### Table 1: Empirical results of the number of rectangulations for non-separable permutations

2000

Cited by 1

### Table 1: The Rana and Schwefel base functions. No- tice the Rana function is non-separable.

### Table 1: The Rana and Schwefel base functions. No- tice the Rana function is non-separable.

### Table 2: Line Search Performance on Nonlinear Non- separable Functions.

1995

"... In PAGE 3: ... This leaves only F2, F8, F9 and F10 among the prob- lems that are both nonseparable and nonlinear. Prob- lems F2, F9 and F10 are not solved by a single pass of line search and multiple passes do not yield compet- itive solutions as illustrated in Table2 . All of these problems are solved by the CHC algorithm using dra- matically fewer evaluations (Eshelman, 1991; Mathias and Whitley 1994).... ..."

Cited by 45

### Table 1: A non separable BBA with its commonality and weight functions.

"... In PAGE 2: ... Example 1 Let = fa; b; cg be a frame of discern- ment. The commonality and weight functions for a BBA m on are shown in Table1 . We can see that m is not separable, since we have w(fbg) gt; 1.... ..."

### Table 5.6: Numerical values of the normalized tracer di usion coe cient for the non-separable surface potential system.

### Table 1: Parameters of a separable and a non-separable mapping. HDm: Hypercube Dimension, GCM: Separable Mapping, based on binaryreflectedGraycodes(23 23 ! 8 8), FGM:Non-separable, mapping based on two folded grids (26 ! 64).

"... In PAGE 3: ... HDm: Hypercube Dimension, GCM: Separable Mapping, based on binaryreflectedGraycodes(23 23 ! 8 8), FGM:Non-separable, mapping based on two folded grids (26 ! 64). In Table1 GCM, the Gray code mapping, is obtained by Gray coded processor indices for two ring mappings. The hypercube dimensions alternate in both mesh dimensions.... ..."

### TABLE IV DIFFERENT RESULTS OF SEPARABLE AND NON SEPARABLE LIFTING SCHEME FOR IMAGES OF THE JPEG2000 DATABASE

### Table 1. Classi cation accuracy (percent) with 1:0% false positives for varying message sizes (the maximum message size for EzStego and LSB is 194 194). Classi cation is from an earlier Fisher Linear discriminant (FLD) analysis [6] and a linear non-separable SVM, Section 3.2. As expected, the classi cation accuracy is comparable. See also Fig. 7.

2002

"... In PAGE 10: ... 6 The SVM parameters were chosen to yield a 1:0% false-positive rate. The trained SVM is then used to classify all of the remaining previously unseen steg images of the same format, Table1 . In this table, the columns correspond to separate classi cation results for JPEG, GIF and TIFF format images.... In PAGE 10: ... In this table results are shown for a 1:0% and 0:0% false positive rate. Note the signi cant improvement over the linear classi ers of Table1 . Shown in Fig.... In PAGE 12: ...3. Note the signi cant improvement in accuracy as compared to a linear classi er, Table1 .... ..."

Cited by 41