最新刊期
卷 7 ， 期 1 ， 2002
- 摘要：It is well known that Foley Sammon transform is a very effective method for feature extraction and that face identification is a typical small sample size problem. In the problems, the dimension of original features is so high that the within class scatter matrix is always singular. In the singulas case, how to calculate Foley Sammon optimal discriminant vectors(FSDV) is a very difficult problem. This paper presents a fast algorithm for calculating FSDV for small sample size problems. The main idea of the proposed algorithm is to map the problem of calculating FSDV in the original feature space to another problem of calculating FSDV in a( c-1 ) dimensional (or less) Euclidean space(where c is the number of pattern classes). In the transformed space, the FSDV can be calculated directly. Generally speaking, the number of classes is much less than the dimension of original sample, so our approach needs less time for feature extraction. We do Experiments on Olivetti Research Laboratory(ORL) face database. The experimental results show that our approach is better than previous method such as Perturbation and Complementary Space in terms of discrimination power and computing time, and also superior to classical "Eigenfaces" and "Fisherfaces" method with respect to recognition rate.4289|479|0更新时间：2024-05-07
- 摘要：For the speed of detecting human faces in single grayscale scene imgae is slowly, an approach of knowledge based fast face detection is proposed. In the method a tri part mosaic image model is supplied, which is consistent with physical structure features of human face. Basing on this model, knowledge base, including information of edge and gray in images, is established by analyzing fairly enough human face images. Moreover, in order to enhance the speed of detecting face, hierarchical detection processes is adopted. In the process of rough face detection, a new coarse detection method--extended projection which is according with the physical structure features of eyes is presented and applied, which makes it possible to get different scale face candidate region, thus reduces the computation time of face detection greatly. Our experimental results indicate that the approach is high robust, and it is fit to solve the problem of multi face detection in complex background. Because of the algorithm's simpleness, it can be easily achieved by hardware, and this will shorten the time consuming on face detection. This made it possible to detect faces in real time. So it has wide application perspective at fields of visual telephone, intelligent monitor etc.4967|673|0更新时间：2024-05-07
- 摘要：This paper presents an improved real-time head tracking method, which can realize robust tracking of a person's translation and turning around with a black/white camera. This method consists of two main steps: block feature based head tracking, and head geometry based correction. The block feature based tracking only uses low-level image information and does not rely on the models of different objects, thus it can be used to track the free motion of the head. The correction step is introduced to address the shift caused by errors accumulating of block tracking. This step measures the displacements of the head contour in the tracking window of current image, and hereby to correct the results of block tracking. Moreover, to improve the performance of block tracking during person's turning around, we introduce a cylindrical model to approximate head, and extract and track features in the warped cylindrical surface. The resulting system provides robust and precise tracking over long sequence on a 350 megahertz microcomputer, and operates at 30 frames per second with the tracking window size of 120×180 pixels and the tracking sets of 80 features.4951|512|0更新时间：2024-05-07
-
Estimation of Human Motion Based on A Relative Deformation Model and the Technique of Regularization
摘要：In order to improve the creditableness and robustness of the motion and structure parameters of brachium and leg which are estimated based on dynamic monocular image sequence of human motion, this paper proposes an approach to estimate the human motion using a relative deformation model and the technique of regularization. Based on the motion description mode under the object centered coordinate system, this paper proposes a non rigid motion model by incorporating deformation parameters into rigid motion model using the notion of relative deformation. In addition, based on this non rigid motion model, a new approach of the estimation of motion and structure parameters is proposed based on the technique of regularization. In the technique of regularization, by using the prior knowledge of human motion, the robust solution of motion estimation can be obtained. Results of experiments show that the method of this paper is effective for reflecting the pattern of human non rigid motion, and the motion pattern can also be reflected to some extent by the relative elasticity parameters of the non rigid motion model.2527|398|0更新时间：2024-05-07 - 摘要：Segmenting character regions in an image is very important because these characters contain clear clues of retrieving and browsing images from video/image databases efficiently and effectively. In this paper, We propose a method to locate character regions of video/image in DCT compressed domain directly. With the distinguishing characteristics of character's texture (such as horizontal lines, vertical lines, or slant lines in a character) that can be extracted directly in DCT compressed domain, the character regions are segmented from their backgrounds quickly, and the image noises rising during the processing period can be removed by morphological filter. With this method, the compressed bit streams, which are encoded by DCT based encoding algorithm such as JPEG, MPEG 1/2, etc., can be processed directly to locate the character regions in image, just a very small amount of decoding is required (Huffman decoding only). So, the amount of data which want to process is smaller, the processing speed is faster and the demand of computer memory is less. The experimental results show that the correct localization rate of this algorithm is higher.2701|410|0更新时间：2024-05-07
- 摘要：Virtual reality is now becoming matured for serious engineering applications. This paper describes VDVAS, an integrated multi modal virtual design and virtual assembly environment, within which the designers are enabled to create both the solid model of a part and the assembly model of a product in 3D space by direct 3D manipulations and voice commands intuitively and easily, to perform assembly planning interactively in 3D space, and to simulate and check assembly process. One important feature of VDVAS lies in that it allows designers to modify components of an assembly(if any design flaw is found) during in the process of assembly modeling and simulation without the need of time consuming data exchange between virtual environment and other CAD applications. This is, of course, achieved by integrating virtual assembly with virtual design. Experiments show that designers can perform assembly planning and verification in VDVAS as if they were, to some extent, in the real physical world.3387|468|0更新时间：2024-05-07
- 摘要：Technique of computer graphics and virtual reality, applied to medical endoscopy system, come up to a new virtual medical image application: virtual endoscopy. In order to apply virtual reality technique into medical image processing, make convenient for surgeons to access virtual and mini invasive surgery, this paper, presents a full interactive virtual endoscopy system framework, analyzing and discussing system architecture and algorithm models. We explain some key procedures in virtual endoscopy system such as definition of world coordinate, virtual camera, surface detection, normal approximation, volume rendering and surface rendering, and also some interactive control method. And also we put forwards some new methods such as object cache method and extended volume growing method to solve the problems encountered in system, such as real time volume rendering and accurate and smooth image result. Then we applied it in medical image processing, following satisfied result. This improves that these methods are useful and effective. Our system provides an effective tool for medical image visualization, successfully fulfilled the aim of real time displaying and good result in virtual realty and visualization application. It has certain practical meanings in algorithms and system analysis.3494|402|0更新时间：2024-05-07
- 摘要：In the compression of some particular kinds of image sources, such as human face, vector quantization should naturally be considered to exploit its statistical properties. In this paper, a new vector quantization method in the wavelet transform domain is proposed. We use tree structure to organize coefficients. In each tree, nodes are pruned or retained according to their importance. The survived nodes are serialized to be a vector, which will be the input of vector quantization; a map indicating the positions of these nodes is also stored, which is to be used in decoding. It embeds together the principles of EZW and SPIHT, exploiting fully both the parent children dependencies and brother dependencies. In our algorithm's framework, SFQ's tree pruning algorithm can also be embedded to increase the PSNR of reconstructed images, though we simply choose the threshold pruning method to reduce complexity of algorithm. Using the proposed algorithm, the reconstructed facial image in very low bit rate(about 0 08bpp) is superior to that of EZW, SPIHT, SFQ in both perception and PSNR. The algorithm is very suitable for the compression of particular kinds of image sources.2387|434|0更新时间：2024-05-07
- 摘要：To meet the requirements of the three dimensional(3D) GIS community, it is necessary to study sophisticated 3D data structures, which exerts very important influence on GIS function and efficiency. An efficient and compact raster data structure in 3D GIS called octree is presented. Arbitrary 3D objects can be represented to any specified resolution in a hierarchical 8 ary tree structure or octree. To save memory, E octants are not registered. A linear octree data structure that lists only octree codes of F Octree is used. This paper focuses on discussing the compression technique of linear octree. According to the peculiarity of morton code, Counting sort algorithm occupying linear time is applied to sort code. Sorting time is faster 21 times than quick sort with n =7. An effective way of compacting sorted octants having the same attribute is given, and temporal and spatial analysis is described. Experiments on PC demonstrates that the memory storage for octree code using proposed algorithm is considerably low in comparison to others(octree/raster=4 32% with n =9). The study also indicates that octree is an ideal data structure for representing the sophisticated geographic spatial information. Program "3D Octree" has been written to develop, verify and demonstrate the octree encoding algorithms.4162|505|0更新时间：2024-05-07
- 摘要：Volume rendering is a very effectived method to help user to find problems inside a very complicated structure. Compared with surface based visualization methods, it keeps more detail information in the final 2D image. This advantage just brings up another big problem. Usually some important information is obscured or overwhelmed by the data standing closer to the observer. To solve this problem, two algorithms are put forward in this paper. Cutting the structure with an arbitrary convex polyhedron to bring forward the focused part. By using the hardware supported stencil function in OpenGL, the cutting operation become very fast and easy. Another method is the definition of multiple transfer functions. The integration of these two methods makes the user's inspection more flexible. To achieve nearly real time response, several optimization methods are taken in this paper. The rendering time in each slice is decreased to the minimum due to the combination of stencil reference values by using binary tree. The integration of the stencil operations between the front and back faces of polyhedron speed the algorithm greatly by saving lots of complicate normal computations.2968|480|0更新时间：2024-05-07
