Skip to main content
Figure 2 | Journal of the Brazilian Computer Society

Figure 2

From: Findings on ranking evaluation functions for feature weighting in image retrieval

Figure 2

Average P&R graphs for all evaluation (fitness) functions. Analyzed on (b) Vistex-167 data set, (c) Corel-1000 dataset, (d) DB-10000 data set, (e) Scenes-1044 data set and (f) Caltech101-8872 data set. The legend for the graphs is given inside (a). The parameters used were the same as those found in studies from the literature: k1=6, k2=1.2, k3=2, k4=3.65, k5=0.1, k6=4, k7=27.32, k8=7, k9=0.982 and A=10. P&R graphs for Vistex-167 and Corel-1000 have been obtained using all the images of the data sets as queries. P&R graphs for DB-10000 data set have been obtained using 1,000 query images, originating from Corel-1000 data set.

Back to article page