最新刊期
卷 12 ， 期 1 ， 2007
- 摘要：Image inpainting is an important research topic in the area of image restoration.Its objective is to restore the lost information according to around image information,which can be used to restore old photo,remove text and conceal errors in videos.Based on many literatures of digital image inpainting,this paper attempts to make an overview of digital image inpainting.First,it describes image inpainting from mathematics background.Then two kinds of important image inpainting schemes are introduced in this paper: one is image inpainting based on the geometric image models;the other is image completion based on texture synthesis.The former is suitable to inpaint the small scale scratches in images and the latter is very good at completing the large objects.Then this paper demonstrates the applications of the two kinds of methods.At the end,the future trend of digital image inpainting is pointed out in personal opinion.关键词：image restoration;image inpainting;image completion;variational method;partial differential equation(PDE);total variation;texture synthesis20517|1744|0更新时间：2024-05-07
- 摘要：Image segmentation and boundary extraction are very important in the fields of image understanding,image analysis,pattern recognition and computer vision et al,while active contour model is one of the most important tools in the areas of image segmentation and boundary extraction which mainly includes parametric active contour model and geometric active contour model.Geometric active contour model has many advantages over parametric active contour model,such as computational simplicity and the ability to change curve topology during deformation,et al.Therefore,significant advances have been made in theories and applications of geometric active contour model recently.In order to show the general idea of this technique,a novel classified mode is developed to describe the parametric active contour model,geometric active contour model and the relationship between them at first in this paper.Moreover,by analyzing several classical active contour models,this paper summarizes the research,development and applications of active contour model.Finally,this paper points out future research orientations on the theories and applications research of active contour model.关键词：image segmentation;active contour model;variational calculus;level set methods;additive operator splitting(AOS) scheme18770|1266|0更新时间：2024-05-07
-
Highly Scalable Video Codec With Wavelet Domain Motion Compensation Based on Regular Triangular Mesh
摘要：Video scalability is referred to that the encoded bitstream can be decoded partly and the reconstruction video quality is proportional to the amount of received information,and the spatial and temporal resolutions can be changed according to the decoder property.This paper presents a highly scalable video codec with wavelet domain motion compensation based on regular triangular mesh,which decomposes the reconstructed reference frame by RDWT,carries out motion compensation in wavelet domain by mesh-based motion model,so that the efficiency of motion compensation is improved.In order to keep not only PSNR scalability but also spatial and temporal resolution scalability,an amended SPIHT method with subbands-scan order is proposed.The experimental results prove that the codec can produce encoded stream with good PSNR and highly scalable functions.关键词：scalabilities;redundant discrete wavelet transform(RDWT);mesh-based motion compensation;set partitioning in hierarchical trees(SPIHT)3895|404|0更新时间：2024-05-07 - 摘要：In video coding system,motion estimation at half-pixel accuracy can obviously enhanced coding efficiency compared to the motion estimation at integer-pixel accuracy only.However it requires more computation at the same time.In order to reduce the computation and while enhance the speed of motion estimation at half-pixel,we proposed a novel half-pixel motion estimation algorithm named half-pixel motion estimation based on linear prediction(BLPHME) for video coding in this paper.The key point of the algorithm is setting up a linear model by analyzing the relativity between the results of integer-pixel search and half-pixel search.Then we can modify the threshold dynamically for each frame.By doing so,it can skip over the blocks that can't be benefited from the half-pixel search based on the block size.Experimental results show that significant reduction in computation of motion estimation is achieved together with the increment in bit rate by using the proposed method,but without obvious increment in bit rate and visible loss in video fidelity and increment in bit rate.Moreover,the proposed algorithm can be combined with any of the fast motion estimation algorithm at integer-pixel or half-pixel accuracy to further reduce the complexity.关键词：motion estimation;block-matching;linear prediction;4-MV;half-pixel search3346|278|0更新时间：2024-05-07
