最新刊期
卷 15 ， 期 9 ， 2010
- 摘要：Performance-based animation has received significant attention because of its naturalness, intuition, and its easy-to-use for novices and non-professional animators. However, most existing performance-based systems either require special and expensive device or can only animate low degree of freedom objects such as a rigid body. This paper presents a new performance-based animation technique for novice animators and develops a 2D character animation system. It uses a pen as the prop and maps its movement to the character in real-time, the character’s motion is recorded to form an animation. The evaluation result shows that the system lowers the barrier of animation authoring and allows novices to create animation simply and quickly.3213|250|0更新时间：2024-05-07
- 摘要：With the development of mobile devices, the pen-based input mode has provided the possibility to use and share chemical knowledge easily. As an application of ubiquitous computing, research on online handwritten chemical formulas becomes a hot area. In this paper, we conclude common relations between chemical symbols and segment a chemical formula after linking the broken strokes. Then a two-level classifier is used to recognize the isolated symbols and the formula is understood with some chemical rules. The experimental results show that our method is robust and feasible when used on Tablet PCs.关键词：online handwritten chemical formulas;two-level classification;structure-based analysis3399|359|0更新时间：2024-05-07
- 摘要：In order to extract the high-precision sub-pixel feature parameters，this paper presents a high-precision sub-pixel feature extraction method based on the multichannel image by use of the phenomenon of the bright pupil. Firstly, the pupil area is obtained by filter on a difference image. Then the edge of the pupil area is detected and cornea reflection area is searched nearby the eye area based on gray level. Then corneal reflex center is located by centroid and the edge of the pupil is filtered to eliminate the impact of corneal reflex to the contour of the pupil. Ellipse fitting is used to locate the pupil center. Finally, multiple parameters of facial features and face position are extracted.We establish a multi-feature extraction process, which provides parameter basis for gaze estimation which is in the next step. The results of feature extraction and the ultimate gaze estimation results made by gaze tracking system verify the effectiveness of the method.关键词：gaze tracking;feature extraction;pupil location;gaze estimation;pupil center cornea reflection (PCCR) technique3257|197|0更新时间：2024-05-07
- 摘要：Vision interaction is one of important aspects of human-computer interaction, and the facial feature extraction is crucial to vision interaction. This paper applies the curvelet transform to the face processing to extract facial feature more effectively. It overcomes the weakness of the wavelet transform which is unable to efficiently extract curve features of face images. An optimized method based on Adaboost and curvelet transform is proposed for face detection. A new approach combining SVM and curvelet transform is designed for facial expression recognition. Experiments on face detection, face recognition and facial expression recognition are carried out. The results reveal that curvelet transform has distinct advantages in facial feature extraction, and lays a good foundation for the further work of the natural human computer interaction.关键词：curvelet transform;face detection;face recognition;facial expression recognition;human-computer interaction3199|218|0更新时间：2024-05-07
- 摘要：How to design digital image denosing filters with good ability of preserving edge and texture information is an international and popular issue in image processing. The unified filtering mechanism of digital TV filter (DTV) and digital bilateral TV filter (DBTV) is described firstly, and then a novel non-local adaptive bilateral filtering mechanism is established. The definition of pixel’s short dependence and long dependence measures are given, and a new non-local digital TV filter (NLTV) is proposed which can deal with Gaussian noise and impulse noise respectively. Experiments show that our proposed algorithm has quite good ability of noise suppressing as well as edge and texture preserving.关键词：image de-noising;short distance dependence;long distance dependence;non local filter;total variation4786|458|0更新时间：2024-05-07
- 摘要：Due to the existence problem and complexity on correlative detection, traditional digital fingerprinting schemes have low efficiency in construction and tracking. To solve this problem, a new fingerprinting scheme based on the FDM(frequency division multiplexing) technique is proposed in this paper. In this scheme, the digital fingerprinting, which is a random sample from Cosine signal with no spectral overlap, is embedded in the middle-frequency coefficients of the original image through the spreading sequences. The performance of anti-collusion by embedding algorithm is also analyzed so as to theoretically deduce the relationship between code length and user size. The experimental results demonstrate the feasibility and efficiency of the scheme.关键词：digital fingerprinting;copyright protection;FDM;Fourier transform2961|179|0更新时间：2024-05-07
