42 results about "Cut" patented technology

The invention provides a wetland ecosystem health evaluation method based on multi-source remote sensing data, and belongs to the technical field of wetland ecosystems, and the method comprises the steps: carrying out the dense dark pixel and Kalman filtering cooperative atmospheric correction of a multi-spectral remote sensing image to obtain a surface reflectance image, and achieving the precise segmentation of a wetland landscape through a graph cut theory, a spectral unmixing model based on an improved Gaussian kernel is utilized to embed physical constraints to invert water quality parameters, a laser radar canopy height priori constraint three-dimensional radiation transmission model is combined to invert vegetation parameters, and a mixed pixel decomposition result is optimized through a Markov random field. A water quality, vegetation and landscape three-dimensional comprehensive health evaluation system is constructed, degradation dominant factors are identified, and the technical problem that it is difficult for multi-source remote sensing data to cooperatively invert wetland ecosystem multi-dimensional health state parameters is solved.

The invention provides a multi-target person image face replacement and skin color coordination method and system, and relates to the technical field of artificial intelligence image processing, and the method comprises the steps: receiving a source face image and a target image sequence containing a plurality of target persons; performing character detection on each target image, and retaining a character detection frame of which the confidence is higher than a threshold value; cutting the figure area based on the detection frame and expanding pixel redundancy to form an expanded cut image including a head and an adjacent skin area; zooming all the expanded and cut images to a uniform resolution ratio and then splicing the expanded and cut images into an optimized long image; and performing semantic segmentation on the optimized long graph, generating a multi-layer mask comprising a core mask layer, an extension mask layer and a coordination mask layer, and performing adaptive edge processing. Through a preprocessing strategy, multi-layer mask control, a style guiding mechanism and a skin color and illumination coordination technology, the quality, efficiency and practicability of a face changing result are improved.

The invention discloses a mask layout processing method and device, electronic equipment and a storage medium, and the method comprises the steps: obtaining a plurality of mask layouts including cutting mask layouts and to-be-processed mask layouts, each cutting mask layout including a cutting pattern, and the to-be-processed mask layouts including isolated cutting patterns; adding an initial pseudo graph corresponding to the isolated cut graph in the mask layout to be processed based on the cut graph in the cut mask layout; predicting a simulation ADI result of the isolated cutting graph based on graph information of the isolated cutting graph in the to-be-processed mask layout added with the initial pseudo graph; when the simulation ADI result does not meet the preset condition, the initial pseudo graph is adjusted, and the graph information is updated based on the adjusted initial pseudo graph until the simulation ADI result meets the preset condition, and a target cutting mask layout corresponding to the to-be-processed mask layout is obtained. The problem that the process manufacturing requirement cannot be met due to the fact that the etching deviation of isolated cutting patterns is small is solved.

Implementations described herein relate to methods, systems, and computer-readable media to cut between coplanar convex polygons. In some implementations, a method includes obtaining the polygons, constructing a corresponding initial coplanar graph, adding that graph into a queue, and cutting each graph in the queue, the cutting including identifying whether the graph includes at least one concave boundary vertex. If there is no concave boundary vertex, the graph is added to a set of result graphs. If there is at least one concave boundary vertex, the cutting selects a starting vertex, finds an optimal cut from the starting vertex to another boundary vertex, splits the graph into subgraphs along the optimal cut, and queues the subgraphs for cutting. The cutting repeats until a termination criterion is satisfied. The method then outputs the result graphs. Using these result graphs may reduce the use of computing resources, such as in real-time applications.

The invention discloses a plate cutting method based on a 3D laser line scanning device, and the method comprises the following steps: arranging the 3D laser line scanning device above a to-be-cut plate, enabling the scanning visual field of the 3D laser line scanning device to cover the to-be-cut plate, and enabling the scanning visual field to cover the thickness of the to-be-cut plate; scanning the point cloud of the to-be-cut plate, then converting the point cloud into a height map, and then converting the height map into a grey-scale map; extracting edges of cut-off parts in the point cloud on the grey-scale map, independently establishing an X array and a Y array for each cut-off part, and storing the X array and the Y array into the arrays according to the edge sequence; the multiple arrays X and Y obtained in the third step are converted into dxf format files, the dxf format files are transmitted to a control system, graphic software is used for opening the dxf format files so that a user can determine an effective area by observing the collected height map, and further cut graphs are determined through the graphic software; and the control system stores the graph as coordinates and imports the coordinates into a cutting device for cutting.

