Introduction - Multimedia Database Retrieval: Technology and Applications - page 3

￿ Multimedia retrieval, including efficient database index, compact storage, and

quick and accurate retrieval performance. Since large-scale databases consist
of millions of images, the computational efficiency of both off-line and on-line
retrieval processes is crucial.

￿ Integration with cross platform-based applications . With the emerging

technologies of mobile devices and cloud computing, a lot of desktop-based
multimedia applications need to be migrated to cloud and must find suitable
positions in the mobile domain.

Multimedia database retrieval has attracted researchers from the fields of com-
puter vision, machine learning, database technology, and multimedia for almost two
decades. It still remains a popular research direction, especially when considering
how to cope with the vast size and increasing growth of multimedia data. In the
beginning of this millennium, Rui, Huang, and Chang stated that there are two
major difficulties with large-scale image datasets [ 6 ]. One is the vast amount of
labor required in manual image annotation. The other is how to understand different
human perceptions towards the same image content. Moreover, the question of how
to efficiently index large-scale image archives for fast retrieval was also raised as a
fundamental consideration in designing large-scale image retrieval systems [ 6 , 7 ].

1.2.3 The Development of Multimedia Database

Retrieval Technology

This topic presents the development of multimedia database retrieval technology
from two perspectives. The first perspective presents up-to-date methods and
appealing topics in multimedia retrieval. It shows the state-of-the art technology
used for small, medium, and large-scale databases. The second perspective provides
an application-oriented view of multimedia technology. This will inspire the reader
towards innovation in developing new applications, and towards the practice of
multimedia technology.

1.3 Technology Perspective

1.3.1 Human Centered Search and Retrieval

Visual seeking is the process of communication between a user and a computer
system, and requires a decision-making method for both sides. This visual seeking
involves interpretations of visual content by both the user and the computer. For
successful retrieval, it is necessary that the user and the computer system use the
same interpretation criteria.