最新刊期
卷 7 ， 期 9 ， 2002
- 摘要：Virtual endoscopy is a new method of diagnosis using computer processing of 3D image datasets(such as CT or MRI scans) to provide simulated visualization. In order to obtain a physically realistic surgery simulation, it is needed to generate the accurate 3D human organ meshes for finite element analysis(FEA) to simulate serials of actions in the surgery. In this paper, a new algorithm is proposed to create the tetrahedral mesh directly from the segmented volume. Because Delaunay triangulation guarantees the well-shape of the final mesh. We follow the idea and classify our method as an incremental insertion algorithm in Delaunay triangulation category. It is composed of three phases: placements of mesh vertices, Delaunay triangulation and restore of tissue boundary. The tissue boundary contained in the original dataset is kept accurately by the featured point selection. An automatic self-adaptive method is presented to vary the density of mesh nodes according to local features of the segmented volume. The adaptive model generated has the attributes of accurate, small scale and well-shaped which is very suitable for complete 3D finite element solvers.2685|372|0更新时间：2024-05-08
- 摘要：This paper proposes an improved motion retargeting algorithm based on existing captured motion data and inverse kinematics. The basic notion of proposed algorithm is to estimate the end-effector tracjectories of target character by exploiting the ratio between skeleton lengths of the original character and the target character. Since important aspects of the original motion are identified as end-effector constraints which is satisfied during the retargeting process, the retargeted motion preserves desired main features. Moreover, the retargeted motion data behaves quite naturally, since the anthopometric proportions are incorporated in the algorithm. Several empirical tests are given to demonstrate the effectiveness and quality of our algorithm.2818|313|0更新时间：2024-05-08
-
- 摘要：In order to overcome the difficulty in applying the Catmull-Clark subdivision surfaces in engineering, a new algorithm about Catmull-Clark subdivision surfaces based on C-B splines is presented. C-B splines curves are extension of B splines, they depend on a parameterα. By use of characteristics of C-B splines, for example, they can provide exact reproduction of circles and cylinders and they can be generated by subdivision scheme while keepingC2, a new surface scheme is generated. The limit surfaces generated by this surface scheme areC2except at extraordinary points. In conclusion, this method not only solves the problem of the precise representation of standard analytic shapes such as circle encountered by Catmull-Clark subdivision surfaces, but also overcomes the difficulty of generating surfaces on arbitrary topological meshes faced by NURBS. Meanwhile, the shaapes of the subdivision surfaces can be adjusted using controlling parameterαand the particular case (α→0) of this scheme is Catmull-Clark subdivision scheme. An application in engineering garphics demonstrates freedom and efficiency of this algorithm.2561|250|0更新时间：2024-05-08
- 摘要：In the areas of computer vision and CAD/CAM, it is often needed to represent an image or a 3D surface from discrete measured data. A novel algorithm for the shape representation and image reconstruction is presented in this paper, which integrates the theories of optimal approximation and data smoothing. A positive definite functional is set up according to Lagrange multiplier method, and solved by finite element method and Newton iteration method. The shape or image is then constructed on the basis of finite element interpolation. This algorithm combines the smoothing processing technique with finite element method, the influence of the noise in input data is eliminated and reconstructing precision is improved. The formulations to calculate Lagrange multiplier and the relevant equations of eight-node isoparametric finite element were dervied. Effects of the variations in smoothing factor, in the finite element mesh and in the amount of imput data on the reconstructed results were investigated. A Gauss surface and two images of sphere and saddle surface were represented from discrete data with imposed noise, the results show the effectiveness of presented method. To illustrate the applicability of the method, a Morie fringes image of a tensile composite plate containing a hole was reconstructed. The method is conceptually simple and relatively easy and expedient to apply. The number of input data required in the presented method is less than that in numerical interpolation and fitting and the method can be used to the problem of irregular region with coved boundary.2242|300|0更新时间：2024-05-08
- 摘要：Mathematical morphology in its original form is a set-theoretical approach to image analysis. It studies image transformations with a simple geometrical interpretation and their algebraic decomposition and synthesis in terms of elementary set operations. Mathematical Morphology has been applied in many fields, at the beginning, it is only applied in binary images called binary-scale morphology, and then it has been developed to gray-level images called gray-scalemorphology, but there are few researches in color image. In this paper we present a new color vector morphological edge detection methods using a multi-scale approach for detection edge under noisy condition. The goal of edge detection is to detect and localize edge points even under noise condition. Not all edges with various fineness regarding spectral contrast and spatial geometry can be detected by a single operator. In fact, some details that seem to be freak and noisy in one scale may become relevant in other scale. Hence, edges of different fineness are detected using operator at different scale, and then they are judiciously combined to produce all the edges of interest in an image. The experiment has proved this proposed method can detect detail edges in color image. Its superiority has been revealed by comparing with the traditional edge detection methods such as LOG. The experimental results have shown this method is robust to noise.2267|373|0更新时间：2024-05-08