The application discloses a kind of weak supervision learning-based hair defect image detection methods, the method includes the following steps: build detection device, obtain the image of the silk ingot product to be detected, using sliding window cuts out all possible hair image sample, using segmentation model to infer one by one to image sample, obtain sample segmentation graph;Sample segmentation graph is spliced, and the overall segmentation graph after processing integration is handled using morphological method;Hair block in overall segmentation graph and its area, bounding box information are counted;Integrate segmentation graph, the hair information existing on single silk ingot is counted, and the hair defect information of silk ingot is output.The application introduces weak supervision learning paradigm, overcome the performance bottleneck caused by the poor consistency of artificial label under supervised learning paradigm, improve the accuracy of silk ingot hair defect detection, while saving the cost required for manual annotation.

The application relates to a key power transmission section optimization identification method under an extreme event N-K broken line fault scene, which comprises the following steps: S1, converting a power grid into a non-directed graph, searching and deleting a hanging node in the non-directed graph, and finding a cut point of the simplified non-directed graph through a Tarjan algorithm; S2, performing power flow calculation on a large power grid to obtain power flow of each line in the power grid, and establishing a network flow graph of each partition of the large power grid according to a power flow calculation result and a maximum transmission power rating of the line; S3, iteratively searching by excluding each edge in an initial cut set in turn, identifying and deleting redundant cut edges and invalid cut edges, until the search of all minimum cut sets in the network flow graph is completed, and a power transmission section is obtained; and S4, obtaining a key power transmission section with an overload tripping risk when N-K lines are broken. The application can quickly and effectively identify the key power transmission section of the large power grid under an extreme event such as a natural disaster, and ensures safe and reliable operation of the large power grid.

The present disclosure provides a feature point matching method and device for images and a storage medium. The method comprises determining a plurality of candidate matching point pairs of feature points in two images, mapping the plurality of candidate matching point pairs into a network graph, establishing a first energy function representing the energy required for marking the plurality of candidate matching points in the network graph as correct matching point pairs or incorrect matching point pairs based on image disparity information, solving the minimum cut of the network graph as the optimal value of the first energy function by using a maximum flow / minimum cut algorithm, and determining the correct matching point pairs and the incorrect matching point pairs according to the minimum cut. The correct matching point pairs in the images can be determined by using the graph cut principle. This process does not need to rely on the setting of a threshold, thereby improving the accuracy of feature point matching. In addition, the present scheme does not need to perform model training and has strong scene generalization ability. Thus, the effect of feature point matching is improved.

The invention discloses an image segmentation method based on superpixel feature optimization graph segmentation. Comprising the following steps: firstly, acquiring and obtaining an original scar image through a camera, and then preprocessing the original scar image to obtain a preprocessed image; carrying out superpixel segmentation on the preprocessed image and extracting mean feature vectors corresponding to different superpixel areas; performing mask processing on the mean value feature vectors corresponding to all the superpixel regions to obtain rough segmentation masks; and finally, taking the rough segmentation mask as an initialization seed, and performing pixel-level edge refining on the original scar image by using an image segmentation method to obtain a refined segmentation mask. The method has the advantages of high segmentation precision, strong anti-noise capability, no need of manual seed marking and high automation degree, and can meet the actual application requirements of clinical diagnosis and quantitative analysis.

