最新刊期
卷 9 ， 期 3 ， 2004
- 摘要：Recently, the problem of packet loss and errors in image communication over erasure channels, such as Internet and wireless networks, becomes an important topic. This creates the need for new approaches to image coding, which combine high compression efficiency and robustness. As a new approach, multiple description coding (MDC) is attractive in this field, which separates the source into several equally important bit streams(descriptions) and then transmitted over different channels. Based on many related literatures, an overview on the state of the art of MDC is given in this paper. Firstly, MDC's history and background in information theory are introduced. And then MDC is classified into several categories, including quantizion based, transform based, space expansion based and so on. Each is described respectively from the mathematic model and its applied conditions. Comparisons between MDSQ and MDPCT are made followed. At the end, some conclusions are given, including problems in MDC and what will MDC be in the future.2925|194|0更新时间：2024-05-08
- 摘要：The restoration of rotational motion blurred images with the boundaries of non zero gray values is a study hotshot in the field of image processing. In order to quickly fetch the gray values of pixels along the blurring paths in the case of discreteness, the idea of Bresenham algorithm is incorporated into the process of restoration of rotational motion blurred image. The restoration algorithm based on deconvolution along the blurring path by mean of filtering in frequency domain is proposed. Meanwhile, in order to effectively restore the information of pixels in the boundary area, the less cost restoration algorithm based on knowledge guide is also proposed?Thus, the restoration of the blurred image with non zero boundary consists of two stages, in which the two proposed restoration algorithms are employed. On the other hand, the iterative calculation and complicated time consuming operations are avoided to speed up the image restoration. The effect of the proposed algorithms is validated by comparison experiments and a lot of test experiments in the microcomputer, and experimental results illustrate that the proposed algorithms are efficient and capable of resisting noise.2877|224|0更新时间：2024-05-08
- 摘要：The requirement of the images registration of pathologic slices is more demanding than others in quality, but some of the present methods of images registration will produce unmatched phenomena. In order to solve this problem, a new stitching algorithm of images based on relativity is presented here. In this algorithm, the character of the image is got from the two parallel columns (or lines) in the overlap area of images, which will be stitched up each other. To avoid the accumulation of the"disturbance", the maximal envelop is taken as the element of template ,which is the strongest character in the curve of difference data. And according to the relativity of the image content, the method of images registration will be proceeded by using the template which is made by the distribution of the maximal convex envelop, then the effect of the "disturbance" will be reduced, thus the method of images registration mentioned here is more steady than some former similar methods. Experiments demonstrate that this algorithm not only is more regular and stable, but also has a faster calculating speed. It's a more practical method of images registration.3349|212|0更新时间：2024-05-08
- 摘要：Histogram shows the statistic relation between every gray level and the probability of its occurrence, giving the global information of the image. Gray scale transformation based on the technique of histogram modification is one of practical and effective ways of image enhancement. Histogram processing includes equalization and specification. The aim of equalization is to create an image with equally distributed brightness levels over the whole brightness scale; The aim of specification is to transform the histogram of an original image to a specified histogram capable of highlighting certain gray level ranges in the image. In order to improve the efficiency of the algorithm of histogram processing further, the shortcomings of several current algorithms of equalization transformation and specification transformation are first analyzed, then a new binary mapping law (BML) is proposed. That law is based on optimal control theory making gray scale mapping abide by the principle of the minimum error of histogram, and haves a simple algorithm. Whether for histogram equalization or for histogram specification, it is testified more efficient than other mapping laws.3572|253|0更新时间：2024-05-08
- 摘要：In order to realize fast and high-quality rotation transform for image , this paper introduces the approach of decomposition of a two-dimensional rotation into three-step one-dimensional shift by decomposing the traditional matrix of image rotation, which succeeds in simplifying the interpolation of image in one-dimensional space. In order to ensure high quality of rotated image, the paper uses the cubic B-spline interpolation to calculate the gray of pixels after each image shift. What is more, the paper develops an IIR(infinite impulse response)-based and a FIR(finite impulse response)-based digital filter to realize fast cubic B-spline interpolation, which turns out to be a significant advantage for hardware implementation and parallelization. And it designs a FPGA(field programmable gate array)-based fast and high quality image rotation system for the 256×256 pixels images in 256 gray levels. By the means of coordinates transforming, the calculation time and storage will not increase evidently when the images rotate at a large angle in the system. In a word, the image rotation system is able to meet the request of real-time and high-quality rotation of image at any angle.4100|209|0更新时间：2024-05-08