- 摘要：Spread-spectrum embedding for video, which can resist many attacks such as video compression and noise addition, is a representative steganography method. In order to effectively detect the hidden information embedded by the method, the steganography procedure is modeled as the addition of additive Gaussian noise. By analyzing the influence on temporal correlation and spatial correlation of video owing to noise addition, a steganalysis algorithm based on spatial-temporal correlation is proposed. The algorithm uses a difference image histogram of four directions to measure the influence of spatial correlation, and temporal correlation is measured by the histogram of frame difference. Experimental results show the algorithm is easier to implement and better in detection performance than other methods.关键词：video steganalysis;spread-spectrum steganography;spatial correlation;temporal correlation3365|184|0更新时间：2024-05-07
- 摘要：A new unique encoding tools are adopted in H.264/AVC. Fully utilization of these tools in error concealment will greatly improve the concealment performance. The tree-structure macroblock division mode with variable sizes is used in H.264/AVC so that the amount of the divided blocks directly adjacent to the lost macro-block may be up to 16. The motion vector of each divided block surrounding the lost macro-block is projected to one point in 3D space. Then the plane fitting based MV recovery algorithm is proposed, in which one plane is constituted to present the MV changing tendency in small regions. Experimental results demonstrate the proposed algorithm is superior to other common used algorithm in both subjective and objective quality review.关键词：plane fitting;error concealment;temporal domain;H.264/AVC3350|219|0更新时间：2024-05-07
- 摘要：Nonparametric region-based active contour models have been widely used in the field of image segmentation. The proposed new model which is based on the use of the structure tensor and the Wasserstein distance belongs to this category. First, the classical tensor structure method is extended in order to express the gradient and color information of the original images. Using the extended tensor structure method, an image can yield more feature channels and more features of the texture can be extracted. Then, we use the Wasserstein distance between each pixels’ local distribution functions of the image intensities to segment the image domain so that the local histograms within each region are approximately homogeneous. Some numerical mechanisms are introduced into the approach, which obviously reduce the computational complexity of the model. Comparing with related studies, experimental results show that proposed model is more precise and stable.关键词：level set methods;structure tensor;Wasserstein distance;histogram;texture segmentation4964|282|0更新时间：2024-05-07
- 摘要：A new image segmentation algorithm combining region growing with C-means Clustering is proposed based on human visual model. Initial seeds are selected automatically and the similarity principle of region growing is adjusted adaptively according to human visual threshold effect. By considering the global influence of region growing to error sum of squares, as well as the local similarity information of seeds’ neighborhood, the proposed algorithm can limit seeds to grow within object boundaries. Experimental results show that the proposed method can produce better segmentation performance with less computational complexity than traditional methods.关键词：image segmentation;human visual model;region growing;C-means clustering3697|299|0更新时间：2024-05-07
- 摘要：A new algorithm for license plate position in complex backgrounds is proposed in this paper. The algorithm is decomposed into two parts, identification of candidate region and elimination of pseudo-region. The edge information and the texture features of license plate region are used to identify the candidate plates. The complexity of the algorithm is reduced and the accuracy of positioning is also improved. Then a sliding strip window is used to select connectivity blocks in the binary image of the candidate plate. And plate syntactic features are combined to evaluate the candidate plate. The low precision of license plate position in complex background and fuzzy image is solved effectively. Moreover, the results of the positioning provide important prior information for following character segmentation. Experimental results prove that the position method has good real-time performance and high positioning accuracy. In particular, it has strong robust performance for complex backgrounds and fuzzy images.关键词：license plate position;pseudo-region;connectivity block;plate syntax;complex background4488|315|0更新时间：2024-05-07
