CiteSeerX — Table search results

Results 1 - 10 of 25,460

Table 7. Dense-codes: Optimizing the variance of the Gaussian kernel

in From Margins to Probabilities in Multiclass Learning Problems

by Andrea Passerini, Massimiliano Pontil, Paolo Frasconi 2002

Cited by 4

Table 7. Dense-codes: Optimizing the variance of the Gaussian kernel

in From margins to probabilities in multiclass learning problems

by Andrea Passerini, Massimiliano Pontil, Paolo Frasconi 2002

Cited by 4

Table 1: Distortion rate performance of DCT, KLT, Wiener transformed KLT, and the proposed system. N = M = 8 is used, and the source coded is the Gaussian AR(3) source coded in Figure 4.

in Minimum Mean Square Error Transform Coders

by Are Hjørungnes, Tor A. Ramstad 1998

"... In PAGE 5: ... This system will be called the Wiener transformed KLT. The performance of the DCT, the KLT, the Wiener transformed KLT, and the proposed trans- form coder is shown in Table1 using N = M = 8 and the input source is the same Gaussian AR(3)... In PAGE 5: ... N = M = 8 is used, and the source coded is the Gaussian AR(3) source coded in Figure 4. From Table1 it is seen that the performance of the proposed system is better than all the other methods for all rates. The proposed system per- forms close to the Wiener transformed KLT, but is better for all rates.... ..."

Cited by 1

Table 1: Examples of possible BCH codes and the corresponding FNMR and FMR results for non- Gaussian and Gaussian assumptions.

in Privacy Enhancing Technology for a 3D-Face Recognition System

by Xuebing Zhou, Emile Kelkboom, Michiel Van Der Veen, Tom Kevenaar, Christoph Busch, Alexander Nouak

"... In PAGE 9: ... With 255 bits codewords, only discrete set of the secret code length s and the correctable errors length e is possible. Several examples and their corresponding bit error rate(BER), FNMR and FMR is given in Table1 . The FNMR under the assumption of non-Gaussian distribution is significantly better than under the assumption of Gaussian distribution,... ..."

Table 2: Compression results in bytes per number required to encode 10,000 random gaussian numbers with different coding schemes

in Optimal alphabet partitioning for semi-adaptive coding of sources with unknown sparse distributions

by Dan Chen, Yi-jen Chiang, Nasir Memon, Xiaolin Wu 2003

"... In PAGE 8: ... To show the applicability of our approach to a wider class of applications, we then generated random gaussian numbers with different variances and used the greedy heuristic, Huffman coding of bytes, Huffman coding of pairs of bytes and Huffman coding of four byte words. Table2 and Table 3 show the results obtained with a source sequence of length 10000 and 100000 respectively. As before all costs including that of storing the Huffman table are included in the results.... ..."

Cited by 7

Table 1. Coding gain at a BER equal to 10-6 for a turbo code over a Gaussian Channel [28].

in Acknowledgements

by Abdallah Said Alahmari 2003

"... In PAGE 88: ... For 2L2PPM modulation, every four signals share the same two positions and they have different amplitude level combinations. Hence, we assigned two bits to the amplitude variations and we mapped them as we map QAM Table1 . 128-SCTCM natural mapping signal labels.... ..."

Table 1: Distortion rate performance of DCT, KLT, Wiener transformed KLT, and the proposed transform coder system for N = 8, and the source coded is a Gaussian AR(3) source with PSD shown in Figure 5 (a).

in Minimum Mean Square Error Transforms and Filter Banks for Source Coding

by Are Hjørungnes, Tor A. Ramstad 1998

"... In PAGE 22: ...iener transforms. The bit allocation used is the same as for the KLT system. This system will be called the Wiener transformed KLT. The performance of the DCT, the KLT, the Wiener transformed KLT, and the proposed transform coder is shown in Table1 using N = 8 and the input source is the same Gaussian AR(3) source coded in Figure 8. From Table 1 it is seen that the performance of the proposed transform coder is better than all the other transform systems for all rates.... In PAGE 22: ... The performance of the DCT, the KLT, the Wiener transformed KLT, and the proposed transform coder is shown in Table 1 using N = 8 and the input source is the same Gaussian AR(3) source coded in Figure 8. From Table1 it is seen that the performance of the proposed transform coder is better than all the other transform systems for all rates. The Wiener transformed KLT performs close to the proposed transform.... ..."

