### Table 1: The shortening at scale 9 of an initial circle with radius 10 is considered. For di erent values of , we give the theoretical value of the nal radius (Rtheo), the corresponding values obtained with di erent precisions (fast or slow computation), the number of performed iterations and the used CPU time.

"... In PAGE 18: ...igure 8: Evolution of circles for = 0:4; 1; 2 displayed at scale 0, 1,..., 9 (fast computation). In Table1 , we give the theoretical and computed radius for several values of and for a circle with initial radius equal to 10. We performed two sets of experiments with di erent precisions,... ..."

Cited by 1

### Table 1 Measured BERs (percentages) of the experiment using perfectly matched corresponding points. The numbers are the averaged results of 16 images.

2003

"... In PAGE 12: ... The three described distributions of corresponding points are expected to lead to increasing BERs from the first to the third distribution. The results of this experiment are listed in Table1 . Given the spatial accuracy of the used corresponding points (pixel accuracy) and the number of feature points we use in experimental situations, the polynomial order is limited to the 4th order, needing at least 15 non-collinear points.... In PAGE 12: ... It was also reported that for most image registration purposes the highest needed polynomial order is equal to four [20]. The figures presented in Table1 are the averaged results of the described experiment on 16 different images. For each of the three distributions, three situations are investigated: 1.... In PAGE 15: ... Again, in some cases a uniform distribution of feature points with pixel accuracy may lead to less improvement in BER performance than a non- uniform distribution. This confirms the results in Table1 . It is most significant when image was translated, or a PLM was used to distorted the image.... ..."

Cited by 2

### Table 2: Stereo correspondence results

"... In PAGE 4: ... Any points for which the least squares matching solution di- verges, or does not converge fast enough (a threshold of 20 iterations is currently used), are also invalidated, and removed from the set of correspondences. Results from the stereo correspondence, using the above matching strategy, are shown in Table2 . A graphical depic- tion of the nal point correspondences for the image pair is depicted in Figure 4.... ..."

### Table 1: Number of iterations required to nd an optimal solution using the previous solution as a starting point and using the vector (0; : : : ; 0) as a starting point.

2001

"... In PAGE 24: ... When a new user enters the system, running the MFVA algorithm using as a starting point the optimal solution for the problem prior to the new user apos;s arrival typically results in substantial computational savings. Table1 shows results from a power control problem involving a system of ten by ten cells and approximately nine hundred mobile users. The number of iterations required to nd the optimal solution for an initial problem is given, along with the number of iterations required to nd the optimal solution when additional users enter the system.... ..."

Cited by 10

### Table 3. The approximation error for the corresponding point noise distribution in stereo matching for three dis- tribution models

2000

"... In PAGE 6: ....e. a 5x5 window around the central point. For the training sets, we placed templates around 10 points which were obtained from the ground truth. As one can see from Table3 the Cauchy distribution had the best t to the real noise distribution relative to L1 and L2. Therefore, one expects the accuracy to be the greatest when using Lc (Table 4).... ..."

Cited by 1

### Table 1: Di erent ways to de ne a circle using incidence and tangency constraints. Case points segments circles Con guration

### Table 6: Sample results of edge measurements of circles

"... In PAGE 10: ...27 A 3-D planar circle is fitted to each group of 3-D coordinates of edge points and the radius and center coordinates were computed. Table6 shows sample results using the ASL camera. As expected, the error becomes larger as the angle increases (see results at 30o angle.... ..."

### Table 6: Sample results of edge measurements of circles

"... In PAGE 10: ...27 A 3-D planar circle is fitted to each group of 3-D coordinates of edge points and the radius and center coordinates were computed. Table6 shows sample results using the ASL camera. As expected, the error becomes larger as the angle increases (see results at 30o angle.... ..."

### Table 2. WCEC of FAST vs. Traditional

2003

"... In PAGE 9: ... 6. Results for FAST Framework The WCEC equations for the six benchmarks obtained from the static timing analysis tool and the FAST tool are compiled in Table2 and in Figure 12. The FAST scheme differs from conventional static timing analysis without parametric expressions of frequencies by less than half a percent.... In PAGE 9: ... Since the effects of scaling on WCEC are accurately modeled by the FAST equations, the scaling of the WCET can also be accurately captured. Table2 shows the WCEC for all six benchmarks calcu- lated for four different frequencies using the FAST equa- tions and compared with the corresponding WCEC ob- tained from the static timing analysis tool. Figure 12 plots the ratio of the WCET for the FAST tool and the static tim- ing analysis tool.... In PAGE 9: ... Figure 12 plots the ratio of the WCET for the FAST tool and the static tim- ing analysis tool. As shown in the Table2 and Figure 12, cnt, mm and srt show that the FAST bounds on WCET match the bounds obtained by the static timing analyzer ex- actly. For fft, adpcm and lms the FAST bounds on WCET are very close to the bounds obtained by the static timing analyzer.... ..."

Cited by 25

### Table 1 Measured BERs (percentages) of the experiment using perfectly matched corresponding points. The numbers are the averaged results of the Lena image only.

2003

"... In PAGE 3: ... On the left side of the figure, a watermark is embedded in a test image. Then, one of the distortions listed in Table1 is applied on the image. The distorted image, IW,dist, is registered using the strategy presented in Section 2.... In PAGE 4: ...The results of this experiment for a Lena image are shown in Table1 . The results can be compared with the BERs of the unregistered, thus distorted image (listed in the first column of the table).... In PAGE 4: ... 5. CONCLUSIONS Looking at the figures in Table1 , the following conclusions can be drawn: a0 Quite low BERs for some transforms The results show a large improvement of the BER for distortions that actually are a polynomial transform. For these transformation classes the achieved BERs drop below or around 5% for all cases.... ..."

Cited by 2