最新刊期
卷 8 ， 期 9 ， 2003
- 摘要：Owing to the widespread use of digital images, methods for efficient image access, management and retrieval become urgent requirements to image users and managers. Content-based image retrieval (CBIR) is a good solution for this problem, and has attracted increasing attention from researchers all over the world recently. In this paper, we review the beginning and various applications of CBIR firstly, and then we introduce some key techniques and algorithms for CBIR, such as methods and principals for image feature selection and representation, feature-based similarity computation, semantic features and relevance feedback. CBIR inherits some automatic techniques from traditional computer vision. However, CBIR and computer vision are very different in essence. We put forward our views on both common and distinct characters between them in the concluding section. CBIR is distinguished by the ability of on-line learning through interactive with users. Future research directions are also be presented in this section, including relevance feedback, features fusion, database technique and hierarchical ordered descriptions about the semantic content of images.2466|176|0更新时间：2024-05-07
- 摘要：Solutions with advanced multimedia functionality are becoming increasingly important as individuals are producing more and more digital media, not only for professional use but also for their personal use. With the increasing use of network and multimedia technologies, all kinds of compression standards of multimedia information are widely concerned. In order to push forward the research and development of compression technologies in China, a summary and general overview on matters of MPEG compression standards for coded representation of digital audio and video is provided. At first, all kinds of existed MPEG international standards MPEG-1, MPEG-2, MPEG-4, MPEG-7 and their basic applications are reviewed in this paper. Then, the new\r\n\r\n\r\n\r\ncompression standard element-MPEG-21 multimedia framework is emphasized and discussed about its target,\r\n\r\n\r\n\r\nuser's need, technical elements, key issues, main characteristics, application domain and future work plan.2914|267|0更新时间：2024-05-07
- 摘要：3D object recognition, which requires recognition and localization of the 3D object from 2D images, is one of the main research fields of computer vision. It is well known that the appearance of an object varies with the viewpoint and the intrinsic parameters of a camera, which makes the recognition more difficult. By geometric invariance refer to the unchanged property of the shape of an object under special space transformation. Because projective invariance between 3D object and its 2D image could overcome the problem caused by viewpoint, calibration and feature correspondence effectively, it has been attracting more and more attention and research efforts over the last 20 years. For more comprehension to this field, firstly, the main research contents of the invariance-based 3D object recognition is discussed, including geometry frames and their invariance, as well as the way to apply the geometric invariance. Secondly, a commentarial review of this field is given to show the last evolution of it. Although many kinds of useful projective invariance have been derived so far, the sorts of 3D object to which they could be used are still very limited. So there are some needs to develop more powerful invariance for the 3D object recognition. Finally, in the last part of this paper, the future researches in this field are proposed.3150|164|0更新时间：2024-05-07
-
Automatic Illuminations Detection and Color Correction of Image Using Chromatic Histogram Characters
摘要：Color is used as most an important cue in those studies such as skin detection, face detection and recognition, image and digital video retrieval. But under changing illumination, objects may show different color as the original one it is, especially the lights run out of the range of the camera sensors. In such worse condition, many algorithms become useless. In order to solve this problem, many researchers introduced additional known knowledge about the objects or environment and so have their use limitation. In this paper, an algorithm's architecture is proposed to automatically detect illumination and correct the color of image without use of any additional knowledge and any assumed ideal condition, but the image itself. By analysing the chromatic histogram characters of uncorrected color image in RGB channels, it first detect which channel cause color uncorrected and then use histogram characters above to adjust channel's intension until it reach balance between each channel. A large of different uncorrected color images is used to verify the efficiency of our algorithm, including computer simulating images, face images under different illuminants and images in art and films. The result shows that our algorithm works well in worse condition, and can correct and restore the image to its original color in most circumstance.3915|221|0更新时间：2024-05-07 - 摘要：The bi-directional cross track scanning, attitude fluctuation and the ETM sensor itself of Landsat7 result in the geometrical distortions including underscan, overscan