- 摘要：The complexity of video encoders in the international standards of H.263,MPEG- 4,H.264,depends mainly on the block motion estimation algorithm.In order to make motion estimation algorithm faster and more accurate,a fast motion estimation algorithm based on probability matrix is proposed in the paper.In the algorithm,every possible motion vector's probability of current macro block is estimated by the previous macro blocks' motion vectors,and constitutes the probability matrix whose size is as the same as the search window's.The algorithm restricts the number of pixels to search by probability.From the results of experiment,comparing the proposed algorithm and diamond search algorithm,it shows that the veracity is improved a little,and the speed grown faster at the same time.关键词：video compression;motion estimation;block matching algorithm;probability matrix3709|342|0更新时间：2024-05-07
- 摘要：In low bit-rate packet-based video communications,loss of packet usually results in whole-frame losses.The algorithm dealing with the problem of whole frame loss is proposed in this paper.Rational interpolation is firstly applied to generate accurate dense motion field based on decoded block-based motion field.Then the previous image pixels are projected onto missing frame along the motion vector direction.The identification of overlapped and occluded areas is also used to improve error concealment performance.Experimental results show that the proposed algorithm outperforms other techniques in peak signal-to-noise ratio(PSNR) and provides good visual quality.关键词：video communication;error concealment;rational interpolation4097|286|0更新时间：2024-05-07
- 摘要：To improve the performances of median filter and enhance the adaptability for the diverse density pepper and salt noises an adaptive median filter,based on the double windows and extremum-compressing filtering(DWECAMF),is proposed for images corrupted with pepper and salt noises.The filter is improved from the standard median filter on the three standpoints: accurateness of noise detection,validity of noise removal and practicability of filtering speed.The method employs some strategies,such as using the large window for noise detection and the small window for noise removal,compressing the maximum and the minimum within the window for noise removal,and using the adaptive filter for impulsive noise removal-to improve the performances of the filter,and the floating filter strategy to accelerate the filtering speed.The experimental results show that the filters's three performances-mentioned above of the filter are all improved greatly.Meanwhile,the filtering performances of the filter are excellent for different noise densities,ranging from low densities,to high densities.关键词：impulsive noise;median filter;double windows filtering;noise detection;extremum-compressing;floating filtering3651|294|0更新时间：2024-05-07
- 摘要：Wavelet image denoising is an important method of image denoising.Recently,many different schemes of wavelet image denoising were proposed.Among these,NeighShrink suggested by G.Y.Chen et.al.has been proved very efficient.NeighShrink differs from traditional threshold methods in that it incorporates neighboring coefficients when shrinking wavelet coefficients,and thus avoids killing too much image details.In order to improve the visual quality of the denoised image,a new wavelet image denoising method,namely enhanced NeighShrink(ENS),is proposed in this paper based on the NeighShrink scheme.By changing the way to calculate the shrinkage factor for the wavelet coefficient,ENS achieves statistically better results than original NeighShrink method in denoising. Moreover,by introducing an extra parameter P in our wavelet scale dependent shrinkage factor calculation scheme,ENS can be used to enhance image details while denoising the image.This feature can be used to improve the visual quality of the image,since the original NeighShrink method,like many other schemes,shrinks all wavelet coefficients,which will incur the loss of the image details to some extent.Experimental results show that ENS can achieve better results in both denoising and enhancing of image details than the traditional soft threshold and NeighShrink methods.关键词：wavelet transform;image denoising;neighboring window;image enhancement3696|292|0更新时间：2024-05-07
- 摘要：Image scrambling is an available means for image encryption.Since chaos system is extremely sensitive for initial value,ergodic and aperiod,it was widely applied for image encryption.Because there are certain techniques to decode the encryption method based merely on chaos system,a novel scrambling method based on image blocking and chaos system is presented in this paper.The rough idea of this method is that an image is scrambled by a blocking method,and then the result is further scrambled by a modulating way provided by chaos system.The experimental result proves that the new method is effective.关键词：image scrambling;chaos system;image blocking6260|333|0更新时间：2024-05-07
