最新刊期
卷 11 ， 期 2 ， 2006
- 摘要：As a new technology of biometrics,gait recognition has its predominance among other biometric technologies while in computer vision research field it is a hard problem.Those two characteristics are analyzed at first.Then, the gait recognition problem is expounded in four aspects which are essentially involved in designing a pattern recognition system: the choice of sensors,preprocessing technique,data representation scheme,and decision making model.The emphases are in the section concerning gait data representation,in which approaches to the gait recognition problem from the point of view of computer vision are broadly classified as being either model-based or holistic.Every class is illustrated in detail,and the merits and defects of each example are pointed out.The information of structure component and the information of the motion dynamics of the body extracted in those approaches are also explained.Thirdly,the performance evaluation standard and available database are discussed.At last,a few weaknesses in the gait recognition research field are proposed,and future areas of research are also put forward.关键词：gait recognition;biometrics;computer vision;spatio-temporal pattern representation;structure component information;body dynamics information5253|612|0更新时间：2024-05-07
- 摘要：In this paper, a new SPIHT(setpartitioning in hierarchical trees) image compression coding algorithm based on the list of pixel states(LPS2SPIHT) is presented, which has five characters as follows: Firstly, a kind of extended spatial orientation tree is defined, whichmakes every point include four adjacent pixels of2×2, and the essential signs of EZW (embedded zerotreewavelet) is applied to the extended spatial orientation tree; Secondly, the three lists of SPIHT(LIS、LIP、LSP) are substituted by a low2cost list of pixel states, saving thememory of program. Thirdly, the coding process is once accomplished bymeans of scanning the list of pixel states, making the sorting pass and refinement pass combine to one; Fourthly, using of one tree exponent gets rid of repeading computation and then results in, the speed up of the significance judgement of trees; Lastly, some predictable bits are omitted from the encoderoutputby rearranging the coding procedure, to reduce the redundancy of the coding and improve the compression efficiency. Practical experiments show that comparingwith SPIHT algorithm which is among the best of existing coding methods, this algorithm not only has better performance, but also is easy to implement, especially, itprovides a feasible referencingmodel for hardware design.关键词：list) of pixel state(LPS);set partitioning in hierarchical trees(SPIHT);LPS-SPIHT;embedded coding;successive-approximation quantization;extended spatial orientation tree3425|515|0更新时间：2024-05-07
- 摘要：Because of the better temporal and frequency properties, the wavelet approximation with constant size filter can get better decomposition effects for the smooth image. However, standard wavelet approximation techniques cannot achieve similar results for images which are not smooth. The structures of wavelet transforms based on the adaptive lifting scheme with perfect reconstruction presented by Heijman et al. , this paper proposed the double adaptive wavelet transform which adopts the adaptive transform in both the update and prediction processes. The transform can perform perfect reconstruction without any overhead cost. At last, the experiment results demonstrate that the algorithm not only allows perfect reconstruction without any overhead cost but also can improve the PSNR of the image.关键词：lifting scheme;adaptive wavelet transform;image compression2506|353|0更新时间：2024-05-07
- 摘要：In this paper,a new ROI based color image retrieval using DWT is proposed.Firstly,the ROI are extracted in DWT domain by using the human visual characteristic and K-mean clustering.Secondly,the local energy of wavelet coefficients of ROI is used for the texture feature,the mean values and standard deviations are used for the color feature,and the weight center coordinate of ROI is used for the location feature.Finally,the similarity between images is computed according to the ROI' above features.Experimental results show that our image retrieval is more accurate in retrieving the user-interested images when there are ROI in the image(especially for the image with simple background).关键词：image retrieval;regions-of-interest;DWT2475|320|0更新时间：2024-05-07
- 摘要：Although color structure descriptor contains the structural information of color, it is crude for processing the color contents of the 8×8 structural elements to distinguish two different pictures with the same color structure histogram. In this paper, a method for image retrieval based on structure quantification histogram is presented. The color structure histogram is one specific case of structure quantification histogram, which is endowed an extreme quantification step as 1. While the colorhistogram is the otherone of structure quantification histogram, which is endowed an extreme quantification step as 0. The performances of the three types of histogram are discussed through average retrieval rate(RARR) and average normalized modified retrieval rank(GANMRR) including 2600 flowerpictureswith differentquantification dimensions(32, 64, 128, 256) inHMMD color space. The results show that the performance ofstructure quantification histogram ofstep size 0.5 is the best. Evaluation criterion used inMPEG27 is introduced in detail and is used to evaluate the experimental results.关键词：MPEG-7;structure quantification histogram;color structure histogram;color histogram;average retrieval rate;average normalized modified retrieval rank3044|366|0更新时间：2024-05-07
