最新刊期
卷 10 ， 期 8 ， 2005
- 摘要：The paper studies concept and approach of decomposable Markov network(DMN).Unlike previous studies on DMN,our method directly employs the construct of DMN as the evidence of inference or for problem solving,and enlarges application range of DMN's concept and approach.It analyses DMN's role and importance in spatial data mining,and deeply investigates several score metrics used in constructing DMN between multi-bands remote sensing images,so as to interpret the result of fusion in them,and realizes optimal fusion of bands.Another example of the factual video images with traffic rule violation is also investigated by novel concept of DMN's graininess.Several graininess(nodes) are used to construct DMN between video images for detecting abnormality in them.Hence, the vehicle of traffic peccancy is located and traced.The result of video image simulation shown that our method is feasible and effective.The researches indicate that the DMN may reveal abstract adjacent relations existed in spatial data;and the network itself has capabilities of showing knowledge.关键词：decomposable markov network;multi-bands remote;target detection;probability distance;spatial data mining3120|214|0更新时间：2024-05-08
- 摘要：In this paper,based on the analysis of interlaced image,quadtree and Baptista algorithm,the authors presented an interpolation algorithm for interactive image generation: Firstly,sample evenly on image,then evaluate the pixel's value according to the distribution geometry characteristics of pixel's neighborhood.After the comparison among the four algorithms,the experimental results show that interpolation algorithm can generate high quality images with less data.关键词：interlaced) image algorithm;quadtree algorithm;Baptista algorithm;interpolation algorithm;Image neighborhood2493|143|0更新时间：2024-05-08
- 摘要：The anisotropic diffusion is a selective smoothing technique that effectively employs intra-region smoothing without limitation while inhibits inter-region smoothing.With this technique,noise is smoothed out and the edge-preserving smoothing can be effectively implemented.Two problems in anisotropic diffusion schemes have still not been solved.One is existence and uniqueness of the solution of the partial derivative diffusion equation;the other is reasonable estimate of gradient threshold in diffusion coefficients.The algorithm proposed in this paper is developed on the basis of minimal description length criterion and incorporated with one alternative method for setting gradient threshold based on the simple calculation of the p-norm.The improved model not only detects effectively the detailed edges in images,but also preserves the stability of anisotropic diffusion.Experimental results show that this method has superiority over the previous anisotropic diffusion models and is an ideal edge-preserving filtering method.关键词：anisotropic diffusion;minimal description length criterion;p-norm gradient threshold estimate;diffusion coefficient;image smoothing3722|175|0更新时间：2024-05-08
- 摘要：Firstly,this paper gives the property of wavelet transform of two dimensional noise,analyzes the relationship of wavelet transform modulus maxima to different decomposed class j and Lipschitz exponent,and points out how to determine and protect image edges.Then it explains the orthogonal wavelet transform of denoising based on soft and hard threshold,and puts forward a denoising method based on the wavelet modulus maxima and Neyman-Pearson principle. The method finds the optimal trade off between image denoising and protecting image edges.Based on the assumption that the observed image is the sum of the expected image and irregular corruptive noise,the qualitative and quantitative performance of our image denoising method is compared with others.Simulation results show the proposed method can efficiently denoise,such as increasing Signal-to-Noise Ratio(SNR),lowing Mean Square Error(MSE) and Relative Entropy(RE), while preserving the details of the original image.关键词：wavelet threshold denoising;Neyman-Pearson principle;wavelet modulus maxima;image edge3471|194|0更新时间：2024-05-08
- 摘要：A new finite difference method is proposed to solve the image irradiance equation of shape from shading.Firstly,the non-linear term in the reflectance map is approximated by a Taylor-expansion.Then the resulting linear partial difference equation is solved numerically by a semi-implicit cross difference scheme,combining with initial and boundary conditions.We prove its stability and convergence in detail.Surface height is obtained iteratively by successive over-relaxation method,and the method of selecting optimization factor is also given in the paper.Experimental results on synthetic images demonstrate the new algorithm suits for any light direction and the reconstruction results are better than the traditional ones.关键词：shape from shading;difference scheme;truncation error;relaxation algorithm2466|196|0更新时间：2024-05-08
- 摘要：This paper does further research for the lifted Donoho wavelets.Through the computing of regularity and the wavelets sampling approximation theory,we explain it has better approximation capability when N=N~=4(denoted as SWE13/7),and it precedes D9/7 wavelets in sampling approximation.The lifting coefficients of SWE13/7 are all simple fractions,so wavelets transform only with integer shift and adder operations can be obtained in computing process,the computing complexity is efficiently lower than D9/7,and the result is equivalent to those using floating number in computing process.The line-based lifting scheme can efficiently decrease the memory usage,and increase the speed at parallel computing case.This paper has realized the line-based lifting scheme of SWE13/7,it has lower complexity than D9/7 in evidence.Numerical experiments indicate that the image compression performance of SWE13/7 precedes that of D9/7 when using EZW coding,and its computing complexity is much less.关键词：lifting scheme;regularity;integer wavelets transform2495|214|0更新时间：2024-05-08
