最新刊期
卷 8 ， 期 7 ， 2003
- 摘要：The abstraction of an article is a short summary of a document. With the development of video processing, it comes into a similar concept—Video Abstraction. The idea of video abstraction is in very much the same way as text: a short summary of the content of a longer video document. This paper first explains the basic concepts of video abstraction, introduces the applications of video abstraction and classifies video abstraction into video summary and video skimming. Then it emphasizes on the realization techniques of video abstraction, including keyframe extraction and multi-features fusion etc. At last this paper makes some conclusions about video abstraction and prospects for some approaches to it.4003|180|0更新时间：2024-05-07
- 摘要：Due to the development of streaming media and emerging of content based functionalities in MPEG 4, object based segmentation technology becomes a popular research in video field. Video segmentation provides an easy and efficient way for video retrieval and coding, whereas it is a difficulty issue in computer vision. From shot based to extraction of video object planes (VOPs), video segmentation develops to extract video object(VO). In fact, shot based segmentation is one of the primary steps in the true object based video segmentation. Currently, segmentation technology for content based representation is still premature. This article gives an overview of existing techniques for object based segmentation. The performance, relative merits and limitations of each of the approaches are comprehensively discussed and contrasted. Different from the traditional classification approaches, this paper presents a new view of analyzing the recent research in this area. Some important and advanced video segmentation algorithms are analyzed and compared both in theory and experimental results. According to different information used in those algorithms, they are classified into several classes. This classification criterion is more revelatory and useful to developing new segmentation algorithm. Finally, some existing problems worth discussing and directions need to further research are proposed.2831|198|0更新时间：2024-05-07
- 摘要：The performance problem of spatial database limits its application and development seriously. Spatial join is the most complex and time consuming operation in spatial database system. Its efficiency determines the performance of the whole spatial database system to a great extent. Although there are many spatial join algorithms already, cost estimation and query optimization of spatial join operation need further study. Most spatial join algorithms are implemented based on R tree index, but if the corresponding relations have no indices, or only have partly indices, special algorithms should be used to handle the situation. Cost estimation models of each algorithm need a relatively uniform calculation method. Considering the characteristic of spatial database, it's reasonable to use I/O cost to estimate the complexity of each algorithm. Based on the above approaches, and because complex spatial queries may include multiple relations for spatial join, dynamic programming algorithm should be used to choose the proper join order which has minimum cost. It becomes a universal algorithm framework. Through the complexity analysis of the algorithm framework, the spatial database query optimization system implemented base on this approach will have better spatial and temporal efficiency, and can handle very complex spatial queries.2610|179|0更新时间：2024-05-07
- 摘要：The inverse problem of process tomography is also named imaging algorithms, it not only agree with the interactional principle of the stimulating field and measured substance, also match the spatial array geometry of sensors. The performance of proposed imaging algorithms(including image quality and calculating interval per frame) is a key question whether the process tomography can be applied to the industry process monitoring and control system. In order to get excellent reconstructing image, a sort of Linear Neural Networks method for image reconstruction are proposed. Through building the linear models of forward and inverse problems to optical tomography, this algorithms calculate the forward problem firstly to obtain pairs of modes of image projections relation, then which are used to train and build the Linear Neural Networks; Finally, the inverse problem can be reflected through using the trained linear neural networks. The numerical simulation demonstrated that the method is a robust imaging algorithm, the quality of image is excellent, and the temporal performance of imaging is very good with this method.关键词：Computer graphics;optical tomography;Linear neural network;Reflection;inverse problem2160|187|0更新时间：2024-05-07
- 摘要：It makes a totally disconnected fractal image.The method for getting the IFS code from fractal image using the IFS code to draw the fractal image based on Iteration Function System are discussed.Traditional fractal method has been applied to the simulation of natural scenery.This paper attaches importance to the improvement of traditional fractal modeling.A iterated function systems is lifted liter iterated function systems to use similarity cantor rally and using the method of adjustment parameter make a large amount of fractal image totally disconnected and the relation is found the similarity cantor rally and the fractal image.It uses varying parameter for getting a large amount of fractal image.A simple image technique is given for computer simulation of the fractal image.The experimental result is shown to demonstrate that the model is simple and valid.I has been known that Iteration Function System is a very important method for generating fractal images.It show that Iteration Function System is also an effective way to generate ordinary planar curves.关键词：Computer graphics;iterated funtion systems;fractal image;Similarity cantor rally;Totally disconnected2540|191|0更新时间：2024-05-07