and misalignment, which make geometric calibration more difficult than other kinds of model. The complicated imaging model increase the difficulty of geometric calibration as well. After the geometric model of imaging is studied, and the factors influencing imaging are analyzed, the systematically geometric calibration algorithms are derived, including generation and projection of line of sight, combination of attitude, generation of correction grid, finding corresponding input grid cell for a given output space, resampling, and so on. The model of line of sight are built in detail. Especially, the gap interpolation and detector delay are considered, and simplified methods are introduced to settle them. Finally, three parameters are introduced to evaluate the result of geometric calibration, including location precision, distortion and correlation between two lines. Experimental results show that the algorithms lead to good location precision, small distortion and good correlation between lines which all reach the requirement of the systematically geometric calibration, so validate the algorithms.2809|208|0更新时间：2024-05-07
- 摘要：Image denoising plays an important role in various image-related applications. While serials of wavelet-based denoising schemes fit well to images with Gaussian (white) noise, few of them can handle images with various non-Gaussian noises effectively. This paper deals with the problem from the data mining approach. It treats noisy pixels in an image as isolating outliers that are discernible in color attributes from their neighbor pixels. Inspired by the idea from outlier detection analysis, it first maps the pixels of the image into a metric space and then introduces a distance among the pixels. By making use of the density function on the pixel data set, it formulates an analytical definition of the noisy point. Further, the paper discusses the properties of the non-noisy points and constructs a denoising algorithm. Results of experiments and real world applications show that this novel approach is effective both to Gaussian and non-Gaussian noise. The method can be implemented for mass image denoising with satisfactory efficiency and denoising quality.关键词：Computer image processing;Image and graphics;Denoising;Outlier detection2512|156|0更新时间：2024-05-07
- 摘要：The deteriorated image resulting from imaging system performances can be recovered with inverse filtering by multiplying the inverse filtering function and the Fourier transform of the acquired image providing that the optical imaging transfer function is known and its inverse function or the corresponding i inverse filtering function can be found according to the measure or the priori knowledge of the imaging transferl function. If the inverse function is continuous about the origin it may actually be represented as the Tailor series. The inverse Fourier transform operation of the polynomial series is differentiation of orders corresponding no degrees in the polynomial. Consequently, the inverse Fourier transform of the recovered image is approximately realized in spatial domain by the linear combinations of the image and its derivatives rather than by complicated deoonvolution. It is considerable for the recovery of the spatially variant deterioration, such as the deterioration resulting from curvature of field. For images of space-variant degradation, the linear combining coefficients rare functions of spatial coordinates specified by the spatial variance. Detailed analysis and derivation of equations are presented. Finally, the processing results both in spatial invariant and variant systems are given.2283|205|0更新时间：2024-05-07
- 摘要：The segmentation of lung tumor serial images is one of the key techniques of Computer Lung Tumor Three-dimensional Aided Diagnosis System. The complex relation between tumor and its adjacent tissue makes it difficult to get good result. For providing doctors more accuate lung image, an automatic segmentation of lung tumor in CT serial images is presented based on texture analysis and radial basis function(RBF) neural network. With the correlation of tumor's gray level and position in sequential slices, we got the training swatch of tumor region automatically. Some second-order statistical texture parameters were computed for composing feature space; a classification procedure based on RBF neural network was applied to this space to segment the tumor. Compared with region growth algorithm and the multi-criterion segmentation algorithm, the experiment demonstrates that the proposed method can make full use of three-dimensional information of tumor serial images, and reduce manual intervention as possible. The segment results also confirm the validity and the clinical value of the proposed method.关键词：Computer image processing;Lung tumor segmentation;texture analysis;Radial basis function neural network2692|194|0更新时间：2024-05-07