- 摘要：Motion objects detection in outdoor video surveillance is prone to be disturbed by insignificant motions, such as branches swing and wave, and has low accuracy and bad real-time. So a Region of Interest(ROI) automatic extraction algorithm is proposed in this paper, and ROI has the existing and potential significant motion objects. The algorithm constructs the belt-shaped operators to detect the region existing motion objects, and realizes region growing based on color similarity and region degeneration based on disturbance objects, and then gets the potential significant motion region. Adopting the different detection strategies for different regions can improve the accuracy in real-time. Experimental results show that the algorithm is efficient in extracting ROI. In motion objects detection application, the approach can overcome the influence of insignificant motions, improve the accuracy, and reduce the computation complexity greatly.关键词：video surveillance;region of interest(ROI);significant motion;region growing;region degeneration3869|256|0更新时间：2024-05-07
- 摘要：Texture is usually depicted by a gray-level distribution along with a certain spatial interaction. Gray level co-occurrence matrix(GLCM) is an appropriate candidate to depicted texture because of its capability of blending spatial interaction with gray-level distribution, thus, it can be widely applied in texture analysis. When calculating GLCM, the gray-level quantization would be needed in order to decrease matrix dimension, and certain information would be lose. A membership function matrix is established whereby the distance which between the real gray-level and the mean of quantization gray-levels area, and then, a newly co-occurrence matrix, namely fuzzy gray level co-occurrence matrix(FGLCM) is proposed. After appropriate features are selected based on FGLCM statistics properties analysis, the hidden markov model(HMM) classification is applied to divide the classical fracture surface image to four kinds. It is proved practically that FGLCM in this paper is better than the GLCM in depicting textures and the FGLCM combined with HMM is efficient performance in fracture surface images classification, and the recognition rate is 98%.关键词：gray level co-occurrence matrix;fuzzy gray level co-occurrence matrix;fracture surface image;hidden Markov model3265|260|0更新时间：2024-05-07
- 摘要：Conventional image matching algorithms based on mutual information not only consume large amounts of time, but also ignore the pixels’ utilities and spatial relations. In this paper, a novel image matching algorithm using multilevel features is proposed based on quantitative-qualitative measure of mutual information(Q-MI). Firstly, multilevel features are extracted on the edge image, including edge points of interest, edge points and edge neighborhood points. Secondly, according to the characteristics of multilevel features, the Q-MI joint utility for each pixel value pair is computed. Lastly, an optimizer based genetic algorithm(GA) is applied to effectively search the best matching transformation parameters, with Q-MI as the fitness function. Experimental results demonstrate the accuracy, efficiency and robustness of this algorithm.关键词：quantitative-qualitative measure of mutual information;multilevel features;image matching3470|231|0更新时间：2024-05-07
- 摘要：Although there are many feature matching and tracking methods so far, the side effect of moving foreground object, which will cause global motion estimation error, is still an open problem. In order to avoid features, located on the foreground objects, participating in motion estimation, feature effectiveness evaluation is employed to improve feature reliability for the features extracted by the traditional KLT method. Effective features are utilized to estimate global motion and obtain accurate motion parameter, based on which video frames are compensated. However, motion compensation will cause undefined area. There are some approaches to reconstruct the undefined area; nevertheless, they have not considered the effect of fast moving objects in the foreground of the video, which will decrease the video quality after stabilization and content completion. In our proposed algorithm, optical flow between defined areas of current frame and neighbor frame is first calculated, and then it is used as a guide to erode unknown areas. Finally, mosaicking on the base of reference frame is used to obtain a complete video stabilization sequence. Experiment results show that the proposed method is robust to moving foreground objects and is able to realize video frames stabilization with complete content.关键词：video stabilization;KLT;optical flow3705|272|0更新时间：2024-05-07
