最新刊期
卷 17 ， 期 7 ， 2012
- 摘要：As one of a serial of reports,this paper is a survey on multimedia researches and applications in China,2011.Papers about multimedia technology are distributed on various journals because multimedia is a cross research area.Totally 2841 papers published on 9 Chinese journals in 2011 are checked,and 349 out of these papers are selected for they are about multimedia technology and application.Analysis,including categorying and comparion of statistics with those from 2007 to 2010,is made on the seleted papers.The categoris are the same as last year.The analysis shows that digital watermarking,multimedia data management and retrieval,QoS control,multicast,interactive mode and interface,identification using biometrics earn high attention of researchers in China.Meanwhile,multimedia application systems are diversified and intergraded with daily life gradually.We present here an overview on the progress in multimedia technology in China,2011.This article can be used by researchers as a thoroughly references,and also will be helpful for people in technical planning and management.关键词：multimedia;digital watermarking;multimedia data retrieval;QoS control;multicast;interactive mode and interface;identification using biometrics3167|96|0更新时间：2024-05-08
- 摘要：Retinex algorithm deals with the removal of unfavorable illumination effects from a given image.Solving it is typically done by introducing a regularization that forces a spatial smoothness on the illumination, which is computational expensive.In this paper we propose a non-iterative retinex algorithm based on a recent"guided image filter".Assuming a spatial smoothness on the reflectance, a method using two guided image filters is applied to eliminate artifacts caused by noise.Then, a multi-resolution framework combining guided image filtering and wavelet thresholding,is presented.Our framework is very effective in achieving a trade-off between detail enhancement and color constancy.Compared to other enhancement algorithms,our results verify the new approach's efficiency in eliminating artifacts caused by noise,detail enhancement,and color constancy.关键词：Retinex theory;image enhancement;guided filter;bilateral filter5463|137|0更新时间：2024-05-08
- 摘要：Image gray or intensity value attacks, including the absolute modification of gray values, contrast adjustment,parameter histogram equalization, and combined attacks can make many kinds of watermark detection approaches unsuccessful. Addressing these problems in this paper, a robust zero-watermark algorithm using the direction stability of the first principal component vector is proposed, which can effectively solve the conflict between invisibility and robustness. The direction stability of the first principal component vector is mathematically analyzed, and experiments are performed to observe the difference of the stability in the spatial domain and the frequency domain. On this basis, the input image is decomposed by the dual-tree complex wavelet transform(DT-CWT)and the output of the two low frequency bands is segmented into non-overlapping sub-blocks, respectively. After computing the first principal component vector for each sub-block, some vector angles between that of each sub-block and a secret reference vector can be obtained. By comparing the angles from two sub-blocks with the same position in the left and right low frequency bands, zero-watermark feature extraction is accomplished. In addition, we analyze the security and robustness for the proposed scheme. Sufficient experiments with six tested images, twenty kinds of common attacks and fifty kinds of combination attacks are carried out. Simulation results show the proposed scheme is superior in the comprehensive performance composed of robustness, withstanding diversity attack and practical application.关键词：image watermark;gray value attack;combination attack;DT-CWT transform;zero watermark2994|56|0更新时间：2024-05-08
- 摘要：Anisotropic diffusion(ATD) is a very important method for image denoising.The selection of the optimal stopping-time for ATD is one of the most important problems. Recently, Gilboa proposed an estimation method of stopping-time for ATD in Gaussian noisy images based on an optimal SNR. The method uses a noisy patch to estimate the derivative of the covariance of the noise and the redundancy (the result of noisy image minus the denoised image) with respect to the variance of the redundancy. The patch's noise is random Gaussian noise whose mean is zero and whose variance is the variance of the image's noise. The method has two defects. On the one hand, the method needs the variance of the image's noise,which is unknown in practice. On the other hand, the patch's noise is random and the result may be different because of different patch's noise. Our proposed method is optimized for these problems. First, the noisy image is transformed by wavelets. Then the information of edges and textures in the first coefficients of direct wavelet (HH) is reduced by using the inter-scale correlation of wavelet coefficients. Last, the reduced HH is taken as the patch's noise. Experiments show that the proposed method can solve the two defects and the denoised image by the proposed method has a better PSNR.关键词：anisotropic diffusion;optimal stopping-time;wavelet transform;Gaussian noise;signal-noise-ratio3256|53|0更新时间：2024-05-08