- 摘要：In order to make full use of texture information of all spectral bands,and take into account the characteristic of significant correlation between the different bands,this paper presents a method to describe texture of multi-spectral remotely sensed image with spatial fuzzy texture spectrum.Based on multi-scale character of texture,the method implements two fuzzy texture filtering algorithm,the first filtering uses triangle membership function,while the second uses spatial distance in place of plane distance.The spatial membership distribution of the fuzzy image is denoted as spatial fuzzy texture spectrum.With classification by FasART(fuzzy adaptive system.ART-based) networks,experimental results show that the proposed method has higher classification precision,especially to the complex texture area.关键词：uncertainty of the uniform surface;fuzzy texture filtering;spatial fuzzy texture spectrum;FasART networks2843|319|0更新时间：2024-05-07
- 摘要：The effect of varying spatial scale of the Getis statistic on the standardized Z value is studied, it is proposed to employ different spatial scales for evaluation of the spatial autocorrelation in the Getis statistic. Multi-scale Getis is analyzed and validated by using passive remote sensing SSM/I data. Furthermore, a novel multi-layer and multi-scale Getis is studied to evaluate the spatial autocorrelation of the pixels with the medium values. An image of active remote sensing ERS-2 SAR gives an example for analysis and validation of this novel multi-layer and multi-scale Getis.关键词：spatial autocorrelation;Getis statistic;multi-layer and multi-scale2820|349|0更新时间：2024-05-07
- 摘要：Face recognition is a new research topic in today's pattern recognition field.With analyzing the advantages and disadvantages of transient image matching algorithm,a new algorithm for fast human face image matching based on self-adaption grid division and vector coding is presented in this paper.On the premise of the edge extraction and two-value characterization,and face contour extraction,the two images' edges are firstly divided into several grids, and then the segments of the edges and character contour within each grid are described with chain and vector coding.Finally,by computing the similarities and shape alteration between the vector codes,the course of goal matching and recognition is accomplished.The simulative experiment is carried out on the standard ORL face database.Satisfied results have been obtained,which show that the proposed algorithm can not only effectively resolve the problems caused by lower recognition owing to image rotation,minute alteration of edge shape caused by pose transformation and different illumination condition,but also greatly enhance the matching precision and reduce the redundant information and the complexity of computation.关键词：edge two-value characterization;character contour;self-adaption grid division;vector coding;similarities;shape alteration of edge2434|344|0更新时间：2024-05-07
- 摘要：Rectification of stereo vision is also called rectification of Epipolar Lines. To search the corresponding points rapidly and accurately, the rectification is used to make the epipolar lines of stereo images be parallel to the horizontal direction and to remove the parallax in vertical direction. A robust algorithm is presented to rectify the stereo images based on the traditional projective rectification algorithm. In this method, the projective matrix is calculated by corresponding points and then is optimized by Levenberg-Marquardt and GA algorithm. And meanwhile the RANSAC algorithm is used to avoid the error caused by the noise. Furthermore, a robust optimization function is proposed for the situation with slight movement of two cameras. Experiments with real images show that the method is an effective rectification method.关键词：epipolar rectification;epipolar geometry;fundamental matrix;random sampling consensus(RANSAC)3567|328|0更新时间：2024-05-07
- 摘要：Image filtering and enhancement play a very important role in digital image processing.The anisotropic diffusion is a selective smoothing technique that effectively employs intra-region smoothing without limitation and inhibits inter-region smoothing.In order to remove noise effectively and preserve edges and key details,a more effective and adaptive diffusion algorithm is to be proposed in this paper.We build a new diffusion coefficient in partial derivative equation(PDE) with the advantages of two existing diffusion coefficients and incorporate the proposed scheme with the nonlinear time-dependent cooling technique for gradient threshold.The proposed algorithm is based upon a selective and improved diffusion coefficient and performs adaptively towards different gradients.The improved algorithm not only preserves image edges,but also smoothes small scale features and avoids distortion of image edges and details. It has been shown from the experiments that the improved scheme has superiority capability over the Perona-Malik scheme and it is a robust anisotropic diffusion.关键词：anisotropic diffusion;diffusion coefficient;image smoothing;image noise reduction and enhancement4635|408|0更新时间：2024-05-07