- 摘要：A bi-watermarking algorithm basing on wavelet packet transform(WPT) and discrete cosine transform(DCT) is proposed in this paper.With the human visual system(HVS) in the domain of WPT and DCT,the algorithm adopts different methods to insert one watermark in the low frequency and high frequency of the host image wavelet packet decomposition(WPD) respectively.Experimental results show that the algorithm possesses the following advantages.Embedding the watermarks in the low and high frequency of the host image WPD improves the watermark resistance to different attacks.Considering the DCT energy compression ability and the ability of ridding correlation,the use of DCT in the lowest frequency band of the WPD enhances the watermark robustness and invisibility.Two watermarks embedded in the host image not only increase the watermark capacity,but also improve the serviceability of the algorithm.关键词：digital watermark;wavelet packet transform;discrete cosine transform;human visual system4413|343|0更新时间：2024-05-07
- 摘要：Stereo video is an important future trend in video technologies and intellectual property rights protection for stereo video should be resolved in advance to avoid piracy.In this paper,based on our previous research work on both digital watermarking and stereo video,a novel adaptive digital watermarking algorithm for stereo video is proposed after analyzing the characteristics of stereo video.The watermark is embedded into disparity vectors of enhancement-layer stream,without affecting the structure of base-layer stream.The proposed algorithm is an adaptive one,which keep the quality degradation of the overall reconstructed video images being the smallest.The whole watermarking process is mainly performed in compressed domain without the need of the fully decoding of compressed stream.And watermark detection can be blindly performed without the need of the original video sequence.As a whole,the proposed algorithm is fast,simple.Experimental results with several stereo sequences show that the embedding of watermark has little affection on image quality and the degradation of PSNR with the adaptive algorithm is smaller than that with non-adaptive algorithm.关键词：stereo video watermarking;disparity vector;adaptive algorithm;compressed domain;blind detection2859|229|0更新时间：2024-05-07
- 摘要：Algebraic reconstruction method(ART) is an effective approach to reconstruction of incomplete projection data and a most powerful key technology especially in nondestructive detection of parts of aircraft and spacecraft which are larger than the size of CT detector.However,former procedures have huge amount of computation and are extremely time-consuming.In order to improve these disadvantages,this paper brings forward a rapid parellel algebraic reconstruction procedure based on SIMD technology of Intel central process units.Having maturely comprehended the feature of ART formula,this novel procedure designs a set of data structures convenient for parallel computation and a procedure pipeline to load a number of packed data in one time and to complete reconstruction computation in SIMD method by MMX,SSE and SSE2 instructions.Proved by the simulating experiment,this method promotes the speed about 4 times with the same precision of ordinary procedures,and solves the bottle-neck problem of traditional ART procedures,which possesses important engineering applicational value of nondestructive detection for large parts of aircrafts and spacecrafts.关键词：computed tomography(CT);algebraic reconstruction technique(ART);SIMD parellel computation;SSE and SSE2 instructions3844|310|0更新时间：2024-05-07
- 摘要：In this paper,a new two-dimensional(2D) thresholding method is proposed in order to improve the efficiency of image segmentation.Based on the characteristic of two-dimensional histogram of image and the requirement of segmentation,one of the two dimensions is the pixel's gray value and the other is its neighboring average gray value.The proposed method utilizes the important edge histogram of the image to segment it,while based on the traditional two-dimensional(2D) Otsu thresholding algorithm.According to the foreknowledge about the relationship of the edge pixel histogram and the threshold vector(s,t),the proposed method derives the optimal threshold vector(s_best,t_best),by looking for the valley value existing between two peaks in the edge histogram.Emulational experiments show that,compared with the traditional two-dimensional(2D) Otsu algorithm,the presented method reduces computation complexity greatly and reduces the running time of the algorithms,while retains the advantage of the traditional two-dimensional(2D) OTSU algorithm,such as nonparametric,unsupervised,high performing quality and so on.It can be seen from the emulational result of cellular images that both the improvement and real-time quality of the proposed method are valid.关键词：two-dimensional thresholding;image segmentation;two-dimensional gray histogram4988|457|0更新时间：2024-05-07
- 摘要：Support vector machine,a research hotspot of the pattern recognition in recent years,performs successfully in solving the nonlinear and high dimensional problems.However,training a support vector machine is equivalent to solving a linearly constrained quadratic programming problem in a number of variables equal to the number of data points.This optimization problem is known to be challenging when existing large number of training data points.Also,it is well known that the number of support vector plays an important role in the classification speed of SVM.So the method of pre-analysis efficient support vectors are used to train classifier becomes a novel task in SVM fields.In this paper,on the basis of a deep investigation into the geometry principle of support vectors and its distribution,we firstly pick out some neighbor vectors by nearest interclass distance analysis,and then select the margin vector by computing its intermixed factor of the neighbor vectors.So this method speeds up the SVM training and classifying synchronously by reducing the number of training samples and trimming the intermixed samples,while the ability of SVM remains unchanged.关键词：support vector machine;neighbor vector;margin vector3240|259|0更新时间：2024-05-07