- 摘要：The traditional color-based image retrieval method is based on color-histogram method, which makes combine the color information with other information difficultly. To resolve this problem, a new method using the color cluster table is presented in this paper. A new method using the color cluster table is presented in this paper. First, cluster the image, then construct cluster table using the clustering information, at last, realize the image retrieval by this table. Here, definition of the cluster table is presented, and the method to achieve is discussed. Then based on this definition, a new color-based image retrieval method is put forward. With this method, the retrieval can be realized by cluster result, making it convenient to combine the color information with other information. In order to prove the efficiency of the method, it is applied to the automated warehouse field to realize the recognition of the material. And the results compared with those achieved by the traditional color-histogram method, which shows that satisfying results can be achieved by this method.2395|234|0更新时间：2024-05-07
- 摘要：When training sets with uneven class sizes are used, the classification result based on support vector machine (SVM) is undesirably biased towards the class with more samples in the training set. That is to say, the larger the sample size, the smaller the classification error, whereas the smaller the sample size, the larger the classification error. This paper proposes weighted support vector machine algorithms based on the analysis of the cause of such problem, and this algorithm overcomes the drawback which standard support vector machine algorithm can't deal with each sample flexibly and compensates for the unfavorable impact caused by this bias. Such weighted support vector machines improve classification accuracy for class with small size at the cost of accuracy reduction for large size class, and can be applied to the case of regarding small sort of classification accuracy, such as fault diagnosis. The result of outdoor image recognition shows the effectiveness of this algorithm.2576|200|0更新时间：2024-05-07
- 摘要：This paper describes a new method of the driver behavior surveillance by using a camera mounted on the test car. Recently, one of the important issues to heavy traffic accidence in the world is that many drivers steer tiredly or are absent-minded when steering car. This situation usually decreases the attention of drivers. Hence, traffic accidence occurs. Thus it is very important to surveille the drivers' behavior, especially the eyes' state recognition to drivers. As a matter of fact, investigation statistics indicates the degree of the drivers eyes' opening is different to some extent on the various occasions of the driver normal driving, the drowsing driving and the fatigued driving. So when the car runs, we obtain two-dimensional image through a on-board camera, from which we can get the eyes' state data. Based on this point, we have developed an effective algorithm to count the frequency of blinkling. This algorithm uses Gabor wavelet to pick up the texture character of the canthus. And the eigenvector that composed of eigenvlaue is the input of the ANN. We also have performed outdoor experiments to make sure that our algorithm works effective.2415|179|0更新时间：2024-05-07
- 摘要：An accelerated algorithm of volume rendering with Phong shading based on textures mapping and blending is presented.In the light of phong illumination model an unit spherical surface under the same shading environment as the actual volume rendering is employed to emulate the real reflectional light of per-voxel，and a look一up table of voxels’refleetional lights indexed in their normal N is made，on which the illumination of per-voxel may be quiekly ascertained. Volumetric Data window transform modulated by voxel's gradients is used to compute per一voxel opacitya. The dataset composed of all voxels’illuminations(RGB) and opacitiesais projected from object space in to view space by textures mapping，and is blended from back to front along view direction into ZD image with 3D vision effect.Matrix transform of degree rotation of 3D object is used in observing indifferent viewing directions，and a fast equivalent method of shading eorreetion about it is brought forward to improve its interaction speed in volume rendering proeess.The excel lene eof data proeessings of tware and texture mapping hardware of volume rendering is eolligated in the algorithm.It is experimented by using ZD textures mapping and blending in general personal computers. High quality images may be produced at nearly interactive speeds.The algorithm may be also applieable for volume rendering based on 3D texture mapping.3826|216|0更新时间：2024-05-07
- 摘要：RS and Chi-square analyses are powerful steganalytic techniques for detecting the presence of secret information embedded in the least-significant-bit(LSB) plane of digital images. This paper describes an improved LSB steganographic scheme that can effectively resist both RS and Chi-square analyses without sacrificing perceptual quality of the cover image. In the described steganographic technique, a pair of mutually complementary mappings, F1 and F-1, is used, leading to a balanced behavior of several statistical parameters explored by the two steganalytic approaches, therefore improved security. In other words, not only can the flipping between 2i and 2i + 1 be used, but also changes from 2i to 2i-1 or from 2i-1 to 2i are allowable. Although modification to the host pixels may affect the higher bit planes, the induced distortion to the host image is not increased compared to the simple LSB replacement technique. Extraction of the embedded information can still be accomplished by extracting the LSB plane as in the straight LSB method. The proposed method has low computation complexity and is easy to implement. Analysis of the effectiveness and experimental results are presented.2424|186|0更新时间：2024-05-07