Cited by 1

Table 4: Comparison of the execution times of the hand-written code and Fortran 90D compiler generated code for Gaussian Elimination. Matrix size is 1023x1024 and it is column distributed.(Intel iPSC/860, time in seconds). Processors Hand Written Fortran 90D Compiler

in Compiling Distribution Directives in a Fortran 90D Compiler

by Zeki Bozkus, Alok Choudhary, Geoffrey Fox, Tomasz Haupt, Sanjay Ranka 1992

"... In PAGE 19: ... 7.2 Performance Evaluation Table4 shows a comparison between the performance of the hand-written Fortran 77+MP code with that of the compiler generated code. We can observe that the performance of the compiler generated is within 10% of the hand-written code.... In PAGE 21: ...Figure 10: Speed-up comparison (corresponds to Table4 of the hand-written code and Fortran 90D compiler generated code for Gaussian Elimination.... ..."

Cited by 1

Table 4 Comparison of the execution times of the hand-written code and HPF compiler generated code for Gaussian Elimination. Matrix size is 1023x1024 and it is column distributed.(Intel iPSC/860, time in seconds).

in Compiling HPF for Distributed Memory MIMD Computers

by Zeki Bozkus, Alok Choudhary, Geoffrey Fox, Tomasz Haupt, Sanjay Ranka 1993

"... In PAGE 21: ... Due to space limitations, we do not present performance of many other programs, and some of them can be found in [?]. Table4 shows a comparison between the performance of the hand-written For- tran 77+MP code with that of the compiler generated code. We can observe that the performance of the compiler generated code is within 10% of the hand- written code.... ..."

Cited by 2

Table 4 Comparison of the execution times of the hand-written code and HPF compiler generated code for Gaussian Elimination. Matrix size is 1023x1024 and it is column distributed.(Intel iPSC/860, time in seconds).

in Impact Of Compilation Technology On Computer Architecture

by Edited By, David Lilja, Peter Bird, William B. Baringer, Barry Boes

"... In PAGE 25: ... Due to space limitations, we do not present performance of many other programs, and some of them can be found in [?]. Table4 shows a comparison between the performance of the hand-written For- tran 77+MP code with that of the compiler generated code. We can observe that the performance of the compiler generated code is within 10% of the hand- written code.... ..."

Results 1 - 10 of 25,460

Tools

Table 7. Dense-codes: Optimizing the variance of the Gaussian kernel

Table 7. Dense-codes: Optimizing the variance of the Gaussian kernel

Table 1: Distortion rate performance of DCT, KLT, Wiener transformed KLT, and the proposed system. N = M = 8 is used, and the source coded is the Gaussian AR(3) source coded in Figure 4.

Table 1: Examples of possible BCH codes and the corresponding FNMR and FMR results for non- Gaussian and Gaussian assumptions.

Table 2: Compression results in bytes per number required to encode 10,000 random gaussian numbers with different coding schemes

Table 1. Coding gain at a BER equal to 10-6 for a turbo code over a Gaussian Channel [28].

Table 1: Distortion rate performance of DCT, KLT, Wiener transformed KLT, and the proposed transform coder system for N = 8, and the source coded is a Gaussian AR(3) source with PSD shown in Figure 5 (a).

Table 4: Comparison of the execution times of the hand-written code and Fortran 90D compiler generated code for Gaussian Elimination. Matrix size is 1023x1024 and it is column distributed.(Intel iPSC/860, time in seconds). Processors Hand Written Fortran 90D Compiler

Table 4 Comparison of the execution times of the hand-written code and HPF compiler generated code for Gaussian Elimination. Matrix size is 1023x1024 and it is column distributed.(Intel iPSC/860, time in seconds).

Table 4 Comparison of the execution times of the hand-written code and HPF compiler generated code for Gaussian Elimination. Matrix size is 1023x1024 and it is column distributed.(Intel iPSC/860, time in seconds).