- 摘要：The more priori knowledge used in removing image noise, the better will the effect of removing image noise be, but we can only get a noise corrupted image in general condition, and can't get this priori knowledge, so effect of removing image noise is not good. A novel filter for removing salt & pepper noise of image is proposed in this paper for solving this problem. Firstly, the filter provides an effective approach for estimating histogram of original image, then gets a membership function through the histogram, and then establishes a weighted fuzzy mean filter based on this membership function, so the filter utilizes the information of the estimated histogram to remove salt and pepper noise; Secondly, the filter does an operation of excluding maximum and minimum gray intensity to the pixels of filter windows in order to remove salt and pepper noise points. Experimental results show that the filter gives superior performance compared with conventional filters and other fuzzy filter, especially, the superiority will more obvious when noise probability exceeds 0 3.2975|200|0更新时间：2024-05-07
- 摘要：Comparing function or morphology between individuals requires non rigid matching, because the detail spatial structure difference between the image pair to be matched is too complicated to be modeled by any parameterized transformation. The goal of deformable matching method is to remove structural variation between the image pair to be matched. In this paper, a new method of deformable image matching based on hybrid elastic models (HEM) is proposed. The method, which need not extract features, works directly on grey level images. The algorithm first globally aligns images with a principal axis method, and then utilizes the linear spring net model for the correspondence and the thin plate spline for the non rigid mapping. This method takes multiresolution strategy to approach better matching. The elastic constant of the spring model will decreases as the process proceeds. Some experiments are performed on both synthetic and segmented medical images. It is shown that our hybrid elastic models can be successfully applied into the deformable image matching to remove the detail structural variation, and achieve good results.2494|201|0更新时间：2024-05-07
- 摘要：In this paper a novel algorithm is proposed for automatic segmentation of moving objects. Quick and precise abstraction of the moving objects enables a high level understanding and analysis of video contents and is very important in context based video coding (MPEG 4). Different from traditional approaches in which only luminace information is used, color information is also used in this algorithm as a reference in preprocess and the preprocess result is good. Traditional canny algorithm is improved here and is added a new ability to process two temporal correlative frames by which we get objects' refined but uncontinuous edges. An approach is proposed to get auxiliary edges by which refined edges are mended to be continuous edges of moving objects. Experimental results show that it's a fast and effective algorithm.2771|185|0更新时间：2024-05-07
- 摘要：Two dimension adaptive threshold segmentation algorithm applied to the segmentation of microscopic cellular image is considered in order to improve the result of segmentation. Based on the characteristic of two dimension histogram of microscopic cellular image and the require of segmentation, one of the two dimension is the pixel's gray value and the other is its neighboring average gray value. usually. at the positions of target or background, gray value of pixel and its neighboring average gray are similar; at the edge of target and background, gray value of pixel and its neighboring average gray are very different, so the pixels of target and background will appear around the diagonal. the sections of the object and the background can being changed, at the same time changing the step's value of searching optimal threshold value, using occur times instead of probability distribution and recursive computation instead of plenty of repeat computation, the improved fast two dimension segmentation algorithms for microscopic cellular image adaptive thresholding segmentation are provided and carried. Simulation shows that the improved algorithms reduces computation complexity greatly and reduces the running time of the algorithms, and the improved algorithms has stronger power against noise and gets clearer edges of targets than original one. From simulation result for cellular image, it could be seen that the improvement and simplification are both valid.2566|229|0更新时间：2024-05-07
- 摘要：In order to provide doctors with comprehensive, visual, and accurate diagnostic information, the 2D and 3D section display method for medical image volume dataset is discussed. For the case of 2D section, by means of human computer interaction, the path of arbitrary section is given by the user, then the operation of quantization and sampling is implemented by virtue of solution train of thought of reverse engineering, each point position along the path is attained, so is each point on the section plane. Through the operation of resampling and interpolation, the 2D arbitrary section display for medical image volume dataset is realized. As to the 3D case, by using the same image sampling and interpolation operation as those in the 2D case to get the texture data array, then in combination with the technique of texture mapping, mapping the texture data into the 3D coordinate space, thus the result of 3D spatial section display is achieved.2668|225|0更新时间：2024-05-07