- 摘要：B-spline wavelets have been widely used in multi-resolution modeling for curves and surfaces, but there are some defects in such modeling using semi-orthogonal B-spline wavelets. First, some existing algorithms for multi-resolution curves use uniform B-spline curves to approximate non-uniform B-spline(NUBS) curves. This will result in an error between the new curve and the original one. Second, the constraints in the existing algorithms specify the shape at a continuous-resolution level by interpolating those at integer-resolution levels. If the control point number of a B-spline curve does not match that of a curve at an integer-resolution, that is, 2 i 32168|338|0更新时间：2024-05-08
- 摘要：Reusing Motion Capture data to instruct Virtual Human is an effective way to improve the reality of Human Animation. However, there are gaps between captured motion clips, and motion transition.is used to fill them. This paper presents novel method to meet the kinematic and dynamic constraints that should be satisfied during transition. An analytical IK algorithm is adopted to solve the position constraints and an additional constraint, "Swivel angle" constraint, is added to guarantee the continuity of solution. Since the algorithm only need to search the feasible rangle of "Swivel angle", it improves the precision of the solution and the computation efficiency. This paper also presents a simplified way to simulate balancing during transiton.To meet the balancing requirement during the motion transition, balance pose is computed at "key points" of transition with a BFGS nonlinear constraint optimizing algorithm, displacement mapping is then built and added to original transition motion sequence.2156|303|0更新时间：2024-05-08
- 摘要：The 3D object representation method is a key problem on computer vision. The existent methods of 3D object representation have applied for 3D object recognition. But, they have themselves drawback. The 3D object representation method based on aspect graph is a kind of means on a center of the observant. It completely describes the 3D object by a finite set of 2D views of the object from different viewpoint. According to geometry structure of object, the set of viewpoint on the viewing space is partitioned into regions such that in each region the equivalent structure of the line drawing is consistent. An aspect is defined for the same of qualitative structure of topology of the line drawing. The 3D object representation method based on aspect graph makes 3D object recognition become problem on matching 2D image and 2D aspect. To begin with, The thesis defines a principle of equivalent structure of topology for the line drawing. In addition, it studies the partition theory and structure method of aspect graph for 3D object of transparent under perspective projection. At last, according to the principle of equivalent structure of topology, the aspect graph for 3D object of opaque and aspect are established. At the same time, it is proved for example of 3D object.2428|336|0更新时间：2024-05-08
- 摘要：A new approach, image registration and superposition of 3D reconstruction based on CT images is proposed in this paper. Compared with the traditional approaches, it is simpler and more effective. It can be realized in three steps. Firstly, correcting the distortion of all the 2D CT images, some images between every two serial CT slices should be interpolated if needed. Secondly, every pixel will be endowed with information of depth and transparence according to its spatial coordinates and gray level. Finally, a 3D image will be displayed on the screen. The reconstruction image includes not only information of the surface of the object, but also the inner of the object. The results of experiments indicate that the 3D image reconstructed has only a little distortion and the results of measure by software are very accurate. The experiment was carried out successfully with 33 pieces of serials image of mandible, a accurate 3D image of mandible was acquired. Compared the results measured by software in our reproduction model with those measured by micrometer in true model, the standard deviation of absolute error is only 0 151mm. At the same time, this registration method can be applied to many other cases for 3D image reconstruction for serial sections.3297|331|0更新时间：2024-05-08
- 摘要：Nuclear medical image reflects form and change of organ organizes, using special exploration device to explore human body's radioactive nuclide distribute outside the human body. As the SNR of Nuclear medical image is obviously lower than other medical image(X-ray, CT and MRI), the picture of nuclear medical image is coarse and the resolving power is low, the processing method of nuclear medical image is obvious much more important than other medical images. In accordance with the characteristic of dynamic nuclear medical image, the self- adaptive enhancing and smoothing algorithm has been proposed, which is used for removing the noise of the nuclear medical image, enhancing the contrast of the nuclear medical image and improving the visual effect of nuclear medical image. After doing that, the reading effect of the nuclear medical image is much better. This method that the paper probe into is a kind of effectual method that is suitable for processing the relatively low gray level, the bigger noise interferes of the dynamic nuclear medicine image. However, the static nuclear medicine image should adopt the proper treatment method according to the statistics characteristics ofγpicture.2007|326|0更新时间：2024-05-08