The present invention discloses a method, system, medium, and device for segmenting subscenes in long-strip drone images, relating to the field of digital photogrammetry. The method mainly comprises: constructing a scene graph using connection point constraints between images, simplifying it, and eliminating gross errors; analyzing the camera parameters of a single image to optimize the gradient direction by combining neighboring images that have a connection relationship with the current image; setting connection edge weights based on similarity metrics between image nodes, and achieving robust subscene segmentation using a normalized graph cut method. Implementing the method, system, medium, and device for segmenting subscenes in long-strip drone images provided by the present invention can improve the robustness of subscene segmentation in long-strip drone images and help improve the accuracy of subscene reconstruction.

Systems and methods are disclosed for identifying a set of internal edges on a representation of a network that satisfy a set of demands on the network. The disclosed systems and methods perform a multi-step process of selecting the internal edges. In a first step, an initial set of internal edges can be selected using a clique graph (or in another suitable manner). In a second step, a second set of internal edges can be selected using stream graph(s) (or in another suitable manner). The second set of internal edges can be used when determining network paths that satisfy the demands. When the representation of the network has a cut of two, the disclosed systems and methods can identify a set of internal edges providing a degree of protection against link failure.

The application discloses a kind of offshore area high-resolution three-dimensional reconstruction method and system, including using perspective size invariant feature to match the multi-view angle image of offshore area, according to the feature similarity guide sampling, obtain sparse feature point cloud;Based on structured local binary pattern, the depth of each pixel is obtained using graph cut method, and the sparse feature point cloud is encrypted to generate dense point cloud;Global error optimization is carried out to the point cloud of all view angles, and according to the depth information of the matching feature points, the three-dimensional coordinates of the feature points in space are generated;Strong and weak connection relationship between plane primitives is constructed, and the surface model is obtained by graph structure and energy function.The application can obtain wide area large range panoramic high-resolution information of sea area, construct three-dimensional model, and can intuitively display the space-time big data of offshore area.

To provide a structural member provided with a protective film having excellent particle resistance. [Solution] A structural member (10) comprises a substrate (100) and a protective film (200) covering a surface (S1) of the substrate (100). When a cross section of the substrate (100) cut perpendicularly to its surface (S1) is taken as a cut surface, the shape of the line on the cut surface corresponding to the surface (S1) of the substrate (100) is taken as a surface shape, a function representing the surface shape is subjected to a Fourier transform, and the transformed function is plotted as a graph with wavelength (μm) as the horizontal axis, the average slope of the graph in the wavelength range of 30 μm or less is 0.0050 or less.

The invention provides a metal pipe mold-free bending method and system based on three-dimensional expansion and segmented V-shaped notches. The bending method comprises the following steps that S1, an expansion graph containing material compensation parameters is generated through three-dimensional modeling; s2, a progressive V-shaped notch set is machined in the surface of the pipe according to the unfolded graph; s3, a segmented positioning device is adopted for conducting segmented bending on the pipe provided with the V-shaped notch set so as to complete angle forming; the bending system comprises a modeling module, an unfolding module, a processing module and a bending module which are in signal connection in sequence. Force is applied section by section through the section positioning device, a special mold is not needed, the mold dependence problem is directly solved, and small-batch production does not need to bear the mold cost; the expanded graph is matched with the geometrical characteristics of the to-be-bent area of the pipe, material compensation parameters are combined with the physical characteristics of the pipe and bending process parameters, the size deviation is reduced from the source, and a foundation is laid for follow-up accurate bending.

The present disclosure is directed toward systems, methods, and non-transitory computer readable media that provide self-supervised object discovery systems that combine motion and appearance information to generate segmentation masks from a digital image or digital video and delineate one or more salient objects within the digital image / digital video. The disclosed systems utilize a neural network encoder to generate a fully connected graph based on image patches from the digital input, incorporating image patch feature and optical flow patch feature similarities to produce edge weights. The disclosed systems partition the generated graph to produce a segmentation mask. Furthermore, the disclosed systems iteratively train a segmentation network based on the segmentation mask as a pseudo-ground truth via a bootstrapped, self-training process. By utilizing both motion and appearance information to generate a bi-partitioned graph, the disclosed systems produce high-quality object segmentation masks that represent a foreground and background of digital inputs.

