"... In PAGE 9: .... The parameter VT in Eq. (22) is chosen to be 5. Experimental results on the ARPs in percentage for the FFrst 30 query images are listed in Table6 . The ARPs in percentage for all the 100 query images are grouped in Table 6 according to the 10 trademark samples.... In PAGE 9: ... It is noted that the symbol rf denotes the number of iterations of relevance feedback in these two tables. From Table6 , we can have some observations: ESOciency: In Table 6, the novel approach of Eq. (23) and the existing ratio approach of Eq.... In PAGE 9: ... It is noted that the symbol rf denotes the number of iterations of relevance feedback in these two tables. From Table 6, we can have some observations: ESOciency: In Table6 , the novel approach of Eq. (23) and the existing ratio approach of Eq.... ..."

"... In PAGE 10: ... A casual inspection of the means would lead us to anticipate positive effects arising from overseas ownership, location in Scotland and patent priority, and a negative effect of small firm size. Table2 provides more detail on the regional differences by reporting the distribution of trademark holdings per firm. Thus for Wales, 96 firms held only one trademark, 54 held two, and so on.... In PAGE 16: ... Firms in Ireland, on the other hand, are significantly less likely to participate in trademark registration than equivalent Welsh firms but, like firms in South West England, when they do participate they have significantly higher trademark stocks. The results confirm the reading of Table2 namely that, compared to the other regions, Wales suffers from a lack of high-trademark firms. However, the Ireland results are somewhat surprising in view of the apparent success of Ireland, relative to SW England, suggested in Table 1.... ..."

"... In PAGE 19: ...or the high and middle-income group is 0.88, while it takes a value of 0.84 for high and low- income economies. As Table4 below demonstrates, in all three income groups the same nine sectors ranked among the top 10 categories in terms of their share in the total number of trademarks registered. These are such R amp;D-intensive sectors as pharmaceuticals and scientific equipment and advertising-intensive industries such as clothing, footwear, detergents and food products.... ..."

"... In PAGE 19: ...7 Notes: For purposes of comparison, calculated shares only refer to 12 high-income, 16 middle-income and 7 low-income countries, for which data in all 5 years are available. There is also some similarity in terms of sectors dominated by trademarks registered by foreigners (see Table5 ). In all country groups these are pharmaceuticals, chemicals and detergents.... ..."

"... In PAGE 13: ... Results Table 3 reports results from the Poisson, ZIP and Tobit models. As there are 55 industry variables included in the analysis, their results are reported in summary form in Table4 . The results for the Poisson model are included largely for completeness but, as is implicit from the discussion given in the previous section, we regard the ZIP models as being more appropriate.... In PAGE 19: ...19 We turn finally to the influence of the industry sector variables. Rather than tabulate the full results for all 55 sectors included in the analysis, Table4 coalesces the sectors into three groups: those for which there are statistically significant positive influences on trademarking (ZIP incidence ratio greater than 1; negative marginal effect; and positive Tobit coefficient); those with statistically significant negative results; and a few for which the indications are mixed. Among the positives, many industries (like food, drink and furniture) are engaged in what would conventionally be regarded as low-technology activities.... ..."

"... In PAGE 9: ...Table2 presents the benchmark results for some uses of key language-level mechanisms. The test machine is a Pentium 166Mhz with 64 MBs of RAM running Linux 2.... ..."