- 摘要：Text is an important feature for computer vision, especially for information retrieval applications. In this paper, the authors have developed a novel algorithm for text background separation, or binarization for color images of complicated backgrounds. In their algorithm, dimensionality reduction and graph theoretical clustering are first performed. Corresponding to each cluster, a binary image can be obtained. Additional binary images are obtained through combination among these cluster related binary images. Then, two kinds of features capable of effectively characterizing binary texture images, run length histogram based and spatial size distribution based features associated with each of these binary images are extracted out. Based on the analysis of these texture features, cooperating with an LDA classifier, the optimal binary image which gives the best text background separation will be found out as the final binarization result. Experiments with images collected from Internet have been carried out, which show that their method can handle color text images with complex background effectively; comparison with existing techniques also presented a notable improvement brought by the proposed method.2537|223|0更新时间：2024-05-08
- 摘要：Image thinning is one of time consuming processes in image processing since it often needs several scanning iterations of the whole image data. Research efforts have been made on this issue to try to reduce scanning iterations for thinning algorithms. Image template thinning algorithms are parallel ones that have been widely applied in image processing with better performance than serial thinning methods. However, for some commonly used template thinning algorithms there are some shortages such as still needing quite a few iterations, incomplete thinning somewhere, etc. In order to overcome these shortages, an improved image template thinning algorithm is proposed in the paper based on a group of modified templates. The new algorithm can significantly reduce the scanning iterations and speed up the thinning process while giving the same or better (complete) thinning result. The new algorithm has been applied to image thinning of fingerprint images and the validity of the algorithm is confirmed by the experiment results in the paper.3340|290|0更新时间：2024-05-08
- 摘要：Enlightened the image compression technique based on 2 D Differential Pulse Code Modulation(DPCM), and taken the correlation between pixel and its neighbors into consideration, a new image quality evaluation method is discussed by adding up the gray errors between every pixel of image and its predictions in the paper. The basic principles are that each pixel of image is first predicted by using linear prediction operator and least square estimate technique and then the mean square root of the difference between pixels' gray and their predictions are added up. This difference reflects the relativity of pixels each other. The larger the difference is, the smaller the relativity of pixels is, and the better the contrast and the definition of image are. The experiment results show that the discussed method is first more sensitive to the change of image quality than traditional mean square error(MSE) method, average gradient method and information entropy method, and then more suitable to the quality evaluation of the images processed by the different denoise methods removing noise. The shortcoming of this method is but that its efficiency is slightly inferior to that of the traditional MSE and information entropy.2767|218|0更新时间：2024-05-08
- 摘要：Object recognition is a challenging problem in the field of pattern recognition and computer vision. Hu's moments are classical tool in the field, which are defined based on the colors or gray levels of objects. This paper is an improvement of Hu's moments. A series of novel moments, which are called contour moments, are constructed based on object contours and their applications to object shape recognition are given in this paper. Some properties of these new moments including the invariance on shift, rotation and scale transforms are studied and proved. A central advantage of the new moments over Hu's moments is that they are independent of the colors or gray levels of objects. They are defined completely by the contours of objects, namely, that they are completely the shape features of objects. To support our new theory, an algorithm for object shape recognition is designed based on the new moments and experiments are conducted. In our experiments, wavelet transforms are employed to extract the contours of objects, therefore, a brief introduction on the theory of wavelet transform as a multi scale edge detector is introduced. Considering that an object may have more than one contour, each of which is a close curve, this paper also gives detailed discussion on how to deal with several contours. Experiments give an encouraging high recognition rates.3541|234|0更新时间：2024-05-08
