### Table 1: MEAN SQUARED ERROR1

"... In PAGE 5: ... If the rectification is successful, the average pixel displacement should be small. 4 Experimental Results The MSE results are shown in Table1 . The proposed rectification method shows substantial improvements in MSE compared to the original distorted image in all test cases.... ..."

### Table 3 Weighted normalised proposed policy values and corresponding square error values

in Abstract

"... In PAGE 7: ... The optimum number of cluster groups is taken as six corresponding to F-statistic value of 10 (Burn, 1989). After fixing the optimal number of clusters as six, the representative policy for each cluster is determined as shown in Table3 . For this purpose the square error values between group mean and the weighted normalised proposed policy values for each criterion in that group are calcu- lated.... ..."

### Table 4. Comparison of Objective Measures of Image Quality and Subjective Ratings for the Images Shown in Figure 24

in â€¢ BAXTER AND SEIBERT Synthetic Aperture Radar Image Coding Synthetic Aperture Radar Image Coding

"... In PAGE 31: ... frequency weighting, the Gabor system was not com- petitive. Figure 24 shows the original SAR image and three reconstructed images, and Table4 gives the nor- malized mean-square-error (NMSE) metric and the distortion contrast (DCON) metric for each image. The objective metrics reveal that the separable-wave- let-based system has the lowest NMSE and DCON values, while the Gabor-based system has the highest values.... In PAGE 31: ... The objective metrics reveal that the separable-wave- let-based system has the lowest NMSE and DCON values, while the Gabor-based system has the highest values. Table4 also summarizes the image-quality rat- ings of four subjects. For most SAR images, the sepa- rable-wavelet system tends to preserve the quality of the isolated point returns quite well, but it tends to reconstruct textures poorly.... ..."

### Table 5 Forecast Mean-Square Errors *

"... In PAGE 9: ... Similar calculations were made for horizons 6 and 12 and for the other lag selection criterion. We see from Table5 that the models with lag lengths specified by SIC, PIC, KAIC, and KSIC did slightly better than the model specified using the AIC criterion for the short-lag lag models (LM1 and LM3) as indicated by ratios which are almost always below 1. There is little difference in the results across lag selection criterion, however.... ..."

### Table 1 : Summary for homogeneous model Posterior estimates of and (with standard errors), and corresponding root mean squared error (RMSE).

1998

"... In PAGE 8: ... For each of these the human eye can make a good job of delineating the boundaries and assessing the relative intensities. Figure 2 shows the posterior means for these two examples using the model de ned in equation 5, that is with constant, but unknown, prior parameter , and Table1 shows posterior estimates of and (with standard errors), and corresponding root mean squared errors.... In PAGE 18: ... how the two texture patterns partition the image. Now compare the single parameter of the homogeneous model, Table1 , with the range of in Table 2. It is clear that though there is little change in the average parameter value, the large di erence between the minimum and maximum mean that the amount of local smoothing varies widely.... ..."

Cited by 10

### Table 6. Forecast Errors (Mean Squared Error)

2005

"... In PAGE 16: ... That is, outside the threshold, deviations from the long-run equilibrium decline more rapidly than most estimates of the PPP literature, while the convergence speed is slower or neatly equal to zero inside the threshold. Finally, comparison of forecast errors is contained in Table6 . The mean squared error for random walk hypothesis is juxtaposed with the mean squared errors based on cointegrating regressions of columns (2) and (3) of Table 2.... ..."

### Table 3: Mean square errors in RGB coordinates.

2007

"... In PAGE 18: ... Only a non local strategy seems to be able to correctly interpolate such images. Finally the mean square error table ( Table3 ) illustrates that the non-local demosaicing and the Gunturk algorithm give the best average errors even if the non-local demosaicing is... ..."

### Table 5. Average Mean Square Errors

"... In PAGE 7: ...In the mean square errors a smaller value is desirable. It can be seen from Table5 that in HIV classification the SVRGA performed the worst as it had the highest error but in the education level it performed the best as it has the lowest error. The following figure, Fig.... ..."

### Table 1 Compression Rates, Median Root Mean Square Error, and Median Maximum Error Obtained from Mammograms

2004

"... In PAGE 6: ...Compression rates ranged from 14:1 to 2051:1 ( Table1 ). Minimum rates were similar for the various types of lesions, whereas maximum rates differed sub- stantially (Table 1).... In PAGE 6: ...Compression rates ranged from 14:1 to 2051:1 (Table 1). Minimum rates were similar for the various types of lesions, whereas maximum rates differed sub- stantially ( Table1 ). Of all mammo- graphic images used in this study, 60% (302 of 500) of the images were com- pressed at a rate less than 100:1 and greater than 20:1.... ..."