30results about How to "Troubleshoot rebuild issues" patented technology

The invention relates to a CT image reconstruction method based on GPU hardware acceleration, which belongs to the field of X-ray CT technology. The software portion of the invention generally comprises a GPU-based CT data preprocessing module, a GPU-based CT data filter module, a GPU-based CT orthogonal protection module, and a GPU-based CT image reconstruction and back projection module. The method can achieve acceleration of CT image reconstruction algorithm by using GPU hardware, and the reconstructed portion and the data processing portion are achieved on GPU. The invention provides a segmentation processing method used for processing larger data, which aims to solve the prior problems of insufficient display memory of GPU and low data transmission speed from memory to the display memory. Different from prior methods, the segmentation processing method only needs a portion of projection data desired for each segment to be reconstructed, thus reducing data transmission and improving the entire reconstruction speed.

The invention discloses a white matter fiber tract reconstruction method based on deep learning, which mainly aims at white matter fiber tract reconstruction. The method comprises the following steps: extracting signal sparsity features of a digital image in a training sample set; importing the signal sparsity features to a convolutional neural network for training, forwardly propagating the classification result, using a back propagation algorithm for the classification error, and getting a best network model; inputting extracted signal sparsity features of an image in a test sample set to a trained network model to get a final prediction result; and finally, describing the result into the orientation and distribution of white matter fibers through continuous curve fitting, and reconstructing a three-dimensional white matter fiber tract. Cross and forked white matter fiber tracts can be constructed accurately to provide help for the clinical research and physiological and pathological mechanisms of white matter fiber tracts.

The invention relates to a method for the three-dimensional reconstruction of a straight-line optical flow field based on an intelligent optimization algorithm, which is characterized by comprising the following steps of: shooting at least three continuous frames of images of an object in a scene by a fixed camera to form an image sequence, extracting the straight-line profile of the object in the scene and taking a straight line as a characteristic; under a perspective projection model, eliciting a relationship between the motion parameter of the straight-line profile of a moving rigid object in a 3D space and a two-dimensional straight-line optical field on a projection image of the moving rigid object, expressing the relationship by a straight-line optical flow equation group and constructing a physical model and a mathematical model of a three-dimensional reconstruction algorithm. The invention has the advantages of realizing the three-dimensional reconstruction and the moving target division of an object with a straight-line profile in a scene quickly, solving the problems of robustness of the three-dimensional reconstruction algorithm and reconstruction of a shielding boundary of the scene effectively and enhancing the accuracy of three-dimensional reconstruction effectively.

The invention discloses a culture-heritage three-dimensional-scene generation method which includes the following steps: step1: a culture-heritage remote-sensing data preparation step: obtaining geographic information data in a culture-heritage region and establishing a three-dimensional base database of the region; step2: a culture-heritage landform modeling and culture-heritage scene modeling step: obtaining texture and geometrical shape data of different objects in the culture-heritage region so as to generate a ground-object three-dimensional model of the region; step3: a culture-heritage three-dimensional spherical-system integrating step: obtaining building process data of the culture-heritage area and insetting landform and scene real-time measurement information of the area into the ground-object three-dimensional model according to a building process sequence and environment background information. The culture-heritage three-dimensional-scene generation method is capable of improving the speed and precision of large-scale culture heritage modeling and providing an integrated dynamic index of a culture-heritage single-body model and an environment model so that the method has complete management and update capabilities.

Provided is a method for acoustically reconstructing a three-dimensional temperature field by taking regard of sound ray bending behaviors. The method includes the steps: arranging a plurality of sound wave transceivers on the surface of a three-dimensional hearth, dividing the hearth into n grids, wherein m effective sound wave paths form among the sound wave transceivers and penetrate the grids in the hearth; calculating the length when each path passes through the grids in a sound wave linear transmission condition, and forming a matrix L; measuring the transmission time of a sound wave on the paths, forming a matrix t, and reconstructing the acoustic velocity and temperature of the geometric centre of each grid; reconstructing a temperature field according to the obtained temperature, obtaining the sound ray track of each patch according to the Fermat principle, calculating the length when each path passes through the grids, and forming a reconstructed matrix L ' taking regard of sound ray bending; reconstructing the acoustic velocity and temperature of the geometric centre of each grid according to the matrix L ' and the matrix t; and repeating the above steps, and obtaining an accurate value that is gradually close to the temperature field through an iteration method. According to the method, a sound ray transmission path can be tracked, and the temperature field reconstruction precision is improved.