- 摘要：In dealing with complicated problems,the characters of the object can be obtained effectively when the disturbing and nonessential attribute can be wiped off by changing the granular space where the problem located,which make it easier to analyze and solve the problems.In this paper,the analysis of clustering is discussed according to granularity computing.It is assumed that the clustering problems are analyzed under the same granularity(the finest granular space of the problem).The essential of introducing the different comparability functions of clustering is to get a series of equivalence species of different granular space.In practice,problems can be transformed into required granular space,by selecting different comparability functions according to the problem.The transformation form multicolor three-dimension space to monochrome one-dimension can be realized by proposing The License Plate Binary Algorithm based on Information Granularity.Experiments show that the results of this algorithm are more suitable to actual image,have broad generality,and are in favor of recognition following.It is especially predominant in inclined plates or asymmetrical illumination plates.关键词：information granularity;granular space;clustering;license plate binary3086|236|0更新时间：2024-05-07
- 摘要：In the texture image,the imperfect part is denoted as defect.Texture is usually depicted by a gray-level distribution along with a certain spatial interaction.Thus,gray-level co-occurrence matrix(GLCM) is an appropriate candidate to depict texture because of its capability of blending spatial interaction with gray-level distribution.However,GLCM is considered as deficient in discriminating the normal and abnormal parts of texture,and in computation efficiency as well.In order to overcome these drawbacks of GLCM,a method of fuzzy label co-occurrence matrix(FLCM) is proposed to detect the textural imperfection.In this method,textural features such as the probability density distribution of the gray levels,the intrinsic dominant orientation and periodicity in the texture,are extracted firstly to set some key parameters of FLCM,and then all gray-levels are classified into several textural tonal classes in a certain rule;the fuzzy membership degrees of each gray-level to each tonal class are computed based on the corresponding posteriori probability,finally the FLCMs are calculated and some simple features are extracted from the FLCMs,and outlier detection is applied to discriminate imperfection from normal texture.It is proved practically that this method is simpler and has better performance in detecting textural imperfection than GLCM.关键词：textural defect detection;fuzzy label co-occurrence matrix(FLCM);fuzzy membership function;gray-level co-occurrence matrix(GLCM)4061|363|0更新时间：2024-05-07
- 摘要：The phase correlation method provides straightforward estimation of rigid translational motion between two similar images. It is often claimed that the original method is best suited to identify integer pixel displacement. Experimental results show that the phase correlation matrix is rank one for a noise-free rigid translation model. The property leads to a new low complexity method for non-integer translational motion. This method based on singular value decomposition estimates the slope of phase by a least-squares fit and well-known Fourier shift property whose speed is nearly three times as quickly as original method. The method is robust in the presence of noise and shelter because of using phase information in the frequency domain. Experiments prove that this method can effectively restrain noise and accurately detect the shift of object sheltered, which resolved the problem of object shift detection for images seriously polluted by noise and cloud in the space. It provides a strong basis for accurate object orientation.关键词：phase correlation;Fourier transform;singular value decomposition;subpixel3369|248|0更新时间：2024-05-07
- 摘要：Driving fatigue is one of the main causes that result in highway crashes.The information from lane departure detection can evaluate the state of driving fatigue.This paper presents an algorithm combining scan line and region-growing method based on video images analysis,in which lane-marking recognition and driving deviation detection are realized.We use the region of interest(ROI) to figure out the current track of the driving vehicle,and give a frame disposal strategy concerning the track's parameter so as to meet the need for real time management.The result of our algorithm will provide necessary data for the future analysis of driving fatigue.关键词：Driving fatigue;lane departure;scan line;region-growing3771|319|0更新时间：2024-05-07