- 摘要：Linear discriminant analysis is one of the classical and popular methods used for feature extraction.In recent years many researchers have been absorbed in the problem of how to extract the optimal Fisher discriminant feature in small sample size case.By making use of the principle of projection transformation and isomorphic transformation,in this paper,we have solved the problem of how to gain the optimal discriminant vectors in the singular case.In fact these optimal discriminant vectors can be derived from a low dimension transformed subspace.Fulfilling the need of application,a new model for feature extraction has been put forward and a corresponding algorithm,called PPCA+LDA in this paper,has been established.The experiments on three kinds of resolution grayscale image for ORL face image database have been performed.The results of experiments show that the set of the discriminant vectors extracted by the proposed algorithm has powerful ability of feature extraction and the recognition results are very robust by the general minimum distance classifier.关键词：small sample size problem;PCA(principal component analysis);LDA(linear discriminant analysis);compressed transformation;face recognition2902|199|0更新时间：2024-05-08
- 摘要：In order to get new methods on the digitization of the traditional Chinese medicine(TCM) tongue diagnosis and classification and retrieval of tongue pictures,a new image retrieval method based on multi-feature(color and texture) is presented in this paper.The authors used a 20-dimension eigenvector including both color and texture as the whole(image's) features to retrieval the resembled tongue images.The experimental results are satisfactying,whicn shows that this method can be used in retrieval of tongue pictures successfully.关键词：color;texture;tongue image;traditional Chinese medicine(TCM) tongue diagnosis;image processing;content-based image retrieval(CBIR)3295|202|0更新时间：2024-05-08
- 摘要：The edge detection of intravascular ultrasound(IVUS) image has great significance in the diagnosis and treatment of the coronary artery disease. In this paper,a new method used for automatically detecting the intima and adventitia edge of IVUS image is presented.The method bases on active contour model,the contrast and Rayleigh distribution characteristics of IVUS image,effectively uses dynamic programming and heuristic graph searching,and automatically detects the intima and adventitia edge of IVUS image with the different cost functions.Experiments show that,compared with the former method of edge detection,our method is algorithmically simple,statistically accurate,reproducible and robust in sequential IVUS frames,and the algorithm is a kind of global optimal one.关键词：intravascular ultrasound;active contour model;dynamic programming;contrast;Rayleigh distribution3047|301|0更新时间：2024-05-08
- 摘要：The paper presents a novel feature matching method based on Queen-bee evolution.The features are the corners extracted by using the improved Harris corner detector.Aiming at their corresponding relationship,a new chromosomal encoding is proposed and for it effective crossover and mutation operators are developed.To evaluate the encoding chromosomes,five matching constraints are considered.By taking advantage of the queen-bee evolution scheme,the linear ranking selection,and the elitist replacement,the global optimal solution can be achieved quickly.The experimental results show that the proposed approach perform corner matching accurately and quickly.关键词：computer vision;feature matching;genetic algorithms;queen-bee evolution;epipolar geometry2805|229|0更新时间：2024-05-08
- 摘要：In recent years,the fields of active-contour based image segmentation have seen the emergence of two competing approaches.The first is based on parameter models and the other is based on geography models.Snake model can segment objects quickly,but can not deal with topological changes and sensitive to initialization.Level Set model can deal with topological changes but it's time efficiency is low.After comparing these two approaches,this paper presents a new active contour model:S-L model,which combines the virtues of Snake model and Level Set model.The new model uses the energy equation of Snake model to evolve the curve and uses a symbol table,which is based on the soul of Level Set model,to change the topology of the curve.To reduce the effect of the noise,the new model constructs a new outer force on the basis of the region information.With the new outer force,the initial curve can be made in a large space. With the region information,the new model can find the edges powerfully,even if in case of complex topology,avoid local minima from Snake model.The experiments to segment cardiac magnetic resonance images show that comparing with Level Set model the new model can get the similar results which has mach higher speed.2603|192|0更新时间：2024-05-08
- 摘要：On the basis of the gray-level distribution in the relative half-neighborhood of image pixels,refering to the method called "Gap statistic" proposed by Hastie and Tibshirani,a concept called inverse distribution function is brought forward,and a multi-scale edge detection model based on Gap of random variable was established.By analyzing consistency between a grayscale distribution Gap and random variable Gap,the edge detection algorithm for Gap statistic model is optimized.The paper analyzes the correlation between the Gap statistic model and two operators(Prewitt operator and Sobel operator),discusses the anti-noise and multi-scale properties of the edge detection model,and an investigation is made to analyze the difference of edge detection at different scales.Finally,experimental examples verify the capacity of the model.关键词：image processing;edge detection;Gap statistic;model;multi-scale2708|172|0更新时间：2024-05-08