- 摘要：Traditional line feature detection methods based on structureless algorithms of the Beamlet transform not only suffer from overlapping and ambiguities, they also can not detect the target information effectively. Moreover, they can not describe the detail information when extracting the line features of a complex image. Therefore, we propose a new line feature extraction algorithm based on an improved Beamlet transform and the Canny operator. First, the Beamlet transform is performed. There is at most one optimal Beamlet in a dyadic square after improving the Beamlet structureless algorithm and using the new drawing rule and the new energy function. Second, the Canny operator for edge detection is used with a larger Sigma in order to detect only obvious edges. Finally, line feature are detected by a combination of both. The algorithm is evaluated under several aspects, such as the continuity of the line feature extraction, the false detection rate and the miss detection rate. Moreover, this method is compared to existing methods. The experimental results show that our proposed method not only overcomes their weakness such as fractureing, overlapping, sambiguities, false edges and so on, but also effectively improves the accuracy and continuity when extracting line feature of complex image.关键词：Beamlet transform;Canny operator;complex image;line feature extraction3144|83|0更新时间：2024-05-08
- 摘要：Gait is an important biological characteristics in the long distance video surveillance field. Nowadays, almost all gait recognition researcher focus on gait recognition only under one single condition.However, the gait recognition rate rapidly decline in blended conditions, for example when somebody is wearing a coat or carrying a bag. Based on our analysis of the gait timing characteristics during the human movements,we propose a new gait recognition approach that expresses dynamic information and static information by using a dynamic Bayesian networw(DSIF-DBN). The DSIF-DBN contains three levels of states and for every time slice of the DSIF-DBN model is expressed by the fusion of dynamic information and static information . This model can exectly express the timing characteristics of the gait, which are the body posture and the range of motion, as well as other gait rhythmic change characteristics. Experimental result show that the DSIF-DBN model recognizes gait with high rates and good robustness to noise and lost of information. The DSIF-DBN model can fuse the dynamic information as well as static information and can greatly reduce the impact of gait recognition rates when somebody is wearing a coat or carrying a bag.关键词：gait recognition;video surveillance;dynamic Bayesian network;information fusion3403|60|0更新时间：2024-05-08
- 摘要：Multi-modality based adult video detection is an effective approach for filtering pornographic information.However,existing methods lack accurate representation methods of audio semantics.Therefore,a novel method is presented in this paper to fuse audio-words with visual features for adult video detection.First,we propose a periodicity-based segmentation algorithm of units of energy envelope (EE).Audio streams are divided into sequences of EE.Second,audio semantics representation method based on EE and BoW (Bag-of-Words) is presented to describe the features of the EE as the occurrence probabilities of audio-words.Integrated weighting methods are used to fuse the detection results of audio-words and visual features.Furthermore,we propose a periodicity-based decision algorithm to judge adult videos to cooperate with the preceding periodicity-based segmentation algorithm.Therefore,we make full use of the periodicity.Our experiments show that our approach remarkably improves the detection performance compared with the method based on visual features.The true positive rate achieves 94.44% while the false positive rate is 9.76%.关键词：adult video detection;multi-modality fusion;audio-words;visual features;units of energy envelope3985|67|0更新时间：2024-05-08
- 摘要：An automatic method for the measurement of an object's position and attitude in an image sequence is proposed. First, a 2D template of an object's planar region is reconstructed with brightness and the real geometry size, by using the homographies and measurement results from the previous frame image; second, with different position and attitude parameters, the template is projected onto the image plane through the imaging formulations. When the template's projection matches the object region of the current frame properly, the position and attitude parameters are considered to be the measurement results in the current frame image. An optimization model is built and solved to achieve the measurement for the object's position and attitude. The experiments show that the method can measure the position and attitude of an object involving a typical planar region automatically from image sequences.关键词：image sequence;optimization model;position and attitude measurement;object tracking2966|65|0更新时间：2024-05-08
- 摘要：In 3D image analysis, detection and reconstruction of boundary surfaces is a very important problem. Some methods have been developed for extracting or approximately computing continuous implicit boundary surfaces from 3D images. However, they have the drawback of incorrectly classifying some zero-crossing surface patches as boundary surface patches. In this paper, we present a new method to detect and trace boundary surfaces from 3D images. First, all cubes containing boundary surface patches are divided into two categories: cubes containing one connected boundary surface patch, called and first class of edge cubes, and cubes containing two or more disconnected boundary surface patches, called the second class of edge cubes. Then, according to the continuity and the connectivity of the boundary surface, we can track all the edge cubes from both classes. Finally, the boundary surface patches contained in the first class of edge cubes can be extracted directly, and the boundary surface patches contained in the second class of edge cubes are extracted based on the adjacent first class of edge cubes. Experimental results show that the proposed technique is feasible and effective, and can effectively overcome the shortcomings of existing methods.关键词：3D image analysis;edge detection;zero-crossing surface;edge surface tracing2630|69|0更新时间：2024-05-08