- 摘要：This paper analyzes the process of lane departure detection approach.In our vision-based Lane Departure Detection system,we use a single camera as input.In this paper,we discuss how to detect the lanes marking on the road and get the relationship between vehicle and road.Some measurements are derived to calculate TLC(time to lane crossing),for measuring the position of the vehicle relative to the lanes.Besides,the forward-looking predicting mode is introduced to establish the relationship among driver,vehicle,and road.Further more,the criterion can be got to tell whether the vehicle is going to depart the lane without conscious.Simulations show that our vision-based lane departure approach does provide an effective alarm when the state of driver goes wrong.Experiments are taken on vision navigation system for HONGQI prototype,using lane change rather than lane departure.Images obtained from forward looking camera are preprocessed to gain lane markers,and TLC curve is then gained from the alteration of the lane markers position.The results prove that the analysis and simulation above can be feasible.The hazard of lane departure can be forecasted through the alteration of vehicle-road relationship and TLC parameter.关键词：lane departure detection;time to lane crossing(TLC);driver modeling;active safety system3342|263|0更新时间：2024-05-07
- 摘要：A novel approach,curve tree,is proposed for curve representation and encoding.Binary trees composed of directed relative height are adopted to describe curves.Any curve corresponds to a curve tree.We may take the first n levels of the curve tree,which compose a binary tree roughly describing the curve.The more levels we get,the finer it describes the curve.The representation is invariant to rotation,scaling,and translation.Based on this descriptor,curve distance is defined to weigh the similarity between curves.关键词：curve representation;curve tree;binary tree;directed relative height;curve distance4211|268|0更新时间：2024-05-07
- 摘要：This paper describes a new approach to image matching.Edge detection uses SUSAN(Small Univalue Segment Assimilating Nucleus) method at low level image processing.Integration features matching can complete object recognition and tracking based on invariant moments in combination with configuration and intensity information.Feature detection with SUSAN method locates precisely and is not sensibly for local noise.Seven moments of image have translation invariant,rotation invariant and scale invariant.Simulations also show that the algorithm is efficient for image with intensity variety,geometry aberration and noise.关键词：SUSAN method;integration features;invariant moments;object detection;matching and tracking11878|942|0更新时间：2024-05-07
- 摘要：Video clip retrieval plays a critical role in the content-based video retrieval.However,existing clip retrieval methods mostly focused on retrieving similar video clips from pre-segmented clips.To automatically segment and retrieve similar video clips from a continuous video database,this paper presents an efficient method for video clip retrieval in which equivalence relation theory is applied to video clip retrieval for the first time.Matching function is defined in terms of equivalence relations corresponding with shots of given video clip.Several similar video clips are automatically segmented from video database using sliding shot window according to matching degree between two video clips.Afterwards,equivalence classes are mapped into matrix representation.Therefore various factors are computed to rank the similarity of the selected video clips by characters of the matrix.Experimental results showed that the method could segment similar video clips from continuous video database quickly,exactly and automatically.关键词：video clip;content-based video retrieval;equivalence relation;matching function;sliding shot window3878|259|0更新时间：2024-05-07
- 摘要：For the problem of remote sensing image fusion,we propose a new scheme based on residual error in this paper.This scheme restores the high-resolution residual errors for multispectral images by fusing the high-resolution residual error extracted from panchromatic image and the low-resolution residual errors extracted from multispectral images based on principal component analysis(PCA).The assessment of experimental results from subjective visual effect and objective statistical analysis indicates that the proposed scheme has better performance in preserving the spectral characteristics of the multispectral images,while improving the spatial resolution,than the conventional image fusion methods such as the methods based on HIS(hue-intensity-saturation) transform,PCA,and wavelet transform(WT).关键词：image fusion;principal component analysis(PCA);high-resolution residual error;low-resolution residual error4818|432|0更新时间：2024-05-07
