最新刊期
卷 8 ， 期 8 ， 2003
- 摘要：Facial features localization, an important technique in human face analyzing, is specialized in searching for facial features (eyes, nose, mouth, ears, etc) within a given region in an image or image sequence. It finds applications in various areas, such as face detection, face recognition, gesture recognition, expression recognition, face image compression and reconstruction, face cartoon, etc. This paper attempts to give an overview of the latest development in this field by classifying the newly proposed methods into five categories, namely the ones based on knowledge, geometry information, color, appearance and relative location. All these methods appeared in related papers or works published on international, as well as Chinese, journals and conferences in recent years. Then their performances in accurate rate, robustness and computational expense are roughly estimated and compared while some discussions about the criteria on which the estimation and comparison bases are provided. At the end of this paper, a prospect of the future development in this field is presented based on the observation of how an effective face feature location system should be build.4207|195|0更新时间：2024-05-07
- 摘要：This paper presents an effective algorithm for frontal face verification based on the wavelet decomposition technique and Support Vector Machines (SVMs). The process of the proposed method for face verification of M clients consists of two stages.(1) Training stage: by the wavelet decomposition, extracting the appropriate features from the facial images in the prepared training database of faces, training M SVMs for the M clients by the extracted facial feature vectors. (2) Verification stage: selecting a trained SVM from the M SVMs based on the identity claim (such as a name or a password) of an unknown person and using the trained SVM to classify the facial feature vector extracted from the facial image of the unknown person by the wavelet decomposition, and the classification result will show whether or not the identity claim of the unknown person is valid. The ORL database of faces is selected to test and evaluate the proposed algorithm. The results of the test are encouraging and the SVMs in the proposed algorithm are shown to perform very well in classification capability when compared to the traditional radial-basis function networks applied as classifiers in the algorithm.关键词：Computer image processing;Face verification;face recognition;support vector machines;wavelet decomposition;ORL database of faces2647|215|0更新时间：2024-05-07
- 摘要：The poses of 3D object in different viewpoints are different, so the 2D digital images in different viewpoints are different. Therefore, the 3D object recognition is difficult. In this paper, the approach to recognize 3D moving object by image sequence and the transition relationship of the images in terms of the principle of "the winner is the king" is produced. Polar-exponential grid technique maps the image in Cartesian coordinate into polar-log coordinate. The profile of object in polar-log coordinate has translation, rotation and scale invariant by Fourier transform and it can be used as invariant to recognize object. In this paper, the profile invariant is obtained by polar-exponential grid technique and Fourier transform, and 3D moving object is recognized by neural network and aspect graphs. An object recognition system is carried out. The experimental results show that this method is effective to recognize 3D moving object.关键词：Pattern recognition;Polar-exponential grid technique;Polar-log coordinate transform;Profile invariant;Aspect graphs2590|171|0更新时间：2024-05-07
- 摘要：In this paper, the methods for the uncalibrated two-views 3D reconstruction is proposed. The methods employ geometric constraints available from geometric relationships that are plentiful in manmade structure-such as parallelism and orthogonality of lines and planes, these constraints lead to simple method to calibrate the intrinsic parameters of the camera. This is done by determining the vanishing points associated with parallel lines in the world, under the assumption of zero skew and known aspect ratio, three mutually orthogonal directions are exploited to give the camera calibration matrix . It is possible to obtain only a projective reconstruction from the fundamental matrix. If each image is calibrated, it be able to convert from the uncalibrated fundamental matrix to the essential matrix. A Euclidean reconstruction would be preferable. Once the essential matrix is recovered, if the first camera is assumed to be at origin of the coordinate system, then it is a simple matter to calculate the rotation and translation of the second camera relative to the first. After the camera intrinsic and extrinsic parameters had been estimated, camera projection matrices may be recovered and used to estimate the structure. The 3D reconstruction process has two stages: the first to recover the camera positions and motions, the second step involves triangulation to recover the 3D points. The validity of the proposed algorithm is confirmed by experiment for a number of multi-plannar scenes. The reconstructed scene is modeled. New images are generated of the model for new view points. The geometry agrees with our perception of scene. The angle between the two reconstructed planes looks just like a right angle.关键词：Computer image processing;Vanishing points;camera calibration;fundamental matrix;essential matrix;Triangulation;3D reconstruction3024|177|0更新时间：2024-05-07