- 摘要：Recognizing subjective contour figures is an intrinsic capability of human vision and an important visual psychological phenomenon.This paper presents a computation approach for subjective contour which utilized the cooperation-competition mechanism of biological vision system.The goal is to simplify the neural-computing of biological vision process modeled in the Boundary Contour System of FACADE(Form and Color and Depth) theory,and to extract subjective contour in a new way.The procedure consists of DOG filtering,cooperative filtering,post-processing of cooperative cues and regrouping of endpoints,corresponding to the oriented edge filtering by simple cells,cooperative filtering by bipole cells and end-stop effect formulated by hypercomplex cells in biological vision system respectivly.As shown by experimental results,this approach is efficient to process both kinds of classic subjective contour figures,the collinear completion information and endpoints.关键词：subjective contour;FACADE;boundary contour system3019|135|0更新时间：2024-05-08
- 摘要：This paper introduces the theory and training algorithm of the support vector machine which is applied in nonlinear classification and recognition by the way of bringing in the concept such as structural risk minimization principle and optimal hyperplane,then a set of nonlinear binary samples are successfully classified by using different kernel functions,followed by discussion to the results.After that current multi-class classification algorithms and application areas are reviewed.Finally future developments are prospected.关键词：support vector machine;optimal hyperplane;nonlinear classification;quadratic programming2720|175|0更新时间：2024-05-08
- 摘要：B To meet the demand of exact localization of video retrieval in Media Asset Management(MAM) system,this paper presents a novel method for logger and retrieval,which takes advantage of both the retrieval based on content and the method based on hand-logger.First of all,the high dimension features are reduced by PCA,which are used to detect the video shots.The semantic of the repeat news shots is auto annotated.The semantic of shots is updated automatically finally.It is verified by the experimental results that the proposed approach is feasible and efficient.关键词：video retrieval;semantic;PCA;shot detection2704|202|0更新时间：2024-05-08
- 摘要：In this paper,an object-of-interest based video transmission approach is presented for vision-based traffic surveillance systems using the background subtraction with one reference image.This background is pre-stored at both the image-capturing side and the monitoring center;at the acquisition side,using the badcgroand sabtraction with only one reference image the target objects are extracted,deriving a minimum rectangular grid-based region and a mask matrix for each of the extracted object.The monitoring center,after receiving the necessary codes,can reconstruct the video frames by fusing the objects or their wavelet-transformed versions into the preset background.The advantages are:(1) the networking resources are greatly saved;(2) some high-level processing functionality like collision avoidance,fault alarming can be improved;(3) operators' attention can focus on important events;(4) new services can be added or developed.Experimental results have proved its effectiveness.关键词：traffic surveillance systems;background subtraction;video transmission;wavelet transform;image fusion2625|205|0更新时间：2024-05-08
- 摘要：With the development of information science and the computer technology,the application of video in commerce on the Internet,such as video on demand or video on line has become more and more common.At the same time,the demand for security becomes more and more urgent.According to the characteristics of the latest video compression standard H.264,this paper proposes several DCT-based encryption algorithms,conducts simulation on two standard video sequences,and analyzes and compares the quality of those proposed algorithms.The experimental results show that the proposed algorithms have no impact on the compression efficiency and have gained good encryption results with low complexity.The simulation results are all tested on JM(joint model,version 7.3) reference software,which is especially programmed for testing the coding standard of H.264.关键词：H.264;encryption;JM(joint model);transform2517|158|0更新时间：2024-05-08
- 摘要：HS method is one of the most classical methods for optical flow computation.In practice the variation of gray level between two points is nonlinear,while in the original HS method,it is assumed to be linear,which leads to the inaccuracy of the convergent point of LSM iterations.In the 2nd section,we discuss the reason why the classic method isn't accurate enough in detail.And in the 3rd section we present a new algorithm to improve accuracy of gray gradient by using the offset,which is generated in the previous cycle,to produce the new gradient.The following experiments,including synthesized image sequences and images sequences extracted from real video,compare the original algorithms with the improved one.The objects in the synthesized sequences involve some stereotype movements such as translation,rotation,and zooming.The results show that the veracity of optical flow field has been enhanced,and the convergence speed is accelerated to a great extend.关键词：optical flow computation;HS method;gray-gradient3244|216|0更新时间：2024-05-08
- 摘要：It has been known that Iterated Function System is a personification methods to produce fractal image,and to form IFS's fractal image by parameters adjustment on the basic of affined map is a usual method.The key is to establish proper parameter models.At present,we have lots of parameter algorithms,but more work need to be done to get over the random of parameters selection and unpredictability of fractals shape.This paper establishes models with exterior parameters of geometry intuitionalism.We adjust indirectly interior parameters of original IFS by introducing exterior parameters which it produces a new IFS,and enables us to control the fractal evolvement under rules certain of designs.Experimental results show that fractal images with different configurations can be obtained intuitively and effectively by this algorithm.关键词：iterated funtion systems;affine transformation;fractal image;immovability point3021|273|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