- 摘要：To bridge the semantic gap between audio feature and high-level semantic concept,an approach for semantic-audio content Analysis is presented in this paper.Hidden Markov model(HMM) is trained for modeling BE.In order to extract G_BE corresponding to a semantic window,Bayesian decision theory is used to eliminate the analysis window not belonging to any predefined HMM.Then,each of the residual analysis windows within the semantic window is classified to BE class by criterion of maximum Bayesian posterior probability.Ignoring the order and repetition of BE,G_BE is got.Logic definition of high level audio semantic concept is the connection of G_BE and HC,through which HC can be extracted.The experimental results demonstrate that the proposal approach could extract HC like human thoughts,and could bridge the semantic gap to some degree.关键词：semantic-audio content analysis;high level semantic-concept;semantic-video analysis;HMM3536|252|0更新时间：2024-05-07
- 摘要：With the improvement of numerical simulation's precision and temperature test's measure methods,better test visualization technique is required.This system integrates numerical simulation and test together,to for visualiz test data,which is a new way for improving test analysis and evaluation.Test data is typically collected in a waveform data type while simulation typically presents results in 3D geometry.In addition,test data locations are sparse compared to relatively dense FEA mesh results.This system uses position matching and mapping method to generate node data on basis of test data and FEA results,and plots generated data on FEA model in color cloud picture mode,thus makes it possible for visualizing all remaining nodes which can't be collected by test. In the system,specific data structure and interactive method such as cutting,probing and comparison are designed.The construction method of this visualization system can also be applied in statics test and acentric test.关键词：temperature test;numerical simulation;visualization;color cloud picture3167|238|0更新时间：2024-05-07
- 关键词：circular segment;spherical surface;B6zier polynomial;approximation3620|312|0更新时间：2024-05-07
- 摘要：With the rapid development of data acquisition techniques,especially 3D laser scanning,surface models of any object constructed by modeling approaches become more and more elaborated,which result in great amount of data to be stored,and decreasing the speed of processing models.It is necessary to reasonably and effectively compress or simplify the surface model constructed from the original scanned point clouds.Based on the data compression algorithm proposed by Garland,a new algorithm for 3D surface model compression with edge contraction is put forward in this paper.The algorithm uses the rule of quadric error to calculate contraction cost of each edge and control the sequence of edges to be contracted.Half space testing is introduced to judge the validity of edge contraction.With this new algorithm,both the triangular faces at the boundary and in the interior of the surface model are compressed synchronously.Experimental results show that the algorithm can preserve the geometric characteristics of the original surface model at high compression ratio,which effectively guarantees the quality of the compressed surface models.关键词：data compression;3D surface model;3D laser scanning;triangulated irregular network(TIN);edge contraction4077|331|0更新时间：2024-05-07
- 摘要：Fragments comparison and reassembly are necessary in art conservation or artifacts restoration.An approach to the problem of object reconstruction from broken fragments of arbitrary 3D objects is proposed in this paper.The triangle mesh model derived from the range data of fragment is preprocessed to remove some unwanted limitations,and then the fragment contour is extracted.The discrete points on the boundary curve are interpolated by quintic B-spline.With the spline curve,the curvature and torsion of every point are calculated and the geometric property is also analyzed.Feature points are detected based on the total curvature of every point,and the contour is segmented accordingly.The similarity between feature segments belonging to different contours is measured according to the variations of curvature and torsion,meanwhile the normal vector is utilized to verify the matchable probability of similar space curves.Thus,the Euclidean transformation of matching contours is computed to achieve the fragment reassembly.Experimental results demonstrate the approach is robust and efficient.关键词：fragment reassembly;contour matching;quintic B-spline;curvature;torsion;normal vector2946|236|0更新时间：2024-05-07
- 摘要：STEP-NC is a new interface standard for data exchanging and sharing between CAD/CAM and CNC,and the CNC based on STEP-NC will be the next generation of CNC controller,which not only holds linear and circular interpolation but also possesses the capability of spline interpolation.A universal NURBS(non-uniform rational B-spline) based interpolator was designed and the interpolation technique based on constant arc increment and interpolation algorithm were inverstigated arc increment.The validity and reliability of algorithm was tested by an instance simulation and machining.关键词：STEP-NC;non-uniform rational B-spline(NURBS);spline interpolation;constant arc increment interpolation2985|247|0更新时间：2024-05-07
- 关键词：fractal;golden section;plant morphologic;attractor;IFS6686|409|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