The invention discloses an urban scene reconstruction method and system, and relates to the technical field of image processing. The method comprises the steps: firstly extracting sufficient and reliable candidate space planes from the spatial points of an urban scene image through employing a spatial plane fitting method based on connected domain detection; secondly distributing the candidate space planes for all segmented regions obtained through the segmenting of each urban scene image under the frame of minimum MRF energy. The method can iron out the defects that a conventional method is poor in reliability and not complete in reconstruction scene, and is higher in calculating efficiency.

The invention discloses a human body geometric reconstruction method based on Euler field deformation constraint, and the method comprises the following steps: taking a target human body grid as a data source, extracting human body feature points through employing a random forest regression method, and obtaining a target human body skeleton node through calculation; driving the skeleton to deformby using the feature points, and driving the grid to deform by using an LBS method, so that the attitudes of the template model and the target model are roughly aligned; on the basis of a local fieldgenerated by the HRBF, performing rigid transformation driven by local shape difference; performing regional expansion and contraction based on local shape driving; based on SDF, optimizing a target energy function based on target data constraint, human body deformation constraint and level set characteristic constraint in an Euler field, calculating a deformation vector field to obtain an SDF deformation result, and according to the SDF deformation result, driving a template grid to be aligned with the shape of a target model, so that human body reconstruction is realized. According to the method, the robustness and efficiency of three-dimensional human body structural reconstruction are improved, and the requirement of performing human body reconstruction on rough data is met.

The invention discloses an effective stream motion vector calculation method. The method is used for carrying out calculation on stream motion vectors by using two-frame images of a video stream, and aims to obtain the two-dimensional vector of a continuous stream motion. The method comprises the following steps: firstly, counting a histogram of an area at which each pixel is located, and by using a main component matching method for area histograms, completing the initialization on motion vectors; then, carrying out denoising processing on preliminary calculation results by using a multi-resolution sampling method; carrying out smoothing processing on the motion vector of a local module by using a statistical analysis method; and finally, carrying out further optimization on the motion vector so as to obtain a stream motion vector keeping the appearance consistency and the motion continuity. The method disclosed by the invention can be effectively applied to the three-dimensional reconstruction applications of stream, has a characteristic of practical applicability, and can solve the reconstruction problem of stream well; meanwhile, the method also can be further applied to the applications of virtual scene building, thereby solving the complexity problem of vision-based stream reconstruction.

The invention provides a reconstruction device and reconstruction method of a digital elevation model. The reconstruction device comprises an arc foundation interference radar, the arc foundation interference radar comprises a rotating platform and an antenna unit; the antenna unit comprises at least two antennas connected with the rotating platform, the continuous arc movement of the antennas isrealized through the rotating platform; the antennas are used for the simultaneous transceiving of the signal, a radar original echo of an observation target region is observed through the antennas, and the synthetic aperture imaging processing is performed, thereby acquiring a single-lock complex image of the imaging region; the height of the rotating plate is adjustable and used for acquiring asingle-lock complex image sequence of the observation target region under different heights, and acquiring an interference image sequences under different heights; the joint phase unwrapping processing is performed on multiple groups of acquired interference image sequences, thereby acquiring the unwrapped phase diaphragm and accomplishing the precise reconstruction of the digital elevation modelthrough the conversion between the phase and the elevation.

The invention discloses a cone beam X-ray oral cavity tomography image algorithm, and the algorithm comprises the following steps: (1), carrying out the discretization processing of a to-be-reconstructed image through a grid, and giving an initial value (shown in the description); (2), carrying out the projection of the image at each angle on the basis of the initial value (shown in the description), calculating and obtaining a projection simulation value, and building a projection coefficient matrix W; (3), carrying out the comparison of the projection simulation value calculated at step (2)and a real projection value at the same angle in a mutual constraint range of a POCS reconstruction method, obtaining a difference value, and carrying out the iterative correction of the pixel valuesof each grid of the image according to the difference value; (4), repeatedly carrying out step (3) till the iteration is finished, and obtaining a reconstructed image. The algorithm achieves the enhanced imaging effect and a three-dimensional reconstruction effect through the improvement of a conventional iteration algorithm.

