Intelligent Image Databases: Towards Advanced Image Retrieval
addresses the image feature selection issue in developing content-based
image retrieval systems. The book first discusses the four important
issues in developing a complete content-based image retrieval system,
and then demonstrates that image feature selection has significant
impact on the remaining issues of system design. Next, it presents an
in-depth literature survey on typical image features explored by
contemporary content-based image retrieval systems for image matching
and retrieval purposes. The goal of the survey is to determine the
characteristics and the effectiveness of individual features, so as to
establish guidelines for future development of content-based image
retrieval systems.
Intelligent Image Databases: Towards Advanced Image Retrieval
describes the Advanced Region-Based Image Retrieval System (ARBIRS)
developed by the authors for color images of real-world scenes. They
have selected image regions for building ARBIRS as the literature survey
suggests that prominent image regions, along with their associated
features, provide a higher probability for achieving a higher level
content-based image retrieval system. A major challenge in building a
region-based image retrieval system is that prominent regions are rather
difficult to capture in an accurate and error-free condition,
particularly those in images of real-world scenes. To meet this
challenge, the book proposes an integrated approach to tackle the
problem via feature capturing, feature indexing, and database query.
Through comprehensive system evaluation, it is demonstrated how these
systematically integrated efforts work effectively to accomplish
advanced image retrieval.
Intelligent Image Databases: Towards Advanced Image Retrieval serves
as an excellent reference and may be used as a text for advanced courses
on the topic.
