### lable/observable either by a direct scheme, or a register-file based

### Table 2: Receptive fields, observables, unwanted transformations and invariants. Physical variable Receptive field(s) Observable(s) Directly measurable Unwanted transfor-

"... In PAGE 5: ... 3.3 Invariants and Relation to Observ- ables Table2 provides an incremental overview of the invariants. 4 Experiments We aim at robust image segmentation and retrieval, invariant to all of the unwanted transformations sum- marized in Table 2.... In PAGE 5: ...3 Invariants and Relation to Observ- ables Table 2 provides an incremental overview of the invariants. 4 Experiments We aim at robust image segmentation and retrieval, invariant to all of the unwanted transformations sum- marized in Table2 . Spectral information is extracted from ^ E and the invariants W and W .... ..."

### Table 1. Fit of the models in o and (u; v) space on a validation set which has an almost uniform distribution in wind speed and direction and 8800 observations.

2001

"... In PAGE 14: ....2. Distance to Model Cone - Validation in o Space In order to obtain quantitative results on the t of the models in o space we have looked at the distance to the cone for a validation set which had an almost uniform distribution in wind speed and direction. The results can be seen in Table1 where it should be noted that for this retrieval only the Jacobians of all the models shown were used to determine the exact minimum distance to the cone using a scaled conjugate gradient minimisation algorithm. As expected Cmod4 ts the observations in o space very poorly if we assume the NWP wind vectors are correct and determine the corresponding location on the model manifold.... In PAGE 15: ... Mlpmod exhibits strong dependency of the mean and standard deviation on wind direction, suggesting the model is rather poor in representing the directional dependency of the o signal. Finally, Table1 also shows the VRMSE of the retrieved winds. This shows that, on this data, the neural network models are more accurate than Cmod4 in wind retrieval terms.... ..."

Cited by 1

### Table 6. Performance of the models on the test set with a typical atmospheric distribution in wind speed and direction and 14500 observations in the range 4{24 ms?1.

2001

"... In PAGE 17: ... The performance as a function of incidence angle is shown in Table 5, and again, the neural network models are more consistent, particularly at low incidence angles (that is on tracks near the satellite sub-point). Table6 shows the results on a test set that was randomly sampled from the atmospheric distribution of wind speeds, and thus re ects the likely performance on unseen data. It can be seen that training the neural networks to optimise their performance on almost uniform wind speed distributions reduces their performance on another distribution.... ..."

Cited by 1

### Table 31.2. Mean Erron of Auditory and Tactile Directional Judgments for Two Conditions of Observation and Three Egocentric Distances of the Sound Source

"... In PAGE 3: ... Although not central to an understanding of the whole of tactual perception and performance, they do constitute an interesting set of facts that require explanation. Katz (1936; cited in Krueger, 1970,1982) and, before him, Weber (184611978) emphasized that most of our perceptual ex- perience is of objects and events external to US rather than of Table31 .l.... ..."

### Table 3: Transformation of observation vectors to a common coordinate system direction position constraints

1997

Cited by 5

### Table 10: The average error for flve observers in x direction, ex, y direction, ey, error magnitude, kek, standard deviation, Std and displacement, Disp, for twenty markers for case 2. All values are in pixels.

2004

"... In PAGE 63: ... This data can be used for validation of nonrigid image registration algorithms. In Table10 , we listed the average error for the flve observers in x direction, ex, y direction, ey, and error magnitude, kek, over the twenty markers for case 2. The errors are difierences between the means over the flve observers and the output of the automated method (which for this purpose can be considered \true quot; data since its accuracy is known).... In PAGE 63: ... All values are in pixels. Table 11 was generated using the same statistical analysis as in Table10 , but only the mean and maximum value over the twenty markers for cases 1 to 6 are given. In all cases, the maximal displacements are between 11 and 22 pixels.... ..."

### Table 4.2: Ideal spherical coordinates of the relative position vector in the Local Horizontal Reference Frame a118a121a119 : Spatial distances, horizontal directions, vertical directions Station Observed Distances Horizontal Vertical

2002

"... In PAGE 61: ... Grafarend and P. Lohse (1991, Table4 , p.108).... ..."

### Table 1: Newton method for minimization of the perimeter. In Table 2 we consider the same problem but with evolutive displacement directions. In that case we observe a superlinear rate of convergence.

"... In PAGE 24: ... The initial point will be given by a piecewise perturbation of a circle of radius 1+ . In Table1 we present the results considering the domains with 128 nodes and using xed displacement directions. We follow the evolution of jjG00jj2 - the euclidean norm of the gradient, jj k ? ?jj1 - the distance between the approximation k and the exact solution ? (which is a circle), jj k ? ~ ?jj1 - the distance between the k and ~ ? (the solution in the set of polygons with 128 nodes) the area and di erence of perimeters.... ..."

### Table 1: Single-observation positional error in the scan direction V Single-measurement error Single-pass error

"... In PAGE 4: ... The total positional error in each observation|which is taken to include a set of consecutive scans in which the star is present|will then be the combination of the metrology and the photon error, divided by p10. Table1 summa- rizes the total single-observation error expected as a func- tion of apparent magnitude for both standard (250 pm) and good (50 pm) metrology. The error in the one-year standard points depends on the number of observations per year, which is a function of the ecliptic latitude: stars at high ecliptic latitude are observed more often.... In PAGE 4: ... Therefore we estimate the error in the one-year normal points to be 0.8 times the error listed in Table1 , probably a conservative estimate especially at high ecliptic latitude. 4.... ..."