- 摘要：Virtual reality technique organizes all kinds of information with a new mode in computer, and give a virtual information space of human nature. Usually, the virtual scene is denoted by panoramas, it has three kinds: spherical panorama, cylinder panorama and cubic panorama. Firstly this paper discusses the environment map briefly. From our point of view, the theory of environment map is a footstone to building virtual environment. Then it presents the basic and speedup arithmetic of browsing the cubic panorama , and introduces an anti-aliased arithmetic of the environment image to eliminate the distortion of the image. At last, the paper implements a cubic panorama browser. The browser has functions as follow: roaming in any direction in the virtual environment and the display speed is in direct ratio to the speed of the mouse; cruising the virtual environment with zooming in or zooming out. The browser also provides anti-aliased functions in following situation: real-time anti-aliased function, actionless anti-aliased function and no anti-aliased function. The browser can make user get the biggest freedom and give a kind of convenient and natural interactive mode.3220|192|0更新时间：2024-05-07
- 摘要：A simple, fast collision detection method is presented, which is based on possible collision set (PCS). The algorithm comprises two main parts: the construction of the possible collision set and the computation of the soonest possible time of collision. First a track bounding box, whose edges are parallel to the global coordinate axes is computed for each rigid body in the environment. Then a spatial tiling technique is employed to obtain a possible collision set P of current frame by searching which bounding boxes overlap. After that, a soonest possible time of collision tmin is estimated for every pair of rigid bodies in the P based on Lin-Canny closest features algorithm. Thus the collision check will only be applied to the element that has the minimum value of tmin. Compared with other methods, this one can reduce times of collision check dramatically in collision intensive environment and real time or near real time collision detection can be achieved easily. In addition, the PCS method assures integrality and uniqueness of collision detection in the environment. Some simulation tests are also designed ?based on the method. The correctness and effectiveness are proved in both theory and practice.2574|184|0更新时间：2024-05-07
- 摘要：Data quality influences directly on correctness and reliability of application, analyzing, decision-making of GIS, at present there is not much research on error of 3-D spatial data. This paper adopts spatial parallelogram to estimate uncertainty of arbitrary points inside and outside the spatial triangle plane, The estimate formulae of uncertainty of spatial triangle plane and normal vector are deducted. It indicates that spatial triangle plane constructed by three spatial points being not in the same beeline, the precision of arbitrary points inside its interior is higher than that of one of the three points at least. The precision of points on the margin of triangle plane is higher than that one of the two extreme points of the beeline at least. The precision of points outside the spatial triangle plane is distinctively lower. Triaxiality radius and direction of error ellipsoid of arbitrary points in spatial triangle plane is relative to variance and covariance, and to its position of inside this plane.2892|174|0更新时间：2024-05-07
- 摘要：MODIS (Moderate Resolution Imaging Spectroradiometer) is a kind of new weather satellite data. Few weather satellite images obtained are all clear sky and they are always influenced by cloud more or less. Cloud is a large obstacle to remote sensing image processing and analysis all the while. In order to extract objective information more effective, cloud should be removed from the remote sensing images, which is an essential sector in the image preprocessing. Cloud detection is the most important processing before removing cloud. Taking it into account that MODIS data includes thirty-six bands, especially the infrared channels subdivided, it has realized cloud detection in MODIS images by multi-spectral synthesis method, infrared difference algorithm and cloud detection index in this paper. Owing to the limitation to a certainty of the above methods, an automatic cloud detection algorithm is applied based on the spatial texture analysis and neural network in this research. At last the cloud detection results gained by different ways are testified each other and analyzed by comparison. It found that the results are consistent, which shows that the cloud-contaminate pixels are detected successfully. It not only lays a good foundation for the cloud removing, but also can improve the precision of remote sensing image recognition, classification and inverse in this study.关键词：Moderate Resolution Imaging Spectroradiometer (MODIS) data;Multi-spectral synthesis method;Index algorithm;Automatic cloud detection4901|237|0更新时间：2024-05-07