- 摘要：To deal of umbra overestimation in some current soft shadow algorithms considering only the outer penumbra effects extended from traditional shadow mapping, a real-time approximating algorithm for plausible soft shadows based on depth peeling is proposed. The algorithm extracts the scene silhouette edges in geometry shader and generates the inner and outer penumbra primitives respectively through the silhouette edges. In the final rendering pass, the shading of each pixel is decided by referring to the shadow map and penumbra map, and plausible soft shadow effect which is more realistic than the results of previous methods is finally simulated. This algorithm runs entirely on GPU. Several experimental results show that, this algorithm generates more plausible soft shadows for relatively uncomplicated scenes than other methods and has real-time rendering speed.关键词：GPU;soft shadow;depth peeling;geometry shader3542|278|0更新时间：2024-05-07
- 摘要：In order to remove clouds and shadows more effectively, cloud and shadow distance model is established in this paper at first. And a novel method is proposed for detecting and distinguishing clouds and shadows to form a fusion region map based on this model. After getting the fusion region map, a new image fusion method based on MGA (multiscale geometric analysis) is proposed for removing clouds and shadows. This new image fusion method decomposes the source images and obtains the coarse and detail band parts, then fuses the different regions by using different fusion rules. Experiments show that this new image fusion method can not only remove thin clouds and shadows, but also have much better performance for thick clouds and shadows without any join edge.关键词：image fusion;multiscale geometric analysis（MGA）;remote sensing image;cloud cover3225|299|0更新时间：2024-05-07
- 摘要：Combined the common properties of wavelet and Contourlet transforms (multi-scale, multi-resolution and the different properties of their adaptivity to point-singularity and line-singularity respectively) a novel remote sensing image restoration algorithm based on two-step iterative thresholding was proposed. Inverse problem can be regarded as a class of the convex unconstrained optimization problem. The objective function is composed of the believe measure to original image and regularization item. The traditional algorithm has only one regularization item, but we proposed two regularization items so as to better use the prior knowledge of the image. The theorem is that the half norm in Besov space is equal to the norm in wavelet domain. Then the two-step iterative thresholding algorithm combined with wavelet regularization item and contourlet regularization item is proposed. The method is received obvious effection in remote sensing image restoration. The experiment results on remote sensing image restoration show that the proposed method achieves significant improvement in improved signal-noise ratio（ISNR） and correlation coefficient（CORR）.关键词：wavelet transform;Contourlet transform;two-step iterative thresholding method;remote sensing image restoration3648|255|0更新时间：2024-05-07
- 摘要：With the availability of terrain data of the TB level, how to render terrain data of TB magnitude based on ordinarily computer efficiently, is becoming a hot topic in GIS academia .Organization and management of massive data is the core mechanism for 3D terrain rendering .The traditional data structures as quadtree, octree and so on are not effective, which cannot support real-time displaying and roaming of large-scale terrain scene. According to the mechanism of multi-level buffer, implementation dynamic dispatching of LOD, graphics memory dispatching and multi-core CPU parallel computing strategy, this paper has good results. Finally, this paper puts forward the solution, a real-time Terrain Rendering Algorithm which has no concern with data size. It is verified by experiment on massive data of 1GB, 10 GB, 200 GB, 500 GB and so on, the rendering rate can reach to 30 frame/s. It is faster than most existing algorithms. The experiment result shows the practicability of the real-time terrain rendering algorithm.关键词：terrain visualization;dynamic dispatching;multi-level buffer;multi-core CPU;parallel computing3350|227|0更新时间：2024-05-07
- 摘要：Digitization of topographical maps is significant for creating geographical information system. In this paper, object-oriented approach is introduced into of segmentation scanned map. The procedure is described as follows. First, linear elements and area elements are separated. Then they are partitioned into image objects which are basic cells to classify map features, not single pixels. And a set of features combining color, shape and topological are extracted from each object. Based on the features, a classification criterion is employed to perform the map segmentation. The proposed method overcomes aliasing and false colors induced by the scanning process. The results seem coherent, with name of the “salt and pepper” effects.关键词：scanned topographical map;segmentation;object oriented;color-alias and false colors3361|223|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