"... In PAGE 30: ... Stage 1 converts outlines into trapezoid sequences, and Stage 2 converts them into grid-restricted outlines. Table1 gives statistics that show how the algorithms perform on a fairly typical sample image.4 The table shows that Stage 2 reduced the vertex count from the Stage 1 midline approximation by a factor of 1.... In PAGE 31: ... All numbers were encoded using a scheme where small numbers require 4 to 8 bits, and roughly two additional bits are required each time the magnitude doubles. Table1 lists the number bytes for this scheme in the \outline bytes quot; row. It ranges from 62% to 123% of the size of the compressed binary image le that served as input to Stage 1.... In PAGE 31: ... A simple bitmap would have required more than a megabyte. Table1 also gives statistics that are relevant to the design of Algorithms 4.... In PAGE 31: ...rom 3.1 for Column H to 4.8 for Column E. The grid spacing and tolerance values in Table1 represent a trade-o between image quality and the compactness of the outline representation. The table shows how reducing the Stage 1 or secondary tolerance or reducing the grid spacing increases the outline byte count, and Figures 22b{d show how reducing the tolerances improves image quality.... In PAGE 31: ... Table 2 shows that the outline byte count generally compares favorably to TIFF bitmaps with CCITT-g4 compression when the Stage 1 tolerance is 1 pixel and the grid spacing and secondary tolerance is 1 2 pixel. Smaller tolerances would increase the ratio of outline bytes to bytes in the TIFF les as suggested by Table1 . Refer to Figure 23 for a full accounting of how this compression ratio depends on the complexity of the test pages with and without halftone pictures.... In PAGE 32: ...32 (a) (b) (c) (d) Figure 22: A magni ed portion of test document h047 reproduced by various methods. (a) pixel replication from the raw image; (b) from outlines generated with grid spacing and tolerances as in column C of Table1 ; (c{d) the same for columns D and H, respectively. without halftones with halftones input vertices vI lt;20k 20-60k 60-200k gt;200k lt;60k 60-200k gt;200k pages 24 227 608 8 16 84 12 output vertices =vI 0.... ..."

"... In PAGE 12: ...ndividuals in organizations or organizations as a whole. Similarly to Cooper and Zmud (1990, p. 123) information technology is defined here as any system, product or process whose underlying technology base is composed of computer or communications software or hardware. Figure 2 below maps the eighteen studies to the IT Diffusion Framework; Table1 provides a high-level summary of each study. [INSERT FIGURE 2 THEN TABLE 1 ABOUT HERE] The four subsections below use the IT Diffusion Framework as a device to structure a discussion of major results and implications arising from the eighteen studies.... In PAGE 12: ...1 Individual Adoption of Type 1 Technologies Five studies examined individual adoption or use of Type 1 technologies. The technologies included a text editor, a wordprocessing package, spreadsheet software, graphics software, personal computers and an expert system (see Table1 ). These technologies qualify as independent-use technologies since they were intended to facilitate self- contained tasks performed by individual users.... ..."

"... In PAGE 6: ...89 The features of differ- ent side-imaging OCT probes reported to date are summarized in Table 1. It is to be noted that Table1 registers the achieved param- eters for each of the side-imaging probes that have been re- ported in literature. While the numbers provide convenient points of reference, they do not represent the achievable lim- its.... ..."

"... In PAGE 44: ... The underlined elements in the matrix denote the assignments from the hungarian algorithm, which gives the mapping set Q0 = f(t1; b4); (t2; b2); (t3; b3); (t4; b1); (t5; b5)g. Table4 lists the results from the modi ed mapping experiment, where the left column lists the original mapping scheme R, the middle column lists the mapping derived from the modi ed semantic matrix T0 in (11), and the right column lists the mapping derived from the mapping algorithm. The results indicate that the mapping algorithm is in fact capable of constraining the mapping with respect to the structure.... In PAGE 45: ...In the rst row in Table4 we see that vertex t1 in the target graph (according to the relabeled target graph in Figure 22) is now mapped to vertex b4... ..."

"... In PAGE 8: ... The processing time has been evaluated for SSP cases, using the NNM algorithm, the Sticks-based algorithm and the GJPF algorithm. The results for a numerically generated image with noise added and for a real image (containing a larger number of scan lines) are presented in Table1 . As can be seen, the NNM algorithm performs 3 to 4 times faster than the comparison algorithms, evaluated in this paper.... ..."

"... In PAGE 1: ...lices with elongated voxel sizes #28e.g. voxel sizes of 1:1:20 #28mm 3 #29#29 and with only a subportion of the brain visible in the set of scans. Table1 shows the parameters of a few sample input image sets. Due to the fact that the voxels of some 3D images can be far from cubical, the algorithm described here is a pixel-based instead of voxel-based 3D image algorithm.... In PAGE 9: ... The angles of the symmetry axes chosen by a neuroradiologist #28curve with dots#29 versus the angles, #12 i = arg max#28C max i #28#12 j #29#29, computed by our algorithm. This is dataset #283#29 of Table1 . The algorithm determined angle is 2 degrees, the mean of the expert chosen angle is 1.... In PAGE 11: ...Figure 15. The mid-sagittal plane automatically extracted from the MR data set #28 Table1 , data set #284#29#29. The detected yaw is 1.... ..."

"... In PAGE 15: ... The column and detected word structures are used as the input of their algorithms. The performance evaluation re- sults of our statistical optimization algorithm are reported in Table5 . Since our performance evaluation is very strict, the result is counted as a correct detection only when the detected result is totally matched the ground truth data.... ..."

"... In PAGE 12: ... Therefore, the LLNL collocated samples were analyzed for VOCs only using EPA Method 524. Table1 shows the results of the collocated samples. As can be seen, VOCs were not detected in either set of samples.... ..."

"... In PAGE 3: ... The algorithm associates descriptive features with the images in the gallery and probe sets and determines the identities of the probe images by comparing their associated features with those features associated with gallery images. According to the imagery utilized in the gallery and probe sets, we can define the following nine recognition settings as in Table1 . For instance, the mStill- to-Video setting utilizes multiple still images for each individual in the gallery set and a video sequence for each individual in the probe set.... ..."

"... In PAGE 21: ... Appendix C tabulates individual sample results for TEQ, TOC, and grain size. Table3 summarizes dioxin results for agricultural soils. The table also includes summary statistics for other previously reported land uses.... In PAGE 21: ... Use of a logarithmic scale is appropriate for log-normally distributed data and spreads the data points, making them distinguishable. Based on data in Table3 , Figure 2, and Figure 3, dioxin concentrations appear to be higher in urban, and forest lands, and comparatively lower in open and agricultural lands (see also Figure 6). Although dioxin concentrations from the four land uses are qualitatively compared, this study was not designed to conduct these comparisons statistically.... ..."