### Table 3 Fingerprint consistency measurements

"... In PAGE 9: ... Using the same data set from which the performance measures were originally obtained, the standard deviation of the MSD values for the fingerprint ensemble from a given radio was computed. The results of this analysis are presented in Table3 . Intuitively, one would expect that if the fingerprints produced by each of the radios were in fact consistent, then the variance of the MSD values should be relatively small.... ..."

### Table 7: Summary of results for fingerprinting of all four standard- cell placement instances. Original lines refer to the initial solu- tions S0. All other lines refer to fingerprinted solutions Si, i gt; 0. Manhattan distance is again expressed in 106 microns.

1999

"... In PAGE 6: ... Again, there is a problem-space metaheuris- tic effect in that the fingerprinted solutions are typically of higher quality than the original solution. A summary of results for all four test cases is given in Table7 . From this table we can see that we can reduce the time to generate the next fingerprinted solution while maintaining the quality as well as producing a unique solution.... ..."

Cited by 14

### Table 7: Summary of results for fingerprinting of all four standard- cell placement instances. Original lines refer to the initial solu- tions S0. All other lines refer to fingerprinted solutions Si, i #3E 0. Manhattan distance is again expressed in 106 microns.

"... In PAGE 6: ... Again, there is a problem-space metaheuris- tic effect in that the fingerprinted solutions are typically of higher quality than the original solution. A summary of results for all four test cases is given in Table7 . From this table we can see that we can reduce the time to generate the next fingerprinted solution while maintaining the quality as well as producing a unique solution.... ..."

### Table 7: Summary of results for fingerprinting of all four standard- cell placement instances. Original lines refer to the initial solu- tions S0. All other lines refer to fingerprinted solutions Si, i #3E 0. Manhattan distance is again expressed in 106 microns.

"... In PAGE 6: ... Again, there is a problem-space metaheuris- tic effect in that the fingerprinted solutions are typically of higher quality than the original solution. A summary of results for all four test cases is given in Table7 . From this table we can see that we can reduce the time to generate the next fingerprinted solution while maintaining the quality as well as producing a unique solution.... ..."

### Table 4. We again see a tradeoff between the quality of the solution and the credibility of the fingerprints. When we recolor more ISs, we can provide more convincing fingerprints, but it will take more time to color the graph. In any case, the savings over the original from-scratch runtimes is still significant.

1999

"... In PAGE 5: ... Table4 : Results for coloring the DIMACS challenge graph with iterative fingerprinting. 5.... ..."

Cited by 14

### Table 4. We again see a tradeoff between the quality of the solution and the credibility of the fingerprints. When we recolor more ISs, we can provide more convincing fingerprints, but it will take more time to color the graph. In any case, the savings over the original from-scratch runtimes is still significant.

"... In PAGE 5: ... Table4 : Results for coloring the DIMACS challenge graph with iterative fingerprinting. 5.... ..."

### Table 4. We again see a tradeoff between the quality of the solution and the credibility of the fingerprints. When we recolor more ISs, we can provide more convincing fingerprints, but it will take more time to color the graph. In any case, the savings over the original from-scratch runtimes is still significant.

"... In PAGE 5: ... Table4 : Results for coloring the DIMACS challenge graph with iterative fingerprinting. 5.... ..."

### Table 3: Number of undetermined variables (Var.), average dis- tance from original solution (Distance), and average CPU time (in 1=100ths of a second) for fingerprinting SAT benchmarks.

"... In PAGE 5: ... The distance of two solutions S = fs1;s2;:::;sng and T = ft1;t2;:::;tng is defined as: dist#28S;T #29= n i=1 jsi ,tij. Table3 reports the results when we maintain 20%, 30%, and 50% of the seed S0 solution. From the last two rows, we can see 9Thus, most entries in the table are non-integer.... ..."

### Table 1: Coding results for standard images with texture - Barbara and Fingerprints.

"... In PAGE 25: ... For all the experiments, we used the factorised [8] 9-7 biorthogonal filters [2] for efficiently computing the wavelet packet transform. Results for the performance of both variations of the CZQ coder - that is, with the wavelet basis (CZQ-Wave) and with the wavelet packet basis (CZQ-WP) - and state-of-the-art SPIHT coder for both the test images (Barbara and Fingerprints) are presented in Table1 . The measure used to describe the performance of each coder is the peak signal to noise ratio (PSNR) given by C8 CBC6CA BP BEBC D0D3CVBDBC BEBHBH CZC1 A0 C1CSCZBE where C1 and C1CS denote the original and the decoded images respectively.... ..."

### Table 1. Effect on bit strength of fingerprint identification system of decreasing input image size. Image Size % of

"... In PAGE 4: ... We also observed that the number of minutia that we were able to extract from the fingerprint images decreased sharply as the size of the input image was decreased. From Table1 , we observe that at 40% of the original image size we obtained bit strengths of 0, 3.... ..."