- 摘要：A novel iterative approach is proposed for precise aligning multiple 3D images captured from difference views of a natural object surface in one reference coordinate system. Based on common characteristic of 3D natural object surface, a hybrid principle is used which minimizes the difference between corresponded cell plane's vectors on different view's 3D images and the distances between vertexes and corresponded cell planes of another view. The process of iterative algorithm is described in detail. The formula derivation of algorithm is given. The ways to accelerate the iterative algorithm and the ways to avoid error accumulation are discussed and analyzed in detail. The examples show that the algorithm can reach high speed of alignment and perfect effect of whole surface made by aligned 3D images taken from different views without any restriction, any additional equipment and any additional operation when capturing 3D images.2252|301|0更新时间：2024-05-08
- 摘要：Multi-spectral information fusion is the combination of information acquired from multi-spectral sensors in the same scene to acquire clearer description about the scene by using the temporal and spacial correlation, and their complementarity. The realistic fusion refers to the conservation of spectral features from original multi-spectral images in certain case in the new fused image. In this paper we propose a new image fusion method by analyzing advantage and disadvantage of several common methods and studying those applicable environments. This method integrates several fusion methods according to those applicable environments. Besides we propose a new HVS based image fusion which is suitable for little correlation among the multi-spectral images. The HVS based method can extrude target spectral character. When the correlation among multi-spectral images is strong, the wavelet transform-based or PCA method is used according to the distance of the targets. Experiments show that our integrated method can be suitable for different type scene, and the fusion results are better than that using single fusion method and the result is advantageous to the detection and recognition for targets.2507|333|0更新时间：2024-05-08
- 摘要：In order to merge information from multi-sensor adequately and to improve abilities of information analysis and feature extraction, we study wavelet packet analysis on image data and give a fusion method in pixel level by means of wavelet packet transform in this paper. Wavelet packet transform decomposes an image into low frequency band and high frequency band in different level. Besides decomposing low frequency band in a higher scale, it decomposes high frequency band in higher scale which wavelet analysis does not do. It offers a more precise way for image analysis than wavelet multi-resolution analysis. Through merging images data of different wavebands from multi-sensor to the same object and different resolution images data by applying image fusion method based on wavelet packet analysis, we get fused pictures. The method can fuse details of input images successfully, so that it display information of the each input image perfectly. Comparing with other fused images and appraising them on way of objective and subjective performance, we can draw the conclusion that using this image fusion method can get more satisfactory result than using others.2468|348|0更新时间：2024-05-08
- 摘要：Based on Delaunay triangulation, This paper describes a novel algorithm of thinning text images for fast nicety recognition. The polygonal approximation of images' boundaries is decomposed into Delaunay triangles that represents the topological features of the object. All triangles are classified into three types that generate different local skeletons which are connected to form the whole text images skeletons. Taking full advantage of text images' whole and local information, this algorithm has the advantages of fast speed and nice effect.2441|347|0更新时间：2024-05-08
- 摘要：In this paper the structural properties of alphabetic and digital characters on vehicle license plate are analyzed in detail and three types of feature are extracted for their recognition. These features include (a) global ones, i.e. closed curve in character, which can be used for character coarse classification; (b) minutiae, which are some key points in thinned image such as endings, bifurcations and crossings of strokes and can be used for character precise classification, (c) corners in strokes which can be used to discriminate similar characters. Algorithms for fast detection of closed curve based on theory of graph and minutiae and for corner finding based on external contour lines of binary image are also discussed. The recognition rate of alphabetic and digit character of the system based on these structural features is approximate to 96% and the time for feature extraction and character recognition is about 20ms/character. This shows that the proposed structural features have good discriminating power of similar characters and can be used for the fast classification and recognition of characters on vehicle license plate.3127|463|0更新时间：2024-05-08
- 摘要：Meshing technology is of great importance in the areas of engineering analysis and scientific computing visualization. This paper proposes a anisotropic meshing method, In order to extends the Delaunay method to anisotropic contexts the elliptical distance and elliptical matrix is introduced and defined, and the circum|ellipse of a triangle is also defined. And then the method is applied to the surface triangulation, which takes the first fundmental form of surface as the elliptical matrix. Consequently, the definition of surface triangulation is obtained and an algorithm is proposed. The speed of the algorithm is greatly faster than the previous methods. Our approach can uniformly copy with quadratic and trimmed NURBS surfaces.2381|323|0更新时间：2024-05-08