- 摘要：This paper proposes a new color image edge detection method, which utilizes quaternion based representation of color images and is based on a locally defined convolution filter that directly operates on the quaternion domain. The filter, which is a convolving kernel as a function of the pixel under consideration, is applied only to a pre defined neighborhood of the current pixel. Based on the properties of vector dot product and vector cross product between color pixels in quaternion domain, the convolution results are used to determine if the current pixel lies inside a locally homogenous region. A pixel that does not lie inside a locally homogenous region is classified as an edge pixel. Experiments with both synthetic and real color images are presented, along with comparisons with some other existing methods, illustrating the advantages of the proposed method.2610|189|0更新时间：2024-05-07
- 摘要：Real time segmentation of moving objects in image sequence is a crucial step in traffic surveillance which include many different sub modules such as vehicle detection, vehicle statistic, real time tracking, speed measurement, etc. A typical method is background subtraction. Many background models have been introduced to deal with different problems at present. In the paper, we propose an adaptive HSV color background model with shadow detection to segment moving objects. We propose to operate in the Hue Saturation Value (HSV) color space, instead of the traditional RGB space, and show that it provides a better use of the color information, and naturally incorporates gray level only processing. At each instant, the system constructs three Gauss distribution for a pixel and maintains an updated background model, and a list of occluding regions that can then be tracked. However, problems arise due to shadows. In particular, moving shadows can affect the correct localization, measurements and detection of moving objects. This work aims to present a technique for shadow detection and suppression used in adaptive color background model. The major novelty of the shadow detection technique is the analysis carried out in the HSV color space to improve the accuracy in detecting shadows. The details of the algorithm are outlined and the experimental results are shown and evaluated. The results show that this algorithm combines the advantages of veracity and of runtime, and fit for real time detection.3427|195|0更新时间：2024-05-07
- 摘要：Digital coal petrography analysis system introduced in this paper is an integrated system that has three main functions: image acquisition, composition recognition and analysis of quality of coal. The key techniques of this system are auto focus, auto exposure, texture analysis, pattern recognition and analysis of coal petrography. A price function is used to implement auto focus; it is the sum of the gray grads of all pixels in three small windows in the same image, and is determined by the Z Axis of the lens of microscope; the need is met when the function takes extremum value. Auto exposure is achieved by using a test of exposing in a very short time; the correct exposure time can be calculated with the gray histogram of the image obtained in the testing exposure. Studied the features of the composition of coal, the methods of classifying by gray threshold and principal component analysis (PCA) are adopted to implement texture analysis and pattern recognition. Moreover, some other techniques like estimating of parameters by use of artificial neural network are used in this system. The experimental results show that all these methods can greatly improve the efficiency and veracity of analysis of coal petrography. Both the speed and the precision of coal petrography analysis are satisfactory. The system is used in the industrial research successfully.2261|176|0更新时间：2024-05-07
- 摘要：In this paper, a new approach for handwritten Chinese character recognition based on multi-channel PCA (principal component analysis)model is proposed. In terms of the stroke directional characteristics of the handwritten characters, a handwritten Chinese character is decomposed into the four directional sub-patterns at first, namely, horizontal (一), vertical(丨), left up diagonal (丿) and right up diagonal( )sub-pattern, each of which could be modeled by its principal components. Then, based on their four sub-pattern PCA models, a multi-channel PCA model for each category of the handwritten Chinese character is constructed respectively, and the model's reconstruction error is used as a matching measure for the handwritten Chinese character recognition. The method can not only exploit principal components' ability for representing the handwritten Chinese character sample set, but also effectively reduce the training time for modeling. Experimental results on 1034 categories of handwritten Chinese characters indicate that, the proposed method can improve recognition rate by 4.4% comparing to the Euclidean distance classifier, while its training time is much lower than that for modeling handwritten Chinese character directly by its PCA model, showing the effectiveness of the proposed approach.3290|180|0更新时间：2024-05-07
- 摘要：Firstly, this paper studies sameness and differences of VNGIS(Vehicle Navigation Oriented Geographic Information system) and geographic information system.Automatic generation road center lines algorithm based on road inner-outer contour is presented.Then the paper mainly focuses the research of the study of double rectangle template track single road inner-outer contour and generate its center line. one template apply to track road inter-outer contour line, and at the same time triangle network is generated; the other judge if the current position is crossroad.In the end,the paper describe that the center lines of the whole road network are automatically generated when traversing.Road-segment list is created,and through modifying the list,road network is traversed.Experiences show; the arithmetic speediness、precision and sureness acquire road center-lines.关键词：Geographic information system(GIS);Road Center line;Digital Map;Vehicle navigation system2178|204|0更新时间：2024-05-07