- 摘要：In this paper, we propose a new method for foreground object detection based on the Kernel: Density Estimation of a local spatio-temporal model (LST-KDE), which overcomes information redundancy and the large calculated quantity problem in the training phase as well as the manual adjusting time window size and shadow problem in the detection and updating background phase. The LST-KDE algorithm uses the k-means clustering algorithm to optimize the sample set and to choose the key frames in the training phase. Therefore, it can avoid information redundancy and the large calculated quantity problem. In the detection and updating background phase, the LST-KDE algorithm constructs a local spatio-temporal model. This method can not only adaptively set the time window size by using history frame information in a temporal model, but also uses color and texture features described with the local binary pattern (LBP) algorithm to remove shadows in the spatial model. The experiment in a complex environment demonstrates that the proposed method outperforms recent state-of-the-art methods.关键词：kernel density estimation(KDE);local spatio-temporal pattern;K-means;LBP algorithm3128|65|0更新时间：2024-05-08
- 摘要：Since the spectral residual (SR) method of the visual attention model has poor contrast saliency maps and unsatisfactory detection of saliency details. In this paper, we propose a saliency detection method based on spectrum analysis by discussing the relationship between spectral characteristics of the image and the saliency. This method keeps the phase spectrum and tunes the amplitude spectrum using a piecewise non-linear function for the purpose of inhibiting the redundant information and enhancing the saliency image information. Experimental results show that saliency detection method based on phase spectrum and tuning amplitude spectrum (PTA) obtains a saliency map, which has better contrast, and allows for a better deteetion of the details of the saliency information.关键词：phase spectrum;amplitude spectrum;nonlinear tuning;saliency map3165|109|0更新时间：2024-05-08
- 摘要：Currently, image analysis is the central issue in the field of computer vision. It reflects the inclusion relationships between the scene and the object. In the process of scene analysis, properly used contextual knowledge can improve the applicability of the scene analysis model and the accuracy of the object recognition. We do our research on the scene analysis from a hierarchical perspective of "database-scene-object-part-visual words", and we add global contextual information and local contextual information into a generative graph model based on a hierarchical dirichlet process, analyzing the scene at the scene-level and the object-level at the same time. In this way, the result of the scene analysis can be used to constrain object recognition, while the object recognition result feedback effect on the scene analysis.关键词：scene analysis;generative graph model;contextual information;feedback mechanism;object recognition3435|66|0更新时间：2024-05-08
- 摘要：A two-level algorithm that is integrating contextual information to recover the structure of a single image is presented. Due to the structural features of outdoor scenes, we can classify the structure of a scene into three categories: sky, ground, and vertical objects. First, we over-segment the image into homogeneous regions. Then, we recognize the regions with significant features as "definite regions", and the regions we can not classify as "undetermined regions". Next, every nearby definite region with similar features as the undetermined region will vote for an undetermined region. The class with the most votes is assigned to that undetermined region. Finally, we construct a 3D model of the scene. Experiments show that due to the exploitation of the contextual information, almost 92.3% of the pixels can be recovered successfully, which is better than the performance of the existing method, whose result is 88.1%.关键词：scene understanding;pattern recognition;image segmentation;contextual information3255|50|0更新时间：2024-05-08
- 摘要：In this paper,we propose to use the Graph Cut optical flow estimation with a local space term controlling factor to detect and process the translational occlusion in the optical flow field. The occluding mechanism is firstly analyzed, and the discriminating formula for occlusion is constructed. Based on this time of forward and backward optical flow fields, the next the energy functional is established to control the smoothing direction by adjusting the local space term controlling factor, and the network flow graph is reconstructed. In addition, we show that this algorithm could interrupt the transition of misjudging occlusions between interating. Furthermore, it makes the Graph Cut optical flow estimation deal self-adaptively with these occlusions. The experiments show that this algorithm could detect and process invalid region in the optical flow field caused by occlusion.关键词：optical flow estimation;translational occlusion;local space term controlling factor;successive;-expansion move graph cut;frame order2898|43|0更新时间：2024-05-08
- 摘要：An improved algorithm for Harris rapid sub-pixel corner detection is proposed by considering the limitations of inaccurate localization and the low efficiency of the Harris algorithm. The improved algorithm is used to reject non-corners and false corners by screening the corners two times. After detecting the initial corners with the non-maximum suppression of the corners response function, the initial corners are refined to sub-pixel corners by the Euclidean distance between the corner cluster and an ideal corner using a weighted least squares method. The location of the initial corners are considered as the center of the searching corner cluster, and the value of the corners response function is considered as weight. Our tests show that the improved algorithm is effective and practical.关键词：Harris algorithm;corner detection;sub-pixel corner;corner response function (CRF);Euclidean distance4242|90|0更新时间：2024-05-08