The invention discloses a method and device for reconstructing a three-dimensional model from three views, equipment and a storage medium. The method comprises the following steps: acquiring a back projection of each two-dimensional edge in each view in a three-dimensional space; according to the back projection of each two-dimensional edge in each view in the three-dimensional space, three-dimensional candidate edges constructed by all points where the back projections of the three different views intersect are obtained; according to three-dimensional candidate edges constructed by all points where back projections of the three different views intersect, obtaining a ring of an actual surface formed by constructing the candidate edges through a deep neural network; and according to the ring of the actual surface formed by constructing the candidate edges through the deep neural network, obtaining a three-dimensional model obtained by fitting a parameterized model of the ring of the actual surface. According to the method, the problem of reconstructing the three-dimensional model from the three views in the CAD can be solved simply and efficiently, the limitation on the geometric shape type is avoided when the topological information of the object is recovered, and the application range is wider.

A method and apparatus of depth coding using depth modelling mode 1 (DMM1) are disclosed to reduce the wedgelet pattern table size. In one embodiment, a size-reduced wedgelet pattern for a reduced wedgelet pattern table is generated by excluding at least one non-corner adjacent-edge sample for adjacent-edge partition or at least one opposite-edge sample for opposite-edge partition from starting positions or from ending positions. The reduced wedgelet pattern table may also include at least one omitted wedgelet pattern in at least one wedgelet direction category. For the adjacent-edge partition, the starting positions and the ending positions may correspond to every other even non-corner adjacent-edge samples in a first and a second adjacent edges respectively. For the opposite-edge partition, the starting positions correspond to every other even opposite-edge samples in a first opposite edge and the ending positions include all opposite-edge samples in a second opposite edge.

The invention discloses a sparse constraint SAR image reconstruction regularization parameter GCV golden section automatic search numerical computation method. During regularization image reconstruction, selection of regularization parameters is a very important issue; for the selection of the nonquadratic regularization parameters, a conventional selection method is limited; and in order to obtain a high-quality reconstructed image, the regularization parameters are always selected manually. In order to solve the problems above, the invention provides the sparse constraint SAR image reconstruction regularization parameter GCV golden section automatic search numerical computation method on the basis of studying the GCV method. The beneficial effects of the method are that automatic selection of the sparse constraint SAR image reconstruction regularization parameters is realized; calculation amount in calculating the sparse constraint SAR image reconstruction regularization parameters is small through the method, and a better balance is kept between noise suppression and feature preservation; and a more reasonable reconstructed image can be obtained.

A digital camera can only capture the brightness of a real scene in a limited range. Due to quantization and saturation of a camera sensor, texture details in high dynamic range (HDR) data are lost, and great difficulty is brought to subsequent image-based application (depth estimation and the like). As an index of the HDR image reconstruction effect, the depth estimation error caused by the incorrect hue reflected in the LDR image can support the improvement of the learning efficiency of the HDR image generation network. Therefore, an end-to-end framework composed of two connected CNNs is provided, and the end-to-end framework is used for solving the key problem in HDR image reconstruction. A large number of quantitative and qualitative experiments are carried out on benchmark and recently published challenging data sets, and it is proved that compared with the latest single-image HDR reconstruction and depth estimation algorithm, the proposed method achieves extraordinary performance.

The invention provides a lung CT image parameter reconstruction method and system based on deep learning, a terminal and a storage medium. The method comprises the steps: inputting an obtained lung CTimage into a preset progressive up-sampling skeleton network model, and outputting a feature image; inputting the feature image into a 3D convolutional neural network model, wherein the 3D convolutional neural network model is a convolutional neural network model composed of an intra-pulmonary branch, an extra-pulmonary branch and a pulmonary mask branch; classifying each pixel of the feature image through the lung mask branches, determining an intrapulmonary part and an extrapulmonary part, and performing classification marking; respectively inputting the intra-pulmonary part and the extra-pulmonary part of the feature image into the intra-pulmonary branch and the extra-pulmonary branch for feature learning to generate an intra-pulmonary image and an extra-pulmonary image; combining thegenerated images of the intrapulmonary branches and the extrapulmonary branches to form a complete CT generated image; conversion of different parameters of the CT image is realized by using a deep learning technology.

The invention belongs to the technical field of medical CT imaging, and particularly relates to a low-dose CT image reconstruction method. When a traditional algorithm is used for reconstructing sparse low-dose CT scanning data, interference information such as sparse artifacts or noise exists in a reconstructed image due to data incompleteness, and therefore subsequent clinical diagnosis is affected. The invention provides a low-dose CT image reconstruction method. The method comprises the steps of performing image pixel-based feature extraction on a prior image; constructing a sparse prior matrix by using the extracted pixel features; and adding the priori matrix into an orthographic projection process of sparse low-dose scanning data statistical iterative reconstruction, wherein an iterative result is an output reconstructed image of the algorithm provided by the invention. The high-quality CT image is obtained on the premise of greatly reducing the radiation dose.