- 摘要：A novel two-dimensional bar code recognition algorithm based on Fourier transform is proposed in this paper. First, the location and segment technology of two-dimensional bar code is discussed and a single row codeword image is obtained. Then the model of blurred bar code signal caused by the point spread function is given. The first derivative and midpoint of bar code signal is discussed. After analyzing the bar code signal, the standard deviation of the point spread function is obtained. The bar code signal is deblurred based on Fourier transform. At last, after getting the first derivative of signal, the edge strength histogram is used to drop invalid edge caused by noise adaptively. In the edge strength histogram, based-moment threshold selection is used to decide the threshold.The experimental results show that the performance of the algorithm proposed in this paper is excellent. It can achieve higher recognition rate of high density bar code, which suffices to the requirement of the practical use.关键词：Computer image processing;Bar code recognition;Two-dimensional bar code;point spread function;Fourier transform3008|192|0更新时间：2024-05-07
- 摘要：In this paper, a series of techniques are developed to address this problem. First, a new concept, adaptive gradient slice(AGS), is defined for the purpose of transforming a 3D video sequence to a 2D representation. Then image projection and morphological operations are adopted to locate shot transitions in the AGS domain. For the detected transitions, soft templates and corresponding template-matching algorithms are designed to recognize their fine structures. Simulation results show that, with the manually constructed templates, the proposed algorithm can recognize most of the transitions in the experiment very precisely in real time. Only for two transitions quite similar to each other, the proposed algorithm will encounter some difficulties to distinguish them. The reason is that the mapping from 3D sequence to 2D image loses some spatial information. However. In conclusion, this paper provides an effective way to deal with the problem of fine structure recognition of video shot transitions.关键词：Computer image processing;video analysis;template matching;Shot transition recognition2389|182|0更新时间：2024-05-07
- 摘要：Edge detection is an important task in computer vision. It is the front-end processing stage in object recognition and image understanding system. In order to make the detected edges to be well localized, continuous and thin, and robust to noise, this paper presents an adaptive immune genetic algorithm (AIGA)based on cost minimization technique for edge detection. The proposed AIGA recommends the use of adaptive probabilities of crossover, mutation and immune operation, and a geometric annealing schedule in immune operator to realize the twin goals of maintaining diversity in the population and sustaining the fast convergence rate in solving the complex problems such as edge detection. Furthermore, AIGA can effectively exploit some prior knowledge and information of the local edge structure in the edge image to make vaccines, which results in much better local search ability of AIGA than that of the canonical genetic algorithm. Experimental results on gray-scale images show the proposed algorithm perform very well in terms of quality of the final edge image, rate of convergence and robustness to noise.关键词：Computer image processing;edge detection;immune genetic algorithm;Cost minimization2390|175|0更新时间：2024-05-07
- 摘要：Texture segmentation is one of fundamental problems in image processing. An efficient and robust method is needed for various texture samples. Consequently, this paper presents a texton-based algorithm of texture image segmentation. At first, Harr wavelet is utilized as a transform tool so that texture subimages of different directions are produced. And then, we use double threshold technology to get edges and propose a new method to extract textons of four directions. Thereafter, textons are simplified to one dimension vector and reduced to twelve kinds. At last, Statistics and vector field are applied respectively in order to separate texture images from coarse to fine. The algorithm is simple and effective for a wide of textures. Through this way, texture images are not only separated, but also texture structure can be described, which is useful for high-level image processing tasks.关键词：Computer image processing;Texture segmentation;Wavelet transform;Texton;Vector field3930|211|0更新时间：2024-05-07
- 摘要：In the process of building virtual worlds, it is very important to get the textures of irregular objects. Many texture acquirement techniques have been proposed in recent years. However, suffering from costly equipment and the ambiguity of texture presentation, these currently available approaches cannot meet the requirements of the conservation and presentation of digital heritages. To overcome these problems, a new algorithm on generating cylindrical textures of irregular objects is presented in this paper. Based on the homochromatic geometric models of objects and the photo sequences shot around them, the algorithm generates the textures of irregular objects by transferring the color information of the photo sequences to the cylindrical texture maps. It consists of three steps: first, building the mapping relations between the pixels on the photos and the pixels on the texture map by projective transformation, second, acquiring partial textures from respective photos and at last, combining these partial textures into integrated one by mosaicing. The algorithm is runtime-controllable and has low quality requirement of source photos. It has been applied to the Digital Rebuilding and Restoration System of Dunhuang Culture Relics and gets good results.关键词：Computer image processing;Texture acquirement;geometric model;Photo sequence;Digital heritage2375|183|0更新时间：2024-05-07
- 摘要：Shape from shading(SFS) is one of the critical techniques to recover three-dimensional shape in computer vision, which obtains 3D of the visible surface of an object from only one images of its using the shading knowledge in the given picture. The former methods are ill-posed.. In this paper, the algorithm of shape from shading based on BP neural networks is proposed, which is based on Lambertian reflection model. The suggest method has error compensation. We demonstrate the robustness of our approach to strong illumination variations and with significant pose variations. The recovered face shape is then shown along with the original surface .In comparison with the traditional methods, examples show that the algorithm is verified to be accurate and applicable.2640|195|0更新时间：2024-05-07
