Image Processing Reference
In-Depth Information
and − 90 ° of w − 1 pixels. In such a case, each pixel of the image can be seen as the intersection
of four corresponding elements of the partitions Y 1 , Y 2 , Y 3 , Y 4 as shown in Figure 3 .
FIGURE 3 Each image pixel can be seen as the intersection of four elements of the parti-
Since for each partition, it is possible to define a C -set, each pixel can be seen as belonging
to the partition obtained by the product of the original four C -set:
where C i is the composite set corresponding to partition Υ i . We shall define as scale the size
w of each partition element. The product operator is neither idempotent nor increasing. The
fact that this operator is not idempotent allows it to be iteratively applied to the input signal
in order to construct the scales pace. The multiscale construction follows that of a fuzzy NN
[ 33 , 34 ] . Specifically, it consists of two pyramidal-layered networks with fixed weights, each
looking upon an 2 n × 2 n image. By fixing the initial dimension of CRC ( C andidate R egion to be
C ategorized), each pyramidal network is constituted by n R multiresolution levels. Each pro-
cessing element ( i , j ) at the r th level of the first pyramid (respectively, second pyramid) com-
putes the minimum value (respectively maximum value) over a 2 w × 2 w area at the ( r − 1)th
level. The pyramidal structures are computed in a top-down manner, firstly analyzing re-
gions as large as possible and then proceeding by spliting regions turned out to be not of in-
terest. The mechanism of spliting operates as follows. If we suppose to be at the r th level of
both pyramid-networks and analyze a region w × w which is the intersection of four 2 w × 2 w
regions, the minimum and maximum values computed inside are denoted by m s , t and M s , t ,
s = i , …, i + w , t = j , …, j + w . The combination of the minima and maxima values is made up at
the output layer, i.e.,
Search WWH ::

Custom Search