature point extraction of each to-be-spliced image; performing feature point matching on one to-be-spliced image and the other to-be-spliced image by using a particle swarm algorithm; finding an overlapping region for one to-be-spliced image and another to-be-spliced image; finding an optimal suture line in each overlapping region based on the gray value and a graph cut algorithm; creatinga template based on the optimal suture line; and based on the template and a Poisson fusion algorithm, splicing the to-be-spliced images to obtain a panoramic image.Splicing of multiple images can bebetter realized, and the problems of ghosting and gaps in a splicing result are effectively solved.

The invention provides a Manhattan structure building automatic modeling method based on cuboid fitting scanning three-dimensional point cloud. The method carried out modeling through plane extraction, cuboid-based space division and binary classification based on a graph-cut algorithm. The method enables modeling work to be converted into the binary classification problem based on a cuboid element structure; the method establishes coverage index for reflecting fitting efficiency and establishes a global optimization target energy equation by utilizing a point set on a plane and space relativeposition relation of each cuboid; the method solves the target equation through the graph-cut algorithm to obtain binary classification of the cuboids, and operation processing efficiency is high; each surface of the model is flat, the overall structure is compact and a comfortable visual effect is achieved; and since the model result only keeps cuboid vertexes and facet structure, compared withinput dense three-dimensional point cloud, data volume is compressed greatly, and model achieves obvious light weight, which has a good support function for modeling work in a large-scale building scene.

The invention relates to an image deblurring method based on fuzzy region segmentation, and the method comprises the following steps: S1, taking a peak measure to distinguish a fuzzy region B and a non-fuzzy region U so as to represent the feature difference; S2, segmenting the fuzzy region B and the non-fuzzy region U through a Graph-cut algorithm; S3, carrying out the estimation of an original image region L for the fuzzy region B; S4, carrying out the estimation of a fuzzy kernel k of the fuzzy region B; S5, carrying out the image deconvolution of the fuzzy region B, and obtaining a clear region; S6, carrying out the fusion of the clear region and the non-fuzzy region U again, and obtaining a final deblurring result. According to the invention, the fuzzy region and the non-fuzzy region of a fuzzy image are segmented, and the estimation of the original image region and the fuzzy kernel for the fuzzy region is carried out and then the image deconvolution is carried out, thereby achieving the deblurring of the fuzzy region and obtaining the clear region. The clear region is combined with the former non-fuzzy region, and a final deblurring result of the fuzzy image is obtained, thereby effectively improving the efficiency and shortening the time.

The invention discloses an unmanned aerial vehicle remote sensing image mosaic system based on an adjacent relation model, belonging to the field of remote sensing image mosaic technology of unmannedaerial vehicle. The adjacent distance is taken as the weight of the edge, the minimum spanning tree is used to determine the mosaic order, which can maximize the use of the clustering of adjacent images; after the mosaic order is determined, the images can be incrementally mosaic onto the cumulative results in order to complete the mosaic; the mapping between images can be replaced by a similaritytransformation or affine transformation; after the mosaic images are transformed, the images are processed by a Graph-Cut algorithm to solve the overlapping seams, then put the images together directly, the invention realizes a multi-view remote sensing image mosaic system of an unmanned aerial vehicle, and the matching precision and the splicing time are balanced, which can not only ensure the extraction of enough features, but also reduce the amount of calculation, improve the calculation speed, and make the color more natural and uniform. The system is simple in operation, practical, easyto operate, easy to maintain, and the whole process does not need manual intervention.

The invention proposes a stereo matching method based on graph cut. The method comprises the steps: extracting supporting points from left and right images through employing an MSERDoG operator; enabling the pixel gray values of the supporting points to serve as matching cost and fixed windows which are used as cost aggregation for the matching of the supporting points of the left and right images, and obtaining the parallax of the supporting points; calculating a DAISY descriptor operator of each pixel of the left and right images; enabling the parallax of the supporting points to serve as a symbol in graph cut, and enabling the DAISY descriptor operator of each pixel to serve a data item of an energy function in a graph cut algorithm, wherein the data item is used in the in a graph cut algorithm; finally solving a dense parallax graph through solving the minimum value of the energy function, and achieving stereo matching. The method can improve the matching efficiency under the condition of guaranteeing the imaging matching precision.

The invention discloses a method and system for reconstruction of multiframe image super resolution and relates to the technical field of image processing. The method comprises the steps of using temporary results obtained according to geometric transformation and a filter transfer function obtained according to a fuzzy kernel to construct a transfer function for super resolution reconstruction, adopting a graph cut algorithm to perform minimization solving, and obtaining a final high-definition image, so that the reconstruction effect and reconstruction speed are improved.