- 摘要：In the existing IBR systems, they can only provide new view point of the image and have not the function of editing of scene. Because the models of the IBR systems are different from the traditional computer graphics, it is difficult to apply the methods of the traditional computer graphics directly. In this paper, a method is put forward to modify the texture of surface of scene model. This method point out the edit field in the original reference image, define the texture and can modify the texture of every view point. In addition, its advantages also include that it need not the shape, size and so on of the edit field and it considers the brightness of environment to merge the new texture and the original scene better. Proved by experiment, this method can't be limited by the complexity of scene. But because the method can't get the scene's mathematic expression, it can't be rendered by current graphics engineer and can't fit to too large scene.2634|201|0更新时间：2024-05-07
- 摘要：H.26L is a new generation video coding standard being established by ITU-T VCEG and ISO/IEC MPEG for video communication, which is aimed at very low bitrates, real time, low end-to-end delay coding for a variety of source materials. By adopting simple and straightforward approach using well-known building blocks, H.26L achieves coding efficiency higher than the other existing video coding standards greatly. On the basis of mature techniques such as transform on block, motion estimation/compensation, quantization and entropy coding, H.26L employs some new techniques, which ensure the coding efficiency. After introducing the H.26L coding standard, experiments are conducted on different block size and shape combinations, different motion estimation accuracy as well as two entropy coding methods. The experimental results proved that taking compression ratio, PSNR and complexities of encoder and decoder into account, the selection of block partition modes that are 8×8 and larger, 1/4-pixel resolution motion estimation and universal variable length coding can achieve the most economical compression results for the common video contents.2866|180|0更新时间：2024-05-07
- 摘要：To improve the efficiency and robustness of constraint solving algorithm, the paper presents a graph-constructive approach. In constructive phase, the algorithm can classify well-constrained, over constrained and under constrained configurations. By adaptively adding virtual constraints, the algorithm can handle all the above configurations, and can overcome the ill-condition constraint which occurs frequently in practice. Decomposition trees are used to reduce the scale of constraint solving problem, decompose it from global scope into many local sub-problems. By avoid solving large-scale equations, the algorithm can separately calculate the sub-problems geometrically with great improvement in efficiency and robustness. It also improves the efficiency for successive modification on constraints or geometric elements. Comparing to conventional algebra algorithms, symbolic algorithms, and rule-based algorithms, the algorithm is more practical and can be used in real-time constraint solving environment. It has been implemented in VC on Windows/NT platform, and been used as geometric constraint solving kernel of a feature based parametric modeling system GS-CAD, which is a commercial CAD system developed by Zhejiang University.关键词：Computer graphics;Variational design;geometric constraint;constraint solving3121|169|0更新时间：2024-05-07
- 摘要：When graphically modeling plants, the process of L-system changes with different rule definitions. As a general method, a language parser based on L-system rules is presented in this paper. Firstly, this method acquires and constructs the parser of, through induction and abstraction, L-plants which is a language able to define various L-system rewrite rules. Secondly, the start state and the rules of L-system are recognized with the parser and then the rule replacement is executed to produce a string. At last, the shape grammar is used to model a specific plant while it parsing the string. This plant modeling method proves to be highly efficient with experiments and with it users can pay more attention to designing the shape of plants and defining the rules of L-system than to thinking about the details of how to replace in each step. This method also has good expansibility. When users want the parser to recognize other types of rules of L-system, they can simply add some new syntax expressions in syntax definition files.关键词：Computer graphics;L-system;Plants modeling;YACC;LEX2518|182|0更新时间：2024-05-07
- 摘要：This paper applies genetic simulated annealing algorithm (SAGA) to solving the geometric constraint problems. Genetic simulated annealing algorithm itself has many merits, such as implicit parallelism, stability of numerical computation and the global searching ability together with the local fast converging ability, etc. This paper takes the special characters of constraint solving into consideration and integrates the SAGA well with it. After the geometry constraint problem is transformed into the optimal one, it is solved by SAGA that making full use of the advantages of SAGA. This approach can deal with the over-/under-constraint problems because of this conversion. It also has advantages due to its not being sensitive to the initial values over the Newton-Raphson method, and its yielding of multiple solutions is an advantage over BFGS(Broyder|Fletcher|Goldfard|Shanno) for multi-solution constraint system. Our experiments have proved the robustn.关键词：Computer graphics;Genetic simulated annealing algorithm;Geometric constraint solving;Variational design3317|179|0更新时间：2024-05-07