- 摘要：Compression algorithm for screen images, which comes from the Computer Screen, is an important task in the Screen Share technology. Traditional methods are based on the algorithms such as RLE, LZW or JPEG directly, which cannot meet the requirements of both the quality of recovery image and compression ratio to compress screen image respectively. High compression ratio algorithms (JPEG) result in a low quality of image, while no-distortion algorithm (RLE, LZW) leads to a low compression ratio. In order to resolve this contradiction, an effective compression algorithm based on dynamic threshold sub-blocking is proposed in this paper. Before the image compression, Screen image is divided into four sorts of blocks by certain provided rules first, which are named as Pure-color Blocks, Text Blocks, Icon Blocks, and Image Blocks. Then these blocks are compressed separately depended on their characteristic so as to ensure that the quality of recovery image is the overall best and the compression ratio is the highest. Pure-color Block, Text Block, and Icon Block are compressed with no-distortion compression algorithm, and Image Block is compressed with JPEG algorithm. Result of experiments shows that our algorithm has improved quality of recovery image, higher compression ratio, and higher compressed speed than traditional algorithm.2797|190|0更新时间：2024-05-07
- 摘要：The good spatial-frequencial localization characteristics of wavelet transform make the energy of image congregate mostly in lower frequency subimage. This paper utilizes the property of the asymmetrical energy distribution on different multiresolution levels, and introduces a method which is called multiresolution levels compression (MLC) to process the image on the levels of different multiresolution. By the estimate and studying of the quantization coefficients during wavelet transform on every decomposed level, the experiment can get the best quantization coefficient for any particular level. Theoretical analysis and experimental results demonstrate that a method of higher compression rate is implemented. In the experiment, the best quantization coefficient of every level is determined by the data analysis of human. For the aim to process huge amount of images such as the images from the Internet, the technology of artificial intelligence and machine learning can be used to let the computer use MLC method by studying a certain number of image examples and then process more images automatically. The ability to use the MLC method would also improve according to the amount of images the computer has studied. This is the new direction to perform in the near future.2531|211|0更新时间：2024-05-07
- 摘要：In multimedia application systems , recently emerging the need of compressing and transmiting a special video in Internet, which has the characters of full screen, half motion and non-continuous tone, for example, the content of PowerPoint files. In order to satisfied this need, this paper proposed an efficient scheme to acheive the high compressed ratio at the same time protecting the nearly lossless visual quality. This scheme is named "sharing screen" and includes the two main techniques, one is called copy-screen, and the other is the data-compression method called LZ77. The first step of this scheme is to copy the full and high resolution screen contents and save them as bitmap formats. The second step is to reduce the color resolution of these bitmap images, transforming the 24bit color definition into 4 or 8 bit color definition so that we can reduced the quantity of image data. The last step is to apply the lossless method to compress the low color resolution images. As to the non-continuous tone color images, the compressed ratio can reach up to 100:1, and it has real time performance, moreover, the decode process is fairly simple. So it can be easily used in network multimedia applications such as Distance Education and Remote Stock Comment & Analysis System.3406|190|0更新时间：2024-05-07
- 摘要：Neural network is a very efficient method for image compression. It is suited to the problem of image compression due to its massively parallel and distributed architecture. Principle component analysis (PCA) neural network model and self-organizing feature map (SOFM) neural network model are often adopted for image compression in many references. In this paper,the authors propose an image compressing algorithm based on PCA/SOFM hybrid neural network, which has the advantages of both PCA and SOFM. A new method of selecting initial codebook and distortion criterion is presented to improve the efficiency of SOFM neural network according to the statistical feature of PCA transformational coefficient. Simulation results show that compared to successive PCA and SOFM algorithm or basic SOFM algorithm, PCA/SOFM hybrid algorithm has many advantages: lower memory storage; the substantial reduction of computation and the better performance of codebook.2502|159|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