- 摘要：A series researches have been made on how to extract road features from satellite data. The main Methods include edge detection and line finding, line filtering and using spatial relationships to extract road features. Another approach to the extraction of road networks involves the use of GIS and rule and knowledge based algorithms. Several workers have examined mathematical morphology as a means of extracting linear features from satellite data. However, many problems are still remained to solve. In this paper, algorithms of mathematical morphology and analysis of road networks are applied to extract road networks digitally from classified imagery to update digital databases and map. The first step involves road image two valued, removes noise data, break line connection, thin and raster convert to vector. The second is analysis the vector networks, connection and selection and eliminate the road arcs. Finally, the Douglas Peuker algorism is used to smooth the road networks. The test area is Nanjing Jiangning County. The satellite data is SPOT multi spectral image. IKONOS high resolution image is tested too. The comparison is made between the results of the extracted road networks and visual interpretation from location accuracy and extraction accuracy. The results described here show that the technique appears to be most effective in areas of recent road development. It is very useful for land managing and planning and is an effective method to update GIS and map.2748|225|0更新时间：2024-05-07
- 摘要：While executing tasks such as sea surveilling, maritime searching and rescue, sea pollution monitoring utilizing remote sensed images, the coastal line feature should be determined at first. Thresholding method is a type of simple but valid method for image segmentation, likewise, they can be used to detect coastal line feature in remote sensed images. However, while conventional thresholding methods used to do it, they are always short of enough discriminating ability to objects' shadow, weak scattering vegetations, dark artificial buildings, sea gulf blurred by noise along costal line. This paper proposes a multi threshold based morphological approach, which divides the isolated regions by thresholding detecting into intra continent, exterior sea, and along coastal isolated regions at first, and then utilizes two definitions and morphological operators to process along coastal regions further so as to improve the detecting accuracy and decreasing false detecting, especially to enhance detecting accuracy for above objects' shadow, vegetations and dark artificial builds. Experiments are executed and the results exhibit the proposed approach possessing better performance than conventional thresholding approach.2822|197|0更新时间：2024-05-07
- 摘要：Efficient content based image retrieval has developed tremendously in many application areas. Content based image retrieval research areas had established many systems, but these systems have deficiency for actually usage:(1) these systems are expected to process different kinds of image retrieval by same method and (2) these systems are designed without considering user requirement. In fact, different kinds of image retrieval need different retrieval mode. So it presents image interactive retrieval based on integrated region similarity match. This system is an image retrieval system with interaction and semantic classification of based on wavelet feature extraction and image integrated region similarity method. Compared with other retrieval methods of based on image content, the method enhances retrieval efficiency by permitting adaptive search and interaction and narrowing search area. Experiment result shows that the system is more precise and efficient than other retrieval systems.关键词：Computer information management system;Image retrieval;Integrated region similarity;Wavelet feature extraction;Interaction method2449|186|0更新时间：2024-05-07
- 摘要：Behavior animation of virtual human is an interesting branch of computer animation in recent years. Many researchers focused on the local expression animation of virtual human emotion, such as facial animation system, and did not consider the reason of emotion to a specific virtual scene. Emotion is the result of interaction between virtual human and virtual environment, the emotion of virtual human has not gained clear description up to the present in the field of computer animation. This paper presents an animation model of emotional behavior for virtual human on the basis of psychology theory. First, this paper introduces the concept of emotion set and emotion expression sets, and sets up a function from emotion state to emotion expression; Second, this paper pays attention to the reason of emotion and introduces the concept of emotion origin, if the strength of an emotion stimulus is bigger than the resistive capacity of an emotion, then the emotion will occur; Third, the emotion state can be expressed as a finite state machine (FSM), and an emotion transfer flow is presented. As a result, a demo of virtual human emotion behavior is realized on personal computer by calling the Microsoft Direct3D API with the programming visual c.2957|180|0更新时间：2024-05-07
