Classification and Retrieval of Images from Databases Using Rough Set Theory - Distributed Artificial Intelligence, Agent Technology, and Collaborative Applications - page 186

• Inverse different moment ( IM ) is a measure of the image contrast or the amount of local variations

present in an image. It takes the following form:

1

∑

(10)

IM = P i j ( , )

2

1 ( + − i j )

i j ,

This descriptor has large values in cases in which the largest elements in P are along the principal
diagonal.

• Angular second moment ( AM ) takes the form:

= ∑

, (11)

AM P i j ( , )

2

i j ,

For homogeneous textures value of angular second moment turns out to be small compared to
nonhomogeneous ones.

• Entropy is a measure of information content. It measures the randomness of intensity distribu-

tion.

= ∑

(12)

Entro P i j ( , )log( ( , )) P i j

i j ,

Where P ( i ,j ) refers to the normalized entry of the co-occurrence matrices. That is, ( , )

P i j P i j R = ( , ) /

d

where R is the total number of pixel pairs ( i, j ) for a displacement vector ( , ) and image of

d dx dy =

size NxM , R is given by (

N dx M dy − )( − ).

In order to obtain better retrieval results, the image texture features can be combined with the color
features to form a powerful discriminating feature vector for each image. Various color identification
schemes have been proposed and used. The RGB (Red, Green, Blue) model has been widely adopted
because of its implementation simplicity (Ahuja & Rosefeld, 1978; Yining & Manjuncth, 1999 ; Zhang,
H. Gong, Y. Low, C.Y. & Smoliar S.W , 1995). Despite this, the RGB model has proved unable to
separate the luminance and chromatic components; furthermore, it results in perceptually nonuniform,
i.e., perceptual changes in color, which are not linear with numerical changes. The HVC (Hue, Value,
Chroma) color model completely separates the luminance and chromatic components with Hue repre-
senting the color type, Value representing the luminance, and Chroma representing the color purity. The
transformation from the RGB model to the HVC model can be performed in several ways; in this work,
the transformation is obtained through the CIE L*a*b* model. Assuming a 24-bit-per-pixel context, the
RGB components are transformed into the CIE XYZ components using the following equations:

 X = 0.607* R + 0.17* G + 0.210* B

Y = 0.299* R + 0.587* G + 0.114* B




Z = 0.066* G + 1.117* B



Then the HVC values are finally obtained as follows: