First of all a simple and practical rectangular transform is given,and then thevector quantization technique which is rapidly developing recently is introduced.We combinethe rectangular transform with vector quantizat...First of all a simple and practical rectangular transform is given,and then thevector quantization technique which is rapidly developing recently is introduced.We combinethe rectangular transform with vector quantization technique for image data compression.Thecombination cuts down the dimensions of vector coding.The size of the codebook can reasonablybe reduced.This method can reduce the computation complexity and pick up the vector codingprocess.Experiments using image processing system show that this method is very effective inthe field of image data compression.展开更多
The stress concentration problem in structures with a circular or elliptic hole can be investigated by analytical methods. For the problem with a rectangular hole, only approximate results are derived. This paper dedu...The stress concentration problem in structures with a circular or elliptic hole can be investigated by analytical methods. For the problem with a rectangular hole, only approximate results are derived. This paper deduces the analytical solutions to the stress concentration problem in plates with a rectangular hole under biaxial tensions. By using the U-transformation technique and the finite element method, the analytical displacement solutions of the finite element equations are derived in the series form. Therefore, the stress concentration can then be discussed easily and conveniently. For plate problem the bilinear rectangular element with four nodes is taken as an example to demonstrate the applicability of the proposed method. The stress concentration factors for various ratios of height to width of the hole are obtained.展开更多
A fast Cholesky decomposition and a fast inverse Cholesky decomposition method for A T A are presented,where A is an m×n rectangular Toeplitz block matrix,we give the FCD algorithm for computing...A fast Cholesky decomposition and a fast inverse Cholesky decomposition method for A T A are presented,where A is an m×n rectangular Toeplitz block matrix,we give the FCD algorithm for computing R , and the FICD algorithm for computing R -1 ,both allow for an efficient parallel implementation,for solving a least squares problem and require only O(mn) operations.展开更多
文摘First of all a simple and practical rectangular transform is given,and then thevector quantization technique which is rapidly developing recently is introduced.We combinethe rectangular transform with vector quantization technique for image data compression.Thecombination cuts down the dimensions of vector coding.The size of the codebook can reasonablybe reduced.This method can reduce the computation complexity and pick up the vector codingprocess.Experiments using image processing system show that this method is very effective inthe field of image data compression.
基金supported by the National Natural Science Foundation of China (No.10772202)the Chinese PostdoctoralScience Foundation (No.20060400757).
文摘The stress concentration problem in structures with a circular or elliptic hole can be investigated by analytical methods. For the problem with a rectangular hole, only approximate results are derived. This paper deduces the analytical solutions to the stress concentration problem in plates with a rectangular hole under biaxial tensions. By using the U-transformation technique and the finite element method, the analytical displacement solutions of the finite element equations are derived in the series form. Therefore, the stress concentration can then be discussed easily and conveniently. For plate problem the bilinear rectangular element with four nodes is taken as an example to demonstrate the applicability of the proposed method. The stress concentration factors for various ratios of height to width of the hole are obtained.
文摘A fast Cholesky decomposition and a fast inverse Cholesky decomposition method for A T A are presented,where A is an m×n rectangular Toeplitz block matrix,we give the FCD algorithm for computing R , and the FICD algorithm for computing R -1 ,both allow for an efficient parallel implementation,for solving a least squares problem and require only O(mn) operations.