- 摘要：Virtual assembly is one of the most important applications of virtual reality in product design fields. Virtual assembly method based on physically modeling is proposed for further investigating the kinematics and dynamics performance of assembling parts during virtual assembly process in this paper. The automatic constraints mapping method, which maps geometrical constraints implied in assembly relationships onto joint relationships based on degree of freedom analysis, is presented to construct the physical constraint models for physically simulation in virtual assembly environment. A variable stiffness spring model, which maps the geometrical displacement onto assembly force, is devised to realize the interactive input of assembly force. The methods are implemented in the development of VIRDAS (Virtual Reality Design and Assembly System) and the primary process of virtual assembly based on physically modeling in VIRDAS is also given in this paper.2430|247|0更新时间：2024-05-07
- 摘要：LBG algorithm is classical algorithm in Vector Quantization, which was proposed in 1980 by Y. Linde, A. Buzo, and R. M. Gray. Used this algorithm, an acceptable performance codebook can be get in the acceptable time, it has better performance than Scale Quantization that was proved by Shannon. But few outlines vector of the training image effect the distribution of the codeword in the codebook. To few vectors, it maybe noises, but have large codeword number in the codebook. And decrease the body vectors' codeword in the codebook. That decreases the compress ratio and makes the reconstruct image worse, the advantageous of the Vector Quantization can't be explained adequately. Different people used this algorithm with different image can get different compression ratio. Design the Vector Quantizer based on robust statistic can improve it. Decrease the outlines vectors, improve the center vector effect in the codebook, it can decrease the relativity of the codebook, made the distribution of the codeword of the codebook more economical and bring on the compression ratio. Theoretical analysis and simulation experimental results presented in the paper show that this method can obtain good reconstruction image quality and high compress ratio. It is improved in both subjective and objective.12411|176|0更新时间：2024-05-07
- 摘要：The instability of network channels will lead to transmission errors which deteriorate the quality of restored image, sometimes even make compression algorithms invalid. In this paper, a new image compression method is porposed to reduce such severity. Firstly a wavelet transform of the image is taken to obtain wavelet subbands of the image. According to different characteristics of each subband, different compression methods are applied to it. DPCM(Differential pulse code modulation) is employed to lowest frequency subband and multiple description scalar quantizer(MDSQ) to high frequency subband. Because coefficients match Laplacian distribution after classification based on their acitivity prediction, a context based classification and adaptive quantizer(CBCAQ) is used to them. Because there are different correlations in different subband, different compression methods are used to encode high frequency and low frequency coefficients. And during encoding process, Laplacian distribution characteristics of subband coefficients are fully exploited. The results of experiments show that the proposed method of image compression performs well in reducing transmission errors.2266|161|0更新时间：2024-05-07
- 摘要：Efficient encoding algorithm is the key factor for image to store and transmit. In order to make it more convient and efficient for image to store and transmit, several kinds of methods have been practised in the past. Among these methods, wavelet transform alogrithm and fractal encoding alogrithm are the two main methods for image processing and have recently received considerable attention. In this paper, with analyzing the merit and disadvantage of basic fractal encoding algorithm, a new fractal encoding method based on wavelet transform has been proposed. Through the two algorithm, a digital image is firstly decomposed into four subband images with two scale wavelet transform, one of them is high frequency part and the others are low ones, then according to the likeness of structure between different level subband image, we apply the basic fractal coding to the decomposed high level subband image and construct the low level fractal codes with the high ones. Satisfactory and effective results have been obtained by analyzing the course of the experiments with the two related alogrithm, especially in shortening the encoding time and improving the compression rate.2304|231|0更新时间：2024-05-07
- 摘要：Every pixel in the super space is required by K means algorithm to calculate Euclidean distance for clustering. When there are much many class centers, this is a much hard work. In this paper, an improved K means clustering algorithm is presented to accelerate clustering process with more and more classes becoming stable by judging with neighbor centers nearest to the pixel. The inter spectral redundancy and intra spectral redundancy can be eliminated mostly by multilevel clustering algorithm with quickly convergent K means classification and the method clearing redundancy at step through enhancing the intra class pixel redundancy. The multi level clustering process with initial S+P (Sequential transform + Prediction) integer wavelet transformation can not only remove the spatial and structural redundancy, but also delete the residual data redundancy realizing the breakthrough of lossless compression for multi spectral images. Furthermore, the comparison with other lossless compression algorithm and the parameter analysis of the TM (Landsat Thematic Mapper) images show that this multilevel clustering compression algorithm is more reasonable and efficient.关键词：Image Processing;Multispectal images;lossless compression;integer wavelet transform;Clustering3044|191|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