Title:
Generic Multimedia Database Architecture: Towards High-Level Semantic
Retrieval
Abstract: Extraction of semantics has been a challenging problem
in multimedia databases. Common methods for semantic extraction are based
on manual annotation of multimedia data and/or content retrieval based on
local or global feature distribution. Content retrieval based feature extraction
(e.g. color, texture and location) fails to extract the semantics. Manual
annotation of multimedia data for semantics presents promising results but
it is time consuming and will be impossible when considering the huge database
compose of thousands of images. In this paper we have presented the idea
of semantic extraction by using semantic libraries. In our approach we extract
the semantics by considering the dominant objects in the image, the relationships
among these objects, state of these objects and the environment in which
they exist. Design and working along with proposed results of these semantic
libraries have been shown. Our proposed generic multimedia database architecture
could give better results with the improvement in the technologies related
to object detection, region splitting and feature extraction.
Authors: Omara Abdul Hamid, Muhammad
Abdul Qadir, Mohib ur Rehman, Nadeem Iftikhar, Mobin Uddin Ahmed, and Imran
Ihsan