The invention discloses a parallax image splicing method based on multiple pairs of binocular cameras, and belongs to the field of image processing and computer vision. The method comprises the following steps: firstly, solving a position relationship between binocular cameras by using a calibration algorithm, and solving a homography matrix between images by using prior information; transforminga camera coordinate system of the depth images by using the internal parameters and the external object parameters of a camera; calculating an overlapping region ROI of the images by using a homography matrix between the images, establishing an energy model, and solving by using a graph cut algorithm; the graph cut algorithm is high in time complexity and depends on the number of nodes of a graph,the image is layered, solved layer by layer and iterated, and a local optimal solution approximate to a global optimal solution is solved; and finally, performing image coordinate system transformation on the depth images by utilizing the homography matrix, and synthesizing splicing seams to realize seamless panoramic depth image splicing. Requirements on memory and hardware are low; the invention is simple in procedure and easy to implement, and reduces the image registration time.

The invention discloses a digital biological organism modeling method based on three-dimensional super voxel description. The method comprises the following steps of firstly, carrying out format conversion, binaryzation and other preprocessing on an original medical image of an organism so as to acquire a physiological-structure peripheral region to be segmented; secondly, reconstructing a two-dimensional image sequence of the physiological-structure peripheral region to be segmented to a three-dimensional space, and using a super voxel algorithm to divide a three-dimensional image into super voxels whose sizes are similar to each other; and then, through a graph-cut algorithm based on a Gaussian mixture model, automatically segmenting the super voxels into two portions of a foreground and a background; and finally, using a surface rendering method to reconstruct the super voxels which belong to a foreground type in the three-dimensional space so as to acquire three-dimensional display of a physiological structure area to be segmented and three-dimensional display of the area in the whole organism. By using the method of the invention, a negative effect of other portions of the organism on a segmentation result is reduced; time is saved; robustness of the algorithm is increased and computation complexity of graph cut is reduced; and segmentation accuracy is increased.

The invention provides an intelligent water level monitoring method based on image recognition. The intelligent water level monitoring method comprises: receiving image data sent by camera equipment in a tested area; detecting and identifying a water gauge part above the water surface in the image data by adopting a deep neural network, and segmenting a water gauge image above the water surface byadopting a graph cut algorithm on the basis; calculating the height of the water gauge above the water surface according to the segmented water gauge image, and obtaining the height of the water gauge below the water surface in combination with the overall height of the water gauge; the water level value can be calculated by acquiring the height of the bottom of the water gauge and the height ofthe water gauge below the water surface of the actual station. According to the method, manual on-site water level survey is not needed, image acquisition is achieved in a remote mode, analysis processing is conducted locally, the water level value is calculated, the manual operation cost is greatly reduced, and the measurement accuracy is also improved.

The invention discloses a method for optimizing radar weak target detection based on a constant side length gradient weighting graph cut algorithm. The method comprises the specific steps that first, short-term unrelated accumulation is conducted on a distance-Doppler sequence after coherent integration; then a target model and a clutter model are used for conducting global optimization target detection on a distance-Doppler video frame based on the constant side length gradient weighting graph cut algorithm under a mixed Gaussian framework, and a target video sequence is obtained; the video frame is judged to be the last frame or not, if yes, detection is completed, if no, a kalman predication method is used for conducting target feature detection on the target video sequence and following and completing target model updating, and a mixed Gaussian updating model algorithm is used for achieving updating of a background model; at last, a just captured next frame video information is read, and the above steps are repeated until all the captured video information is detected. The method for optimizing the radar weak target detection based on the constant side length gradient weighting graph cut algorithm optimizes and improves the detecting results relative to radar weak targets.

An image editing system comprises an input device for inputting an image, a graphical user interface for selecting background and object seeds for the image, and an image processor for editing the image. The image processor has various editing routines, including a segmentation routine that builds a graph associated with the image and uses a graph cut algorithm to cut the graph into segments. The user marks certain pixels as “object” or “background” to provide hard constraints for segmentation. Additional soft constraints incorporate both boundary and regional information. Graph cuts are used to find the globally optimal segementation of the image. The obtained solution gives the best balance of boundary and region properties satisfying the constraints.

Methods and apparatus for blending regions from multiple images to produce a blended image. An image blending module may obtain multiple digital images of a scene. A base image and a source image are selected, and a stroke is applied to the source image to indicate a desired region which is to be blended with the base image. A region in the source image is identified from the stroke using a segmentation technique such as a graph cut algorithm, and the identified region is blended with the corresponding region of the base image, for example using alpha blending. Additional strokes may be applied to the source image to select other regions to be blended with the base image. A different image may be selected as a source image, and a region from the different image may be selected for blending with the base image.