- 摘要：Point cloud segmentation is widely used in point cloud parameterization, shape recognition, and model editing. A point cloud segmentation algorithm based on a minimum spanning tree is proposed, which includes four steps: generating banded segmentation boundaries,region growing, splitting banded boundaries, and generating the final regions. The Snake model is used to extract the dividing lines, and the lines are expanded towards both sides to generate banded segmentation boundaries. Then the Minimum Spanning Tree is used to extract all interior points in each region using region growing. At the last step, the banded segmentation boundaries are split to several parts, and each part combined with its region to generate the final regions. Experiments show that the algorithm can avoid over segmentation or under segmentation and generate smooth segmentation boundaries. Compared with the Level Set segmentation algorithm, the algorithm can segment point cloud more efficiently.关键词：point cloud;Snake model;model segmentation;minimum spanning tree;K-nearest neighbors;region growing3999|87|0更新时间：2024-05-08
- 摘要：Selecting regional features in ultrasound images of lymph nodes is important for clinical diagnosis. Most of the current feature selection algorithms are time-consuming and lead easily to a premature convergence. In this paper, a new novel discrete differential evolution (DDE) algorithm based on virus-evolution is presented to solve the cervical lymph nodes features selection problem.We call it virus-evolutionary discrete differential evolution (VEDDE) algorithm. Biological virus mechanism and the infection-based operation between host and virus are introduced in the DDE which can maintain the diversity of individuals while retaining the best search information and improve the fitness function value and the speed of evolution. The proposed algorithm has been tested on many clinical ultrasound images of cervical lymph nodes. The classification accuracy is 98% and the average number of iterations is only 30 times,which indicates that the proposed algorithm is valid.关键词：virus-evolutionary;discrete differential evolution;ultrasound images;features selection2846|50|0更新时间：2024-05-08
-
Liver image segmentation algorithm based on the Snake modeland optimized by watershed transformation
摘要：Directional Snake is a classical algorithm in the active contour models, and is widely used in the field of image segmentation and video research in the past few years. Aimed at sensitivity to the initial contour and lack of curvature constraints in the formulation of the function, an automatic contour segmentation algorithm based on an improved watershed transformation and active contour model is presented. First, a modified watershed algorithm based on the marker function and the mandatory minimum technology is proposed in this paper to deal with the over-segmentation. Then,the improved watershed algorithm is adopted for pre-segmentation, and the extracted object contour is taken as the initial contour for the Snake model. Finally, an external force,which is related to the curve shape,is added in the Snake model for making up the lack of curvature constraints in the formulation of the energy function for precise segmentation computation. The improved Snake model can achieve good results in the liver image recognition and segmentation when applied to the MR images of the abdomen.关键词：Snake algorithm;image segmentation;watershed transformation;liver medical image4245|100|0更新时间：2024-05-08 - 摘要：As spatial information plays an important role in remote sensing analysis, more and more researchers pay focus on spectral-spatial endmember extraction. An improved endmember extraction method with a spatial preprocessing module, which uses watershed with normalized cuts to avoid over-segmentation and producing accurate results from spectral mixture analysis, is proposed in this paper. According to the experiment in this study the spatial-spectral endmember extraction method can generate a more accurate pixel un-mixing results during image segmentation.关键词：remote sensing image segmentation;watershed transform;Normalized Cut;endmember extraction;hyperspectral satellite image3170|62|0更新时间：2024-05-08
- 摘要：Based on the non-local means filter (NLMF), we propose an improved denoising algorithm for synthetic aperture radar (SAR)images. In the framework of the NLMF, combined with the characteristics of SAR images, we improve the NLMF using pre-generated similar sets and the two-dimensional principal component analysis (2D-PCA). First, we choose suitable image slices to generate the similar set, and then extract the main features of these image slices by applying the 2D-PCA, which can reduce the effect of the speckle noise on the similarity. Finally, we measure the similarity of the image slices based on the processed similar set. In the end, we show the noise reduction experiments of the simulated SAR images and the real SAR images. Compared with traditional Lee filter, Kuan filter, Gamma-Map filter, and the NLMF algorithms, the experiments confirm that our algorithm can achieve a better result on both: the edge retention and the smoothness of the consistency area. Simultaneously, the image quality is improved in all aspects.关键词：SAR image denoising;similar set;two-dimensional principal component analysis (2DPCA);non-local means filter3880|80|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