- 摘要：With virtual cutting operation for volume data of medical images or reconstructed geometry models from medical images, inside tissues can be observed easily, which can aid doctors to diagnose. The methods can also been used in surgery planning of radiotherapy. In this paper, new methods of plane cutting, stereo clipping and arbitrary cutting operations of 3D surface model reconstructed from medical images are presented. In plane cutting and stereo clipping methods, the 3D model is cut by plane or polyhedron. Lists of edge and vertex in every cut plane are established. From these lists the boundary contours are created and their embracing relationship is built. The region enclosed by the contours is triangulated using Delaunay triangulation algorithm. The procedure of arbitrary cutting operation is to create cutting path interactively at first, then build cutting boundary, and finally cut the model with the boundary. The cut model still maintains its correct topology structure. Experiments show the efficiency and stability of the methods. Good results are achieved by using the methods to cut models reconstructed from medical images.关键词：medical image;surface model;virtual cutting;stereo clipping3332|337|0更新时间：2024-05-07
- 摘要：Image restoration is an ill posed problem and must be regularized.Usually,the difficulty of regularization lies in avoiding the smoothing of edges while suppressing the noise.In order to preserve the edges of restored image effectively,a general wavelet based edge-preserving regularization scheme and the corresponding solving scheme are proposed.To get rid of the Gibbs effects of restoration algorithms in wavelet domain,the natural(image's) horizontal(vertical) continuity in horizontal(vertical) sub-band together with additioned punishment are employed.Experiments are presented which verified the theoretical results.关键词：edge-preserving regularization;Gibbs effects;Markov random field;potential function3225|386|0更新时间：2024-05-07
- 摘要：In PET image reconstruction, regularization methods are usually considered to suppress noise effects in reconstructed images. In this paper, the Mumford-Shah functional model was firstly investigated and the new variational framework for PET reconstruction was then proposed. The resultant linear diffusion equation based reconstruction method was derived accordingly. Based on this linear method, the new nonlinear diffusion method was also presented to provide more advantages for the image edge preservation. Simulation results suggest that both linear and nonlinear diffusion equations based methods have nice performance in final reconstructions. Also, results have verified that the nonlinear diffusion equation based reconstruction method is superior to the linear one in both noise smooth and edge preservation.关键词：positron emission tomography(PET);reconstruction;Mumford-Shah functional;diffusion equation;nonlinear;regularization2828|324|0更新时间：2024-05-07
- 摘要：On the basis of digital image’s information sharing, a new blending approach based on RB curve is presented. The approachmakes use of an degree rational Bézier curve or akdegree2[n/n]RB curve or a[n/n]RB curve with weights to hide information of one image intonimages. These blending schemes have more blending parameters to act as the private keys, which can enhance the security and robustness of the hiding image. Furthermore, the schemes can be combined with the scrambling technology, which is usually used as pre2process and post process of digital image information hiding. Experimental results show that these schemes can be realized expediently and have some degree of robustness. These blending schemes can also be used in the field of digital watermarking and other digital application.关键词：digital image information hiding;blending image;image sharing;digital watermark2893|361|0更新时间：2024-05-07
- 摘要：This paper presents a digital image encryption algorithm based on four-value chaotic array. It uses chaotic maps to generate a four-value chaotic array whose size is as big as the image. In encryption processing, we deal with the different value in the chaotic array separately, and use different keys to encrypt pixels which correspond to different chaotic array value respectively and then put the result to the appropriate position in the chaotic array according to the given random integers. Decryption is the converse of encryption. The analysis and the results of experiments show that the scheme has low time complexity, low space complexity and good security.关键词：image encryption;four-value chaotic array;time complexity;space complexity;security3871|580|0更新时间：2024-05-07
- 摘要：The definition of analytic signal and the computation algorithm for instantaneous frequencies and instantaneous amplitudes have important significance in signal processing domain.The quaternionic analytic signal is an appropriate extension of 1D analytic signal to 2D case.Whereas,the existing phase definition of quaternionic analytic signal isn't fit for extracting instantaneous frequencies of images.According to the properties of quaternionic analytic signal,a novel phase definition of quaternionic analytic signal is proposed in this paper and two decomposition tools are utilized to test the effectiveness of the definition.The experimental results of synthetic and natural textures indicate that the new definition is more coincident with image intrinsic characters and advantageous for image analytic feature extraction and classification.关键词：analytic signal;quaternion;empirical mode decomposition(EMD);Gabor filter;texture analysis3227|380|0更新时间：2024-05-07