The present invention relates to a three-dimensional matching method based on interactive image segmentation. An object in which a user is interested is segmented by adopting a graph-cut algorithm based on the interactive image segmentation; an optimal network graph is generated by a segmentation template; the graph-cut algorithm is called to carry out minimum segmentation calculation on the network graph; a disparity image is generated; and a three-dimensional matching result is obtained. According to the present invention, segmentation information is effectively utilized; the calculation is reduced; the accurate matching result is obtained by utilizing the characteristic of global optimum of the graph-cut algorithm; an experiment result shows that the method disclosed by the present invention obtains excellent effects on shielding, depth discontinuity, illumination variation, low texture, no-texture regions and the like and is obviously improved on the aspects of operation time and operation accuracy.

The invention discloses a change detection method of remote sensing image based on Graph-cut and general gauss model (GGM), which solves the problems that the threshold value for difference image classification in prior art is difficult to determine, and the difference image structured by ratio method is difficult to analyze. The implementation process includes that: (1) the difference images are structured by logarithm ratio method; (2) the difference images are primarily classified through the Graph-cut algorithm; (3) the primary classification result is clustered through FCM algorithm; (4) the EM algorithm is adopted to estimate category parameters of the general gauss model (GGM); (5) the categories for pixels are judged according to Bayesian decision to obtain change detection results. As shown in the experiment, the invention has high detection accuracy, less false detection and strong practicality, and is applicable to hazards assessment and land utilization of sensing images.

The invention relates to a video splicing method based on suture line space-time optimization and belongs to the field of video processing. The method comprises the following steps of detecting feature points through a SIFT algorithm and aligning the feature points, optimizing the feature points with RANSAC, and calculating the tie distance of the feature point pairs in the vertical direction, andpre-aligning a video frame; and based on a graph cutting algorithm, and constraints such as foreground, edge, parallax and the like, calculating the optimal suture line; smoothing the suture line sequence by means of the foreground detection and Gaussian filter; and performing quality evaluation on the suture line sequence; performing linear fusion on images on the two sides of the suture line byutilizing the quality of the suture line to obtain a panoramic video. Compared with methods in the prior art, the method reduces ghost images of the panoramic video, maintains the frame continuity ofthe video, and improves the visual quality of the video.

The invention provides a stereo matching algorithm based on color information and graph-cut theory, and relates to the field of computer vision. The method comprises the following steps: to begin with, under an RGB color space model, carrying out matching by combining template weight and an SAD algorithm and judging an SAD optimal matching point by utilizing color difference constraint conditions, thereby enabling SAD initial matching accuracy to be improved, suppressing noise and meanwhile, keeping scene detail information; carrying out occlusion detection by utilizing left-right consistency check criteria, thereby improving matching precision of an initial matching occlusion region; and meanwhile, for the problem of not high matching precision of initial matching textureless regions and parallax discontinuous regions, constructing an energy gird chart and an energy function through a graph-cut algorithm and carrying out global optimization to obtain a high-precision parallax result. The method improves the situations of not high precision of a local stereo matching algorithm and weak real-time performance of a global stereo matching algorithm, improves matching precision of special regions of the discontinuous regions and the like, and reduces algorithm matching time.

A method for improving classification results of a classifier including receiving classification results for a plurality of elements that have been classified by a classifier as one of a plurality of classes, constructing a graph having a plurality of nodes, each node corresponding to one of the elements, and a plurality of labels, each label corresponding to one of the classes, adding edges between nodes corresponding to related elements, adding edges between each node and each label, and using a graph cut algorithm to cut edges to a node and partition the graph into classes, the graph cut algorithm using as input the classification results for the element corresponding to that node and related elements.

The invention discloses a method for segmenting images by the aid of automatic weight selection, and belongs to crossing fields of computer vision, computer graphics, image processing and the like. The method includes that partial foreground and background pixels are interactively specified via a user interface of an application program; color models of specified partial foreground and background are established, and a graph and a corresponding energy function are constructed; the energy function comprises color constraints, gradient constraints and weights for adjusting the color constraints and the gradient constraints, and each color constraint is defined at a corresponding node in the graph; the nodes of the graph can be pixels of an image or super-pixels formed after the original image is segmented; the validity of each color constraint and the validity of the corresponding gradient constraint is evaluated at the corresponding node of the graph, so that the weight of each node can be determined; the minimum value of the function is solved by a graph-cut algorithm, so that a segmentation result is obtained. The method has the advantages that the method for segmenting the images by the aid of automatic weight selection is implemented for the first time, and a segmentation effect of the method is excellent as compared with the traditional method implemented by the aid of fixed weights on the premise of identical interaction.