- 摘要：Aimed to locating all the instances of multiple templates in one image, a fast and more effective multiple templates locating algorithms based on clustering and synthesizing of templates is proposed. This algorithm can process those multiple templates even if only some of which are similar to each other. But all the templates must be almost the same size. First, a hierarchical clustering algorithm with feedback is applied to cluster the templates into some categories. In each category a mathematical model is applied to synthesize the templates in it. And thus a mother template is constructed. Second, the mother template of each category is used to search and matching in the translation space. And then the matched mother template is guided to check all the son templates. Edge maps are extracted for clustering, synthesizing, and matching. The partial Hausdorff distance matching with fast algorithm is suggested for mother template searching and matching procedure. Our algorithm is tested with difference multiple templates in integrated circuit micro images database. The results show that the scheme is efficient and effective for the task of multiple templates matching and locating.2851|200|0更新时间：2024-05-08
- 摘要：Surface inspection for quality control is the most important part of today's highly competitive industrial production. There are a great number of manufacturing processes where inspection for surface finishing or surface defects is attempted: steel strip, hot steel slabs, plastic plates, painted surfaces, wooden surfaces, profiles and etc. But the surface inspection is usually a bottleneck in many production processes. Reliable and effective detection of machined surface defects is of great importance in automatic industry that needs high surface aesthetics and machined precision, it can obviously enhance the entire efficiency for the automatic manufacturing. A novel image filter has been developed in this paper, it transforms the original image via FFT and then filters out those Fourier areas which Fourier energy concentrates on. These Fourier areas correspond to the background texture on machined directions. The method can enhance the texture defects and reduce the effect of background texture by analyzing the texture features of machined surfaces. Then the texture defects can be easily detected through the segmentation of defects and background. The experimental results have shown that the method can obtain precise results and high detection speed.5335|451|0更新时间：2024-05-08
- 摘要：The wear condition of the artillery pipe is decisive to ballistic performance of the artillery and its detection is important to the using, maintenance and life evaluation. There are two types wear on the inner surface, the one is called friction wear, which is caused by the friction between the bullet and pipe's inner surface, making the rifling edge became smoother and smoother, the another is called ablation wear, which is caused by the high temperature and pressure gas produced by the fire powder, resulting cavities and flaws on the inner surface. These two types wear exhibit different appearance and have different evaluation criterions, it's necessary to divide these two types wear and give a respective detection. In the image of the inner surface of artillery pipe, the property which tell apart the edge of the rifling and the edge of the flaw and cavity is its direction: the edge of the rifling has a certain direction and shows a character of anisotropy but the edge of the flaw and cavity has no certain direction and shows a character of general isotropy. So the degree of the direction can be employed to divide the two types edge and measure the two types wear. Base on this, the mathematic modes of the friction wear and scrape wear are established, the direction image operator and its relative algorithm is proposed and proved, which can divide and detect the ablation wear and scrape wear from the image, realizing the image detection of the artillery body pipe's wear degree.2915|204|0更新时间：2024-05-08
- 摘要：Retrieval and management of the vast quantitative image data needs efficient approaches of content-based image retrieval. Current image retrieval methods in geographic image databases use only one kind of image features, which can not describe image content completely. In this paper, a remote sensing image retrieval approach using color and texture fused features is presented. A given image is decomposed using quin-tree, and each subimage except the center one is separated into 5 sublevel subimages until the size of the subimage equals to or is great than 16×16 pixels. The energy values of the image are calculated via multi-channel Gabor filter, and the mean values and standard deviations of each subimage are extracted as texture features. The color features are calculated too. Similarity between a given query image and database subimages, which sizes are approximately equal to the former, are measured using linear weighted distance of color and texture features. Then the top similar subimages are returned as query results. This approach is applied to retrieve high resolution remote sensing images from database, and its efficiency is confirmed by the experiments. The comparison of precision and recall between texture-based and color -texture-based image retrieval shows that synthesized features describe image content better than lonely one does.2760|198|0更新时间：2024-05-08
