### Table 1: Median of squared errors (Interquartile range of squared errors) for six smoothers. The results for HAS, SS, SureShrink, and MARS are from Luo and Wahba (1997).

2000

Cited by 20

### Table 1: Median of squared errors (Interquartile range of squared errors) for six smoothers. The results for HAS, SS, SureShrink, and MARS are from Luo and Wahba (1997).

2000

Cited by 20

### Table 1 Table 1: Median of squared errors (Interquartile range of squared errors) for six smoothers. The results for HAS, SS, SureShrink, and MARS are from Luo amp; Wahba (1997).

2000

Cited by 20

### Table 3: Numbers of synonymous substitutions per synonymous site (Ks) at hydrophobic (below diagonal) and cross-linking (above diagonal) regions of ELN gene. The number of substitutions was estimated with Li-Wu-Luo method [25] in pair-wise comparisons.

2004

### Table 4: Numbers of nonsynonymous substitutions per nonsynonymous site (Ka) at hydrophobic (below diagonal) and cross-linking (above diagonal) regions of ELN gene. The number of substitutions was estimated with Li-Wu-Luo method [25] in pair-wise comparisons.

2004

### Table 1. PUoW Sets

1997

"... In PAGE 6: ... If any LUoW name matches the name attribute of the LUoW, then the LUoW can execute on this resource. Consider the three example PUoW sets given in Table1 for three resources. The left column shows various LUoWs.... ..."

Cited by 36

### Table 1. PUoW Sets

1997

"... In PAGE 6: ... If any LUoW name matches the name attribute of the LUoW, then the LUoW can execute on this resource. Consider the three example PUoW sets given in Table1 for three resources. The left column shows various LUoWs.... ..."

Cited by 36

### Table 2 Numerical results on Dec ALPHA Acknowledgements The research of Z.-Q. Luo is based on work supported by the National Sciences and Engineering Research Council of Canada under grant OPG0090391; The research of J.S. Pang is based on work supported by the National Science Foun- dation under grant CCR-9213739 and the O ce of Naval Research under grant N00014-93-1-0228; The research of D. Ralph is supported by the Australian Research Council. We are grateful to H. Jiang for providing the computational results in x5.2, and valuable discussion especially regarding stopping rule B. We also thank the School of Mathematical Science, University of New South Wales, for providing access to the Dec ALPHA used in numerical tests; and to two anonymous referees for their comments and suggestions.

"... In PAGE 21: ... We give two tables of results. Table 1 presents computation carried out on a Sparc 10 whereas Table2 is for computation on a Dec ALPHA. The di erence in the performance of PSQP on these machines is due mainly to the di erent behavior of the MATLAB quadratic programming routine qp, from the Opti- mization Toolbox, on these two machines.... ..."

### Table 5: G44CMC color difference of actual and predicted appearance, bold p-values indicate that there is 99 percent confidence that the transform performs as well as the best transform for a given data set. Lam Data Set Sharp BFD CMCCAT von Kries ROMM Prime 709RGB

"... In PAGE 12: ... Table5 : cont. Kuo amp;Luo TL84 Sharp BFD CMCCAT von Kries ROMM Prime 709RGB RMS G44CMC 3.... ..."

### Table 1: Performance comparison of di erent area thresholds on Lena image at 0.25 bpp, using 5 5 structuring element (Fig. 5d). As extremely small clusters usually do not produce discernible visual e ects, and those clusters render a higher insigni cant-to-signi cant coe cient ratio than large clusters, they are eliminated to avoid more expensive coding cost. As the area threshold increases, the number of clusters decreases which results in the reduction of the required cluster positioning information. As illustrated in Table 1 zero area threshold has the worst performance. All other area thresholds have similar performance. In the paper by Luo et al. [20], the authors also propose wavelet coe cients cluster- ing for image compression. They use clustering as a tool for quantization, i.e., wavelet coe cients are clustered together and quantized to the mean value of the given cluster. The wavelet coe cients are then coded by either using traditional runlength coding or Shapiro apos;s EZW algorithm. In SLCCA, we use clustering to register and transmit the signi cance map, i.e., clustering is our tool for data organization. The two algorithms

1999

Cited by 5