- 摘要：When a geometric constraint system can not be fully decomposed, numerical solving methods are used, in which Newton-Raphson iteration method is the most popular. However, Newton-Raphson iteration method is not stable. To improve the stability of numerical geometric constraint solving, an homotopy method, named under constrained homotopy, is advanced in the paper especially for under-constrained geometric system. It can be combined with the decomposition of geometric constraint system and can be used together with other solving methods easily, and thus helps to the solving ability of geometric constraint solver. Some key problems of under constrained homotopy, such as construction of the homotopy function, homotopy path tracing and singularity analysis of homotopy path, are discussed in the paper. A pure homotopy method for under-constrained geometric systems is not very effective. To solve this problem, a hybrid Newton-Homotopy method is proposed. It makes use of both the fastness of Newton-Raphson iteration method and the stability of homotopy method and thus improves both the ability and the efficiency of the geometric constraint solver.2239|359|0更新时间：2024-05-08
- 摘要：Occlusion culling and LOD(Level of Details) both are effective techniques in accelerating the rendering process of 3D large scene. An algorithm framework which integrate occlusion culling and LOD is proposed in this paper. It partitions the scene volume into a hierarchical structure during the preprocess stage, and in runtime, it uses occlusion culling in the high level to cull the invisible scene, while view-dependent simplification are used in the low level to show the detail. This algorithm is mainly based on Luebke's work of mesh simplification and Heinrich's work of lazy occlusion grid culling in the image space, it extends the data structure of 'vertex tree' to support occlusion culling and also uses an occlusion grid buffer to reduce the number of expensive occlusion queries at pixel-level. Experiment result shows that significant speedup is achieved through our algorithm.3049|396|0更新时间：2024-05-08
- 摘要：Combining the embedding process with the performance analysis, this paper presents a new image digital watermarking technique, the closed loop digital watermarking in order to make the technique more robust and utility, and gives an implementation approach in wavelet domain. The embedding process is modified continuously according to the results of the performance analysis of watermarking embedding process. If the watermarking can not meet the requirements of detection process or the embedded watermark renders visual artifact of image, the embedded intensity of watermark or watermark itself is modified so as to obtain optimal embedding effect. The convergence of the closed loop techniques is proved. In the performance analysis, the distribution of likelihood function is analyzed, and the mean and the variance of likelihood function in two hypothesis tests and their relationship are deduced. The probabilities of false alarm and detection are also given which are used as the performance measures for the closed loop system. The conclusions are approved by experimental results that this closed loop watermarking technique takes advatage of traditional methods.1891|341|0更新时间：2024-05-08
- 摘要：In this paper a new digital watermarking technique is proposed based on discrete wavelet transform. In order to enhance the robustness of watermark, watermark was based on the content of the original image. Pesudorandom permutation, division and spectrum spread techniques were used to improve the robustness of watermark. Different from the traditional techniques, in our method, watermarks were embedded for two times, the holistic watermark was embedded first and then was divided watermark, and watermark here was embedded both into the low-frequency components of the wavelet transform of the image but also to the high-frequency components. The extraction process extracted both the holistic watermark and the divided watermark in turn. The watermark based on the content of the original image and the watermarks that given beforehand were compared in the same condition and the experimental results showed that this kind of watermark has a more robust than the techniques used the watermarks given beforehand. The experimental results also showed that the capacity of embedded watermark was related not only to the just noticeable difference of the human visual system but also to the characteristics of the watermark itself.2050|318|0更新时间：2024-05-08
- 摘要：In this paper, a genetic algorithm using multi-species and multi-encoding method is proposed to combat premature convergence inherent in Standard Genetic Algorithms(SGA). It involves with three species evolved simultaneously. By using float encoding method and binary encoding method respectively, the first species has stronger local search ability and the second has stronger global search ability. The third species, which called "elitist species",aims to keep the elitist individuals in the evolution process. And at the same time, it evolves too, which will enhances the convergence speed and improves the perfomance of GA. And the migration strategy adopted in the proposed method which immigrates elitist individuals among the three species can keep the population diversity efficiently. This multi-species method can help genetic algorithms to escape from possible local entrapment and obtain good tradeoff between exploration ability and exploitation ability. The experimental results of this method on a series of classical complex multimodal functions have shown its efficience and superiority.2008|311|0更新时间：2024-05-08
- Postal code：100190
- Tel：010-58887035/58887030/58887418 Email：jig@aircas.ac.cn
- Technical support is provided by Beijing Founder electronics co., LTD 京ICP备05080539号-4
京公网安备11010802024621 - It is recommended to read the content of this site in Chrome&IE9+. Please switch to extreme mode in browser 360.
- Cookies We use cookies to help provide and enhance our service and tailor content. By continuing, you agree to the use of cookies.
0