- 摘要：Different from the former focus on the neighbor region analysis of spatial objects, the purpose of the study is to locate and analyze the competitive region of weighted objects in the Euclidian R 2 space. A new form of weighted Voronoi diagrams named projective weighted Voronoi diagram (PWVD) is introduced to resolve this problem. Based on the model of spatial tessellation of ordinary voronoi diagram and weighted Voronoi diagram, a projective spatial tessellate model is presented, which can produce competitive region in planar space. Thus the interspaces produced by the tessellation i.e. competitive triangles (CT), are examined. Next, three weighted methods: proportional, gravitational and Gaussian are applied to produce different CT respectively. The properties of the triangles and PWVD, such as the inclusion and overlap of CT and the intrusion, inter intrusion and neighbor region sharing of PWVD and CT are analyzed too. At last section, the creation algorithm of CT is discussed and the result of computer simulation shows that it is feasible to use the competitive triangles as the competitive region of the Euclidian R 2 space. From this study, it is concluded that the method might be widely used in future's GIS spatial analyses and practical applications.2965|180|0更新时间：2024-05-08
- 摘要：Digital Elevation Model(DEM) finds its widely using in the areas of Geography, landscape and regional planning. The abbreviation DEM is used to describe a digital data set which is used to model a topographic surface (a surface representing height data). Since this is impossible, a surface model approximates a continuous surface using a finite number of observations. In its simplest form a vector DEM mimics the raster version by using a regularly spaced set of spot heights to represent the terrain surface. A more advanced, more complex and more common form of vector DEM is the triangulated irregular network (TIN). In vector GIS a TIN is used to create a DEM from either regular or irregular height data. However, some systems have their different request of the data volume under different scale for the influence on the speed of operation and analysis. Besides, cartography generalization of the contour can be realized by the generalization of DEM. So the generalization model on DEM grid data is given in this paper by using the multi band wavelet analysis, and a case study is put forward to give the detail explanation. Meantime, its reliability is described from the point of data volume, profile of the elevation and change of curve surface area.3106|204|0更新时间：2024-05-08
- 摘要：A wavelet watermarking algorithm based on the excellent property of chaotic sequence and some characteristics of the human visual model is proposed in this paper. Firstly, the multilevel wavelet decomposition of digital image is involved, and then combined with the white noise statistical characteristics of chaotic sequence and the contrast sensitive function of human visual, chaotic sequence is added to the wavelet coefficients of the high pass bands from the low resolution to the high resolution which are important to the human visual. Finally, watermarked image is obtained by using reverse wavelet transform. Because the chaotic sequence is very sensitive to the initial value, the attackers cannot deduce the whole sequence from the finite sequence and the security of watermark information is ensured. At the same time, combined with the human visual model, our algorithm guarantees the visual impression and protects the low frequency part of image efficiently. Experimental results show that the proposed scheme has good robustness and imperceptibility to common information processing operations and noise attack.2620|210|0更新时间：2024-05-08
- 摘要：For multimedia data is often voluminous and requires real time operations, multimedia encryption algorithms should meet extra requirement compared with textual data encryption algorithms. Traditional algorithms such as DES, RSA, can't meet some real time requirement. Some selective encryption algorithms can't completely keep the file format and compression ratio unchanged. In order to meet such requirement, the algorithms should combined encryption process with encoding process. Due to the wide application of quad tree structure in image or video encoding, two types of quadtree permutation algorithms are proposed here. They are used in quadtree based image encoding, wavelet zerotree encoding (EZW, SPIHT), and quadtree based fractal image or video encoding. The encryption algorithms combine encryption with compression, and have some advantages. Firstly, it is secure, and the encryption result is too confused to be understood. Secondly, it keeps the file format unchanged, and thus it is easy to operate compressed and encrypted data directly. Thirdly, it keeps the compression ratio unchanged, that is to say that the encryption operations do not produce extra data. Lastly, it is fast and can meet real time requirement.2692|195|0更新时间：2024-05-08
- 摘要：The picture scrambling algorithms based on chaos are frequently used for images secure communications. In this paper, a new modified method of chaotic picture scrambling network is proposed by the chaotic mapping interval twice division and the chaotic orbits dynamic quantification. The new scrambling network adds the complexity and speeds up the computation of the network, since it overcomes the disadvantage that it need large number of computation to make the chaotic orbits ergods the whole chaotic mapping interval under the finite computational accuracy, reduces the computation. The paper also analyzed the time complexity and the scrambling performance of the new networks by computer experiments. The results show that this picture scrambling network has good scrambling capability and guarantees the security of the encrypted images effectively because it can work quickly with large cryptographic keys space.2460|199|0更新时间：2024-05-08