The invention discloses a multi-target merging method in a cross-camera scene based on effective cut points, a medium and equipment, and relates to the technical field of image analysis. Through hierarchical clustering combination, particularly similar targets under different cameras are combined into one cluster; according to the method, the feature similarity between clusters is utilized to carry out threshold constraint and merge into an undirected graph, and a graph analysis method is adopted to segment the undirected graph through effective cut points, so that the false merging phenomenon caused by jump noise is reduced, and the accuracy of multi-target merging in a cross-camera scene is improved.

The present disclosure relates to the field of communication technology, and to a method and device for establishing a separation path. The method includes: constructing an initial graph based on source nodes and destination nodes in a communication network topology, and adding the initial graph to a to-be-verified graph set; performing an iterative process until the to-be-verified graph set is empty: obtaining a current graph from the to-be-verified graph set; if a first path of a source node and a destination node is determined based on the current graph and a second path is determined based on the remaining nodes in the initial graph except for the nodes passed through by the first path, forming a current path group based on the first path and the second path to update the optimal path group, and returning to the step of obtaining the current graph from the to-be-verified graph set; if the second path cannot be determined, obtaining a cut point set of the first path, cropping the current graph based on the nodes in the cut point set, adding the cropped subgraphs to the to-be-verified graph set, and returning to the step of obtaining the current graph; determining two separation paths based on the optimal path group corresponding to the i-th iterative process, where i is the total number of iterations.

The invention provides an image retrieval method and system based on a large model. The method comprises the following steps: extracting text information of an original picture; converting the original picture into a visual feature vector, and converting the text information into a character feature vector; respectively calculating the similarity of the two feature vectors in each set to obtain a visual feature fusion similarity and a character feature fusion similarity; determining similarity measurement of visual feature and character feature fusion; constructing data corresponding to the original picture into an initial picture structure; allocating a weight for each edge of the initial graph structure by using the metric to obtain a simplified coarsened graph structure; performing hierarchical segmentation on the coarsened image structure to obtain a plurality of initial cut edge sets; training the automatic edge cutting set optimizer by using the training data; re-dividing the initial cut edge set by using the trained automatic cut edge set optimizer; and taking the nearest partition as a retrieval result. According to the method, the retrieval accuracy can be improved while the retrieval efficiency is ensured.

This invention belongs to the field of path planning, specifically relating to a method for obtaining the shortest path between nodes. The method first obtains the sets of cut vertices on the original graph through several traversals, and then maps the cut vertex set subgraphs corresponding to these sets of cut vertices into a high-level tree of nodes. Preprocessing and parallel shortest path queries are then performed using the high-level tree with the cut vertex set subgraphs as the smallest unit, thereby obtaining the overall shortest path. This method transforms the shortest path analysis for a large number of nodes in a topologically complex original graph into a shortest path analysis for cut vertex sets. It eliminates a large number of useless nodes when obtaining the shortest path, saving related calculations and reducing computational load, thus improving the speed of shortest path acquisition. Furthermore, preprocessing is performed on the subgraphs formed by cut vertex sets, avoiding the computationally cumbersome problem of directly obtaining the shortest path for complex topological structures of large subgraphs.

The invention discloses a method for solving a feedback set of a network, which relates to the technical field of software engineering, logic circuit design and social networking services, and comprises the following steps of: for a graph G, searching a cut edge of the graph G by using a depth-first search algorithm, and if the graph G has the cut edge, deleting the cut edge; if the maximum degree of the graph G is 0, ending the process, and outputting a feedback set; if the B is an isolated point, deleting vertexes in the B, and if the maximum degree of the B is at least 2, firstly putting a vertex v with the maximum degree in the B into a feedback set; according to the method, the feedback set of the network is directly obtained by alternately deleting the vertex with the maximum degree of the network and deleting the cut edge by utilizing the structural characteristics of the graph, and compared with the traditional solving modes such as a linear programming algorithm and a polyhedron combination algorithm, the solving method does not need to construct an adjacent vector and a feedback set polyhedron firstly, so that the solving efficiency is improved. The problem is converted into a linear programming problem, a feedback set result is directly obtained, solving is simpler and more convenient, and the result is more visual.

