Databases Reference
In-Depth Information
1.
Creation of metadata space:
To provide the function of semantic associative search, basic information
on m data items (“metadata for space creation”) is given in the form of a
matrix. Each metadata item is given independently one another. No
relationship between the metadata items need to be described. The
information of each data item is represented by its features. The
m
basic
metadata items is given in the form of an
m
by
n
matrix
M
where each
metadata item is characterized by
n
features. By using M, the orthogonal
space is created as the metadata space
MDS
. These metadata items are
determined as mentioned in the following section.
2.
Representation of metadata for images in
n
-dimensional vectors
Each metadata item for images is represented in the
n
-dimensional vector
whose elements correspond to
n
features. The metadata items for images
become the candidates for the semantic associative search. Furthermore,
each of the context words, which are used to represent the user's impression
and the image's contents in semantic image retrieval, is also represented in
the n-dimensional vector.
3.
Mapping data items into the metadata space
MDS
.
Metadata items (metadata for space creation, metadata for images and
metadata for context representation) which are represented in n-dimensional
vectors are mapped into the orthogonal metadata space. Those data items
are used as context words and target image data items which are extracted
according to users' requests.
4.
Semantic associative search
When a sequence of context words which determine the user's impression
and the image's contents is given, the images corresponding to the context
are extracted from a set of retrieval candidate images in the metadata space.
1.4
A creation method of the metadata space
We introduce an implementation method for the creation of the
MDS
.
The procedure for the creation of the
MDS
is as follows:
1.
To create the data matrix
M,
we can use “General Basic English
Dictionary
16)
”
in which 850 basic words are used to explain each English definition. Those
850 basic words are used as features, that is, they are used for characterizing
metadata as the features corresponding to the columns in the matrix
M
.
The 2,000 data items are used as “metadata for space creation.” Those