- 摘要：Vector Quantization (VQ) is an important technology of image compression research in the recent years. Reducing encoding computation time and cutting down average encoding bit rates are the two important problems of its current research. In the present, many fast search encoding algorithm based on VQ are proposed. In order to get less encoding time and lower average encoding bit rates, a fast encoding algorithm, which integrates advance correlation predictive process with a fast search encoding algorithm, has been presented in this paper. After the current image block has been encoded, the encoding value of near neighbor image blocks is predicted, by virtue of their correlation to the current block. If predictive successes, the encoding value needs few bits to denote, otherwise it needs complex computation by Absolute Error Inequality Elimination algorithm (AEI), and more bits to transfer. By this proposed way, the total encoding time decreases, and the total encoding bits saves. The simulation experiments show that the proposed algorithm consumes less time to encode, and needs lower average bit rates to transfer encoding results, against full search algorithm (FS), while its encoding performance is close to that of full search algorithm.2762|200|0更新时间：2024-05-08
- 摘要：According to the line based wavelet transform algorithm, the column transform is performed progressively, so lower memory is required without affecting the result of wavelet transform and it is suitable for the memory insufficient system very much. In this paper, the improved lifting scheme is used to perform wavelet transform to replace the traditional Mallat method that is used in the original algorithm. The three adder unit is adopted to realize lifting scheme. The in place calculation property of the lifting scheme makes the original data updated directly by the current computation result. So the synchronous buffer and the splitting operation are not necessary any more. At the same time, one three adder unit can be divided into two two adder units so that less memory is required. However, two two adder units need one more multiplication in contrast to one three adder unit. Thus, there is more computation in two two adder units. Furthermore, the three adder uses the same operations for low pass lines and high pass lines while the two adder needs different. This algorithm has been tested and compared with JPEG2000 in items of PSNR and bit rate. Experimental results show that it is memory efficient and excellent in compression performance.2771|186|0更新时间：2024-05-08
- 摘要：How to provide quality Streaming Media service with limited bandwidth is an important issue worthy of discussing and it has become the hotspot of streaming media research. As a popular codec method usually used now, VBR (Variable Bit Rate) often causes the streaming media's coding rate to alter remarkably and results in abrupt data transportation, which will affect the service quality significantly. To handle this problem, many flow control techniques have been brought forward. But on the other hand, the conventional technique of flow controlling requires feedbacks from clients and the bandwidth adjusting lags far behind the actual data transportation so that it is unsuitable for the VBR streaming media service based on UDP (User Datagram Protocol). At the same time, it is yet not necessary to support fastidious real time performance for some kinds of streaming media service, like video service. In this paper, we present an active wave filtering and stream stabilizing model based on the feature of data transportation velocity. Because it does not require feedbacks from clients, it not only reduces the abrupt data transportation caused by VBR but also provides steady streaming data transportation service by the means of filtering and stabilizing.2715|185|0更新时间：2024-05-08
- 摘要：Road image recognition is a key technique for vision navigation of outdoor mobile robot. Because vision navigation of outdoor mobile robot requires good robustness and rigid time limitation for real time autonomous driving, in order to satisfy the request of good robustness and rigid time limitation, in this paper, we applied successfully Augmented Transition Networks (ATN), which is a kind of method used for sentence parsing in the artificial intelligence field, to the road image recognition of outdoor mobile robot, and proposed a road image interpretation algorithm based on the ATN. The characteristic of the ATN is, aiming at the different road condition, the algorithm can make up of different ATN, but all the ATN are constructed by the uniform construction rules. Therefore, the construction of the ATN is very flexible in application, and the essence of the ATN is coherent yet. The algorithm has good robustness and real time performance, and has been verified by the experiment that it can meet the needs for high speed autonomous vision navigation of outdoor mobile robot.2977|196|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