- 摘要：This article expands the method constructing the generalized Mandelbort sets for positive integer index number with the composition of two simple complex mapping which were put forward by Shirriff. Based on a class simple complex mapping system expanded by the author, a series of generalized Mandelbort sets for real index number has been constructed. With the experimental mathematics method of combining the theory of analytic function of one complex variable with computer aided drawing, the fractal features and evolutions of the generalized Mandelbort sets are studied. The results show that the generalized Mandelbort sets for integer index number have symmetry and fractal feature; while the ones for decimal index number have discontinuity and collapse, and their evolution depends on the choice of the principal range of the phase angle. And the author sets forth the physical meaning of the generalized Mandelbort sets.关键词：a class simple complex mapping system;the generalized Mandelbort set;fractal;evolution;physical meaning3762|451|0更新时间：2024-05-07
- 摘要：Bézier curve and surfaces are one kind of the most commonly used parametric curves in CAGD and computer graphics. Developingmore convenient techniques for designing and modifying Bézier curve and surfaces are an important problem. This paper investigates the optimal shapemodification ofBézier surfaces by geometric constraints. A constrained optimizationmethod based on changing the control points of the surfaces is presented in this paper. Ahich modifys control points of the originalBézier surfaces to satisfy the given constraints andmodify the shape of the surfaces optimally. Practical examples are also given. The results indicate thatLagrange multipliermethod is an effective way to solve the problem of shape ofBézier surfaces.关键词：Bézier surface;constrained optimization;shapemodification;control points3041|374|0更新时间：2024-05-07
- 摘要：Through heightening the degree of polynomial function, a class of polynomial function of(n+1)degree that containing an adjustable constantparameterλis presented in this paper. They are an extension ofndegree Bernstein basis functions. Properties of this new basis are analyzed, which have symmetry, linear independence, weighting property and nonnegative property when the parameter λis between -2 and 1, based onwhich a(n+1)degree polynomial curvewith a shape parameter λis defined. The curve, to be called λ Bézier curve not only inherits the most properties of n degree Bézier curve, such as endpoints’properties, symmetry, convex hull property, geometric invariability, affine invariance, convex preserving property, variation diminishing property and so on, but also can be adjusted in shape by changing the value of λ without changement of control points. Whenλ=0, the curve degenerates ton2degree BézierCurve. Using tensor product approach, a surfacewith parameterλis constructed, whose properties are similar to the curve’s. At last, examples illustrate themethod of constructing curve is very useful for curve/surface design.关键词：Bezier curve;shape parameter;curve design3624|355|0更新时间：2024-05-07
- 摘要：In this paper,the basic system structure of the virtual NC(numerical control) machine tool is put forward and the basic conception of OOD(object oriented design) is introduced to the virtual system design of NC machine tool.Based on their respective characters,all parts of NC machine tool are classified into three classes: CStaticClass,CMoveClass and CDynamicClass.The STL files which containing the appearance feature of the undeformed objects are imported from solid modeling software(Pro/E) and feature modeling approach are adopted to describe the deformable objects,then combine with the respective animate functions to compose the classes,and finally accomplish the 3D dynamic simulation.The basic theory of quaternion is introduced and applied to the design of the system and the rotation of virtual NC machine tool.The methods provide a basic frame for studying on virtual NC system.The experimentations prove that the methods are feasible and have good practical value.关键词：virtual reality;CNC(computerized numerical control) machine tool;object oriented;graphics handle2913|347|0更新时间：2024-05-07
- 摘要：The object-based selective attention has been given increasingly importance in psychology domain in recent years,whereas most of the existing models of selective attention in computer vision field are either feature-based or space-based.In this paper,an object-based computational model of selective attention is proposed where "perceptual object" is postulated as the basic attention unit.The proposed attention model consists of following two steps:(1) how to select the first focus of attention in a given image;(2) how to shift the attention within the whole image.Under this model,the contrast,defined as the absolute gray difference between the "perceptual object" and its neighborhood,is used as the measure of the object's salience,by which the attention is determined and shifted.The advantages of this model lie in: Firstly,the model is entirely based on perceptual objects,its results can be easily applied to object detection,image segmentation,and scene analysis.Secondly,the model is of multi-scale,flexible enough to be easily adjusted according to the specific applications.Extensive experiments on real images show that the proposed attention model works reasonably.关键词：selective attention;regions-of-interest;perceptual object;shift of attention3919|530|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