- 摘要：In the field of synthetic information forecast for mineral product, extremum lines (curves) of the first order differential coefficient of aeromagnetic and gravitational maps reflect the ruptures of magnetism and the densities of geologic bodies respectively. In order to deduce the rupture conformation, the similarities of attitude and size of geologic bodies in different deepness, curves of aeromagnetic and gravitational maps in different deepness and directions must be compared; only those ruptures of reflection with similar curves are reliable. In order to deduce and analyze the situation automatically, in this paper, we first introduced the conventional estimation approach of spatial relation between curves. Later on, we also described the Neural Networks approach of reasoning of spatial relation between curves. The result indicate that these estimations of spatial relations between curves obtained by the Neural Networks approaches matched perfectly with the estimations of authoritative geological experts, and achieved triumphs in practical applications.2644|407|0更新时间：2024-05-07
- 摘要：Because the environment of outdoor robot is very complicated, the visual navigation should be intelligent and robust enough. In order to improve the performance of computer vision navigation, a real time road scene understanding algorithm based on uncertain knowledge is presented in this paper. In this algorithm uncertain knowledge reasoning is applied to fuse the various image information and application knowledge, which make the algorithm can recognize the road scene robustly, even with the noise such as strong shadow and water on the road. Color information is used to obtain a coarse road region at first. And then image edge information is used to provide the precise road edge, which can meet the accuracy need of the visual navigation task. And this algorithm is well designed for real time application, which ensures that the visual navigation can make decisions in time. This algorithm has been applied on a real robot which has been tested in the real road environment for half a year. The experiment results are satisfying.2760|377|0更新时间：2024-05-07
- 摘要：The vehicle speed is one of the most important parameters in intelligent transportation system. In this paper, a new algorithm for displacement field estimation is present to get vehicle speed. This method can detect the big displacement vectors between frames. In order to alleviate the over smooth problem between the motion object and non motion object, a discontinuity preserving function, Huber function, is introduced instead of globe smoothness constraint term. We first create an energy function, we can get based on the image gray. The energy function consists of two terms:one is matching error;the other is a smoothness constraint. Minimizing this energy function, we can get the displacement field between frames. Using Euler's equation and partial derivative equation, the displacement field could be calculated. The displacements of corners are computed firstly,and then the displacement vectors are smoothed along the tangential direction of edge. The vehicle speed could be getting through calculating displacement field, time and the distance between camera and vehicle, Experiment shows, using this method result is good.2458|410|0更新时间：2024-05-07
- 摘要：Since condylar movement includes both translatory and rotatory components, each condylar reference point moves in different path. Therefore, the choice of condylar reference point is important for the interpretation of condylar movement traces. Kinematic center was originally suggested by Kohno. It's like a translating rotational center of the condyle disk complex in condylar movements. Its movement traces are mainly determined by the contour of the articular eminence, and independent of additional rotatory component of condylar movements. To date, commercial comdylar traces of kinematic center recording system if not available in market. There are rare explorations in this field. Therefore, the aim of this study was to develop condylar traces of kinematic center recording system. A condylar movement traces of kinematic center recording system has been developed with the combinations of computer graphic, mathematics and oral physiology. The movement traces of any point on the mandible can be reconstructed by this system. Furthmore, the kinematic center of condylar can be located and its traces can be analyzed by this system.2980|434|0更新时间：2024-05-07
- 摘要：In this paper the problems of image processing in welding process have been studied, for example, how to detect the edge position of welding seam and to ascertain the size and shape of welding pool and so on. The traditional method of image processing is sensitive to noise and other disturbances, so it is difficult to process images accurately and timely. In order to overcome this deficiency, microcosmic and macrocosmic algorithms of image processing in welding were compared in this paper, then a novel composite image processing method using fractal theory was put forward. Based on the characteristics of images of welding process, regions were divided up according to fractal characters of image edge, and fuzzy regions including image edges were detected, then image edges were recognized using Laplace operator and curved by LSM (least square method). Experiment results show that the image edge of welding seam or welding pool can be recognized correctly and quickly. The new developed image processing method could provide reliable information for the real time control of welding process.2304|522|0更新时间：2024-05-07
- 摘要：Automatic extraction of regional water resource information from digital elevation data can promote the efficiency of regional water resource prospecting and assessment in GIS. Therefore, a practical method for automatic extraction of natural river system from DEM is introduced in this paper. The method, first, needs to pre process DEM data in order to transform depressions and flat areas in the DEM into extended parts of the corresponding slopes. And then the flow direction of every grid point in the DEM is determined according to the theory of which water always flow along the steepest down slope direction. Third, the upstream catchment's area of every grid point can be calculated on the basis of the data of grid point flow direction. Forth, giving a threshlod value of upstream catchment's area picks out the grid points in river lines. Finally, the river system is pursued from DEM according to the flow direction data of grid points in river lines. The method has been applied to extracting the natural river system automatically from the DEM of a testing area. The natural river system extracted automatically with the method is basically consistent with that drew manually based on the corresponding topographical map. This experimental result has proved that the automatic extracting method is effective.4310|455|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