The present invention discloses a method for detecting defects on the inner wall of a seamless stainless steel pipe. The present invention comprises the following steps: first, performing edge-constrained SLIC superpixel clustering on the collected inner surface image of the seamless stainless steel pipe to obtain a plurality of superpixel blocks; then, taking the pixel mean of each superpixel block, obtaining the mean image block and mean pixel point corresponding to each superpixel block, and forming a superpixel mean image from the mean image blocks of all superpixel blocks; then, clustering all pixels of the current inner surface image of the seamless stainless steel pipe to obtain defect pixel point clustering results and background pixel point clustering results; finally, using a method based on a graph cut model to iteratively optimize the pixels in the defect pixel point clustering results and the background pixel point clustering results until the optimization is completed, obtaining a defect pixel point set and a background pixel point set, and completing the segmentation of the defect. The method of the present invention does not require excessive sample data and has high edge segmentation accuracy.

The invention relates to the technical field of semiconductors, in particular to a layout design method, a layout, a mask and related equipment.The layout design method comprises the steps that an initial layout where a first main pattern and a plurality of reference cutting patterns are distributed is obtained, the first main pattern extends in the first direction, and the reference cutting patterns extend in the second direction; the first direction is perpendicular to the second direction; determining a to-be-processed cutting graph meeting a preset graph parameter condition from the plurality of reference cutting graphs; based on the size of the first main graph, the to-be-processed cut graph is cut in the first direction, multiple target cut graphs are obtained, and the sizes of the multiple target cut graphs are adjusted; the target layout is determined based on the adjusted multiple target cutting patterns, multiple reference cutting patterns and the first main pattern, the lower limit key size and the lower limit interval are not compressed, the size of the cutting patterns can be increased, the cutting patterns are not lost in the patterning process, and the etching deviation is eliminated as far as possible.

A processor-implemented method for implementing graph cuts for explainability using an artificial neural network (ANN) includes receiving, via the ANN, an input. The input is represented as a graph. The graph includes nodes connected by edges. The ANN determines a graph cut between a source node and a sink node associated with the input by solving a quadratic process with equality constraints. The ANN processes a subset of the input based on the graph cut to generate a prediction.

The invention discloses a method for removing a net-shaped shielding object in an image, and relates to the technical field of image processing, and the method comprises the steps: carrying out the superpixel segmentation of a to-be-processed image based on the energy functions of a color item and a structure item, carrying out the iterative exchange of pixels, obtaining a superpixel set, enabling a fusion energy function to be minimized through an image segmentation method, and obtaining a connected region set; sorting and screening according to the mean value and the variance of the connected regions to form a net-shaped shelter sample set; extracting a joint feature composed of a color histogram feature and a rotation invariant local binary pattern texture probability feature from the superpixel, and performing binary classification based on a support vector machine to obtain a mesh shelter mask; and iteratively repairing the occlusion area under smoothness constraint and edge continuity constraint by adopting a total variation model according to the mask, and outputting a repaired image. According to the method, the information of the image is fully utilized, the net-shaped shielding object can be removed without providing depth information from the outside, additional calculation is reduced, and the slender strip-shaped shielding object is repaired to be smoother through the total variation method.

The invention provides a hyperspectral remote sensing image splicing method based on deformation transmission along a splicing line, and the method comprises the steps: S1, constructing a function in an overlapping region of adjacent images based on the spectral similarity information of the images for a registered hyperspectral image, and solving through a graph cut model to obtain an optimal splicing line; s2, in two adjacent hyperspectral images, performing one-dimensional feature point extraction and matching along a splicing line based on spectral gradient information; and S3, based on the matched one-dimensional feature points on the splicing line of the two adjacent hyperspectral images, obtaining a deformation vector describing geometric position difference between the feature points, solving a deformation vector of any pixel point in the overlapping region of the adjacent images through a geometric relationship, and realizing smooth deformation transmission from the feature points on the splicing line to the global image. And extracting and matching the one-dimensional feature points in the two registered images along the splicing line to obtain the geometric deformation relationship of the typical surface feature boundary on the splicing line.

