### Table 4: Measurement invariance results

2006

"... In PAGE 28: ... Finally, measurement invariance is assessed between male and female respondents based on the P2P non-users measurement model (model 3) with ease-of-use, perceived usefulness and perceived risk and its two subconstructs, and the construct of collection-from-near-others. Table4 presents the outcome for measurement invariance. Insert Table 4 Configural invariance is met since confirmatory analyses show that the pattern of how items load on constructs is invariant across groups (users, non-users, male/female).... ..."

### TABLE 1. Measurement of Dynamical Invariants

1994

Cited by 4

### Table 1: Invariance Characteristics

"... In PAGE 5: ...or the measurement process ( e.g. = 0:5 for Poisson processes). This is summarised for three common noise models in Table1 , but more generally, if the mapping required to represent the data with uniform errors is known then a suitable (scale invariant) form for the Bhattacharrya measure can always be constructed by the process presented here. 3 Experiments We have taken a set of eight test images, three simulated with known ground truth and five real MR images with known uniform additive noise and manually verified coil correction.... ..."

### Table 6 presents the invariance assessment results. + and - denote invariance and non-invariance respectively.

2006

"... In PAGE 4: ... Table6 : Invariance properties of the measures 5 Analysis of results The reported invariant properties divide the measures into three clusters. One cluster is constructed from measures non-invariant under the four matrix transformations.... ..."

Cited by 1

### Table 10. Empirical comparison, invariance properties and clustering of performance measures.

"... In PAGE 9: ...Table10 report measure values and the classifler ranks. Recall is omitted because of co-linearity with Sensitivity.... ..."

### Table 1. Densities of acyclic graphs on n points when partitioned w.r.t. binary parti- tion (bp) coe cients and w.r.t number of edges. 1.656. In practice, these expectations transalate into a e ective lter. In a recent application that involved testing 1,034 M random pairs of 8-point graphs for isomorphism (equality), we rst applied the edge cardinality lter; 193 M pairs passed this lter. Of these, only 148,762 had identical partition coe cients, and of these 87,710 were actually isomorphic. The probability of being isomorphic, given equal partition coe cients and numbers of edges was 1.69, compared to 1.66 as predicted by the table. A quick measure of the e ectiveness of invariant partition coe cients as an

1995

"... In PAGE 8: ... Comparing binary partition coe cients is a useful lter for eliminating obviously non-isomorphic pairs. In Table1 we display the expected number of acyclic graphs on n points that have the same identical binary partition coe cients, exp(jGj per bp). For n = 8, there exist 16,999 distinct, non-isomorphic, acyclic graphs,8 having 5,187 distinct partition coe cient sequences; so that an expected 3.... In PAGE 8: ... They frequently do not. As shown on the next line of Table1 , the expected number of graph with identical partition coe cients and the same number of edges, exp(jGj per bp and jEj), drops to 8 The number of distinct n point acyclic graphs, or posets, grows exponentially. It is known that jGnj is: 183,231 (n = 9), 2,567,284 (n = 10) and 46,794,427 (n = 11) [5].... In PAGE 10: ... Readily, the expected number of graphs passing nbr of equivalence classes (a) (b) (c) (d) (e) n jGnj coe +jEj degree ideal +jEj 4 16 15 16 16 15 16 5 63 52 61 63 52 61 6 318 208 285 125 208 284 7 2,045 962 1,570 432 951 1,551 8 16,999 5,187 10,263 1,588 4,932 9,863 Table 2. Comparison of isomorphism lters on graphs with n points any lter, as in Table1 , is the expected number of graphs per equivalence class. The similarity of (a) and (b) with (d) and (e) is striking.... ..."

Cited by 3

### Table 2. Summary of the total edge strength measures for the various color invariant sets, ordered by degree of invariance. The edge strength BXDB is not invariant to any change in imaging conditions.

2001

Cited by 64

### Table 2 provides an incremental overview of the invariants.

"... In PAGE 5: ...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 .... In PAGE 6: ...Table2 : Receptive fields, observables, unwanted transformations and invariants. Physical variable Receptive field(s) Observable(s) Directly measurable Unwanted transfor- mation(s) Invariant(s) Wavelength spectrum G ixjyk Geusebroek[8] Object cover re- flectance Reflectance Illumination intensity W Illumination intensity, shadow, shading, highlights H Illumination intensity and spectrum, shadow, shading N Geusebroek[9] Local geometry G ixjyk Koenderink[16] Object macro-shape Edges, curvature, junctions, corners Object position, pose, distance I Florack [6] Object size, surface/cover granularity maxarg ( xy)Inorm Lindeberg [21] Object distance Far distance Z Pentland [24] Object meso-shape Principal curvatures Koenderink[16] Shading w Lee [20] Object cover re- flectance type Highlights w Spatial frequency ~ Gu0;v0 i Bovik [4] Surface/cover regular- ity Spatial frequency Object position, pose, distance S Object distance Relative near distance (a) Q1 (b) Q2 (c) Q3 Figure 2: Examples of the query sets of a single object for content-based image retrieval.... ..."

### Table 13: Predicted Responses for Invariance tasks for two di erent constraints.

2004

"... In PAGE 67: ...Table13 compares the model-predicted probabilities of response for Invariance tasks in both cases. Because there were very few Non-Ratio responses overall in the study, the probability of a Non- Ratio response is quite small compared with the probability of a Non-classifiable response (see Table 4).... ..."

### Table 2. Number of invariants and perfor- mance for the letters given by Figure 5

"... In PAGE 4: ...ormation. (For details refer to [6], [7].) The second example is a scene composed of two alpha- bets (Figure 5), hence 26 different classes have to distin- guished. The quality measures q 0 and the number of invari- ants are listed in Table2 for all four invariant types. The first two columns contain the number of invariants and the quality measure q 0 for the full set of invariants, the next two columns contain the same information for the re-... ..."