### Table 2. Number of bit errors after JPEG compression (without / with edge enhance filtering). Name Label

1996

"... In PAGE 6: ... Information about the labeled test images. Name Resolution (pixels) Compression ratio using JPEG quality factor of 90% 80% 75% 60% 50% 40% Diver 302 x 323 1:9 1:14 1:16 1:23 1:27 1:30 Mountain 733 x 487 1:8 1:11 1:12 1:18 1:21 1:25 Lena 512 x 512 1:11 1:17 1:20 1:28 1:33 1:39 Kielp 720 x 576 1:7 1:10 1:12 1:16 1:18 1:21 In Table2 , Figure 2 and 3 the bit errors in the label are represented, after compressing the images with the JPEG compression algorithm, with quality parameter set to different values. Table 2.... ..."

Cited by 13

### Table 10: Results of the runs for normal queries for the official runs in the Robust track.

2004

Cited by 11

### Table 10: Sawtooth results, 9 a1 9. Summary of the results Among all the measures studied, those of the two first fam- ilies, GC (derivative-based measure) and SCC (ordinal measure) give good results. In contrast, derivative-based measures are not efficient. Ordinal measures that are efficient in occluded regions are not really efficient in non-occludedareas. Robust measures are the most efficient particularly partial correlations, PSEUDOP, MAD, LMPP, LTPP, SMPDP,

### Table 1: Table showing robust (20% outlier trim) estimates of correlation, along with signi cance test results at the 5% level from Kendall apos;s tau measure of correlation. Columns on the left are based on the raw tests, while columns on the right are based on the tests after a log transform. Est is the robust estimate of correlation with ratings, norm.z is the normal z score, and p is the p-value used for the signi cance testing.

"... In PAGE 9: ... 4.3 Correlations Table1 shows the correlation between the various tests, both with and without log transforms, and Best ratings. Because both log and non-log transformed versions of the same data are being used, clearly some variables are not normally distributed.... ..."

### TABLE I ENERGY SAVINGS OF THE LMS-BASED CORRELATION TRACKING SCHEME

2005

### Table 7: Results for the robust track.

"... In PAGE 9: ... UAmsT03RStFb Language model run on the Snowball stemmed index, using Rocchio blind feedback. Results Table7 gives the results of the five official runs over all 100 robust topics (best scores in boldface). The second Table 7: Results for the robust track.... ..."

### Table 7: Results for the robust track.

"... In PAGE 9: ... UAmsT03RStFb Language model run on the Snowball stemmed index, using Rocchio blind feedback. Results Table7 gives the results of the five official runs over all 100 robust topics (best scores in boldface). The second Table 7: Results for the robust track.... ..."

### Table 3 Normalized Correlation for the Embedded Watermark and max. of the Normalized Correlations for 100 other watermarks.

"... In PAGE 4: ... Then, the SPOMF type of detection is utilized to extract the watermark from the rendered image. The normalized correlation values are given in Table3 for different imagery camera positions and rotations. The normalized correlation for the embedded watermark is clearly higher than the maximum of the other correlation values which shows that the method detects the watermark for the mentioned cases.... ..."

### Table 1: A Comparison of the Signal to Clutter Ratio for Each Tracking System, * indicates a loss of track.

2005

Cited by 1

### Table 3 Regression equations relating satellite imagery to vegetation transects within various vegetation types at Fort McCoy, Wisconsin, USA

"... In PAGE 9: ... Correlation was improved by comparing each vegetation type with the satellite imagery separately (Table 2). Based on the improved correlations, regression equations were computed individually for each vegetation type ( Table3 ). The best-fit regression equations were used to reclassify the satellite imagery to produce a spatially distributed C factor data layer for each study area.... ..."