Classification and Retrieval of Images from Databases Using Rough Set Theory - Distributed Artificial Intelligence, Agent Technology, and Collaborative Applications - page 192

Table 1, Retrieval by image classification, where T, R, and C are the Total, Retrieved, and Correct number
of images in structure class

Distance measure

T,R,C Re=C/T% Pr=C/R% Fm=(2*Pr*Re)/

(Re+Pr)

Rough distance 305,262,235 77.00% 89.697% 82.86%

Histogram
intersection

305,189,152 49.83% 80.42% 61.73%

Euclidean distance 305,155,96 31.47% 42.58% 36.19%

Figure 3. A comparative between similarities measures

r etrieval Analysis

0

00

0

total
retreived
correct number

00

0

00

0

0

Rough distance Histogram

intersection

Euclidean
distance

similarity measure