The invention relates to the technical field of image processing, in particular to a material expanding, cutting and beautifying method and system. The method comprises the following steps: inputting a background image into a preset image external expansion model to obtain a target background image; inputting the main body element information of the target background image and the main body image into a preset layout generation model to obtain a first layout result; according to the first layout result, combining the main body image with the target background image to obtain a combined image; cutting the combined picture according to the target size to obtain a cut picture; inputting decoration element information of the cutting picture and the decoration picture layer into a layout generation model to obtain a second layout result; optimizing the second layout result by adopting a preset refined network on the basis of the cutting graph to obtain a third layout result; and according to the third layout result, combining the decorative image layer with the cut image to generate a target image. According to the method, the semantic consistency and aesthetic feeling of the composition can be ensured.

The invention discloses a large-scale social network graph division method based on a DDQN algorithm. The method comprises the following steps: randomly dividing a social network graph to obtain initial partitions; constructing a graph feature matrix according to the social network graph attributes and the partition information; inputting the graph feature matrix into a Q network for prediction to obtain a division action and an expected Q value; executing a prediction action to obtain a new partition and an actual reward; calculating mean square error loss according to the actual reward and the evaluated Q value, and updating network parameters by using gradient descent; after the division action is executed, the division convergence degree is judged; and if the division converges, stopping division and outputting the partition, otherwise, using the new partition to replace the initial partition to continue. Attribute information and structural features of the social network are fully utilized, and the minimum number of cutting edges is achieved while social network graph partition load balancing is achieved.

Systems and methods are disclosed for identifying a set of internal edges on a representation of a network that satisfy a set of demands on the network. The disclosed systems and methods perform a multi-step process of selecting the internal edges. In a first step, an initial set of internal edges can be selected using a clique graph (or in another suitable manner). In a second step, a second set of internal edges can be selected using stream graph(s) (or in another suitable manner). The second set of internal edges can be used when determining network paths that satisfy the demands. When the representation of the network has a cut of two, the disclosed systems and methods can identify a set of internal edges providing a degree of protection against link failure.

This invention discloses a piecewise planar tank reconstruction method based on iterative graph cut-guided optimization. By embedding high-level geometric priors into the energy function of the graph cut, the selection of facets gains semantic understanding capabilities, and the graph structure and candidate face set are dynamically updated. This allows the graph cut results to drive the update of geometric parameters, and the new parameters, in turn, feed back into the graph model reconstruction, forming a closed-loop refinement. The realized "graph structure-geometric model" joint iterative mechanism can eliminate invalid assumptions round by round and continuously focus on effective facets, upgrading from static local optimization to dynamic global guidance, significantly improving the applicability of graph cut in complex 3D reconstruction. Finally, lightweight topological constraints are applied to the highly refined candidate subset to ensure the watertightness and manifold of the output mesh. This invention significantly reduces the dependence on subsequent topology guarantee modules and the computational burden, making it particularly suitable for robust reconstruction of regular objects in industrial scenarios, such as tanks, pipes, and boxes.

The invention relates to an integrated circuit pin number and geometric coordinate analysis method, which adopts a multi-agent framework, is realized through cooperative workflow of a chart agent, a planning agent and a parameter agent, and comprises the following steps of: receiving an integrated circuit data manual image by the chart agent; a chip packaging graph area in the image is accurately identified and positioned, and a cut packaging graph image is output; a planning intelligent agent receives the cut packaging graph image, identifies the type of chip packaging, and generates a planning instruction for pin coordinate extraction and analysis according to the packaging type; and the parameter intelligent agent receives the cut packaging graph image and the planning instruction, executes geometric parameter analysis according to the planning instruction, and generates a standardized list containing pin numbers and corresponding geometric coordinates thereof. Compared with the prior art, the method has the advantages that the coordinate recognition precision is remarkably improved, and the processing efficiency is remarkably improved.