- 摘要：Uncertainty spatial information representation is a critical problem in qualitative spatial reasoning (QSR) and other spatial information process domains. Bittner put forward the rough location method which is better than other uncertainty spatial theories in special circumstance. But his theory wasn't formalized by relation algebra, thus can not be directly applied to spatial database or other spatial information applications. Based on rough set theory, we put forward a relation algebraic formalization for the rough location model and applied it to geographical informational system. Extending the RCC theory of QSR, basic spatial relation is defined by spatial relation matrix. Based on 249 basic relations and the operations on them the approximate spatial relation algebra ASRA is formed. After discussed its character and axiom, we studied the uncertainty of the maping from ASRA to RCC5. Applied it to GIS, a spatial relation judgement algorithm is given. Comparing with similar algorithms, it supports both crisp and approximate regions and is more efficient.关键词：GIS;Qualitative spatial reasoning;rough location;Approximate spatial relation algebra3131|210|0更新时间：2024-05-07
- 摘要：There are many researches and applications about the shortest path searching in GIS field at present. Algorithmic efficiency of the shortest path searching is problem which has came to front and need to be resolved in application. In this paper, Author analyzes the optimization means based on the Dijkstra's shortest path algortithm and brings forward his optimization schemes beeline optimizing Dijkstra's shortest path algortithm from two aspects optimization of algortithm and data storage configuration. At same time, author discusses necessary testifying and application conditions in this paper. This scheme was applied in China Main City Highway Information Searching System, and has got satisfied effects. This paper also offers concerned testing data about Beeline Optimizing Dijkstra's Shortest Path Algortithm.关键词：GIS;Dijkstra;The shortest path2844|223|0更新时间：2024-05-07
- 摘要：In this paper, the Event-based Spatio-temporal Data Model (ESTDM) put forward by D.J. Peuquet in 1995 was analyzed, and its inability of representing multi-attributes is discussed so that a resolution is reached: to introduce "attribute table and index" into the model to link the ESTDM with "external" attributes. Then, one new spatio-temporal data model is designed for land use change and the corresponded spatio-temporal database as well. In addition, a spatio-temporal database is designed for the data model. Based on the new model, a prototype of temporal GIS for land use is designed and implemented. Then the prototype system is tested with land use change data of Jinan city. The experiment indicated that the proposed new model is in right point for the spatio-temporal change of land use since it is specially aimed at land use change, which qualifies the model to hold both high academic and practical value.关键词：GIS;Spatio-temporal data model;event;Land use;Temporal GIS2644|201|0更新时间：2024-05-07
- 摘要：In this paper, we present a method to simulate the realistic effect of the cloth in the 2D virtual environment. In order to simulate the cloth that is in certain environment, we mainly need to solve the two aspects of the problem. Firstly, we need to establish the model of pseudo-curving surface meshes. Secondly, we need to map the texture of the cloth to the pseudo-curving surface. To model the cloth's pseudo-curving surface meshes, we firstly draw the outline of the curvi-edge quadrangle, and then generate the meshes on the pseudo-curving surface, with the method of bilinear interpolation. To complete the cloth texture mapping, we introduce the algorithm based on the triangle grid. Shading processing is also considered when we want to exactly simulate the cloth surface. With this method, we develop the cloth environmental simulation system. Designers can immediately check the production effect when they are designing the fabric. The customers can easily select their favorite cloth when shopping. So the system based on this method is very useful and valuable.关键词：Computer aided design;Textile CAD;Cloth environmental simulation;Mesh generation;bilinear interpolation;texture mapping2636|192|0更新时间：2024-05-07
- 摘要：Combining the basic principle and characteristics of genetic algorithms with the there basic principles of point cluster selection, which are the selection according to the standard of selection, the selection according to the important meanings, the selection according to the range of distributing and density, we design a model of point cluster selection based on genetic algorithms. Considering that we must do our best to preserve the range of distributing , the principle of arranging, the density of distributing of point cluster, the basic principle of the model of point cluster selection is that divide the point cluster M into a sub-clusters: A 1, A 2,...,A a according to the density, then calculate the preserved amount of every sub-cluster according to the amount of every sub-cluster and the total amount to be preserve, last combine convex hull with genetic algorithms to select. The results of the experiments are compared, one is by the methods of the selection based on genetic algorithms and the other is by the methods of the selection of convex hull. From the experimental output, we can get there conclusions: (1) The model of point cluster selection based on genetic algorithms can be better in the selection of dispersive object. (2) The model can preserve the characteristics of the density and the principles of arrangement well. (3) The model can preserve the outlook of the cluster well.2632|198|0更新时间：2024-05-07
- 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