Databases Reference
In-Depth Information
FIGURE 8.5: Framework for commodity-grid based pattern recognition.
The Karajan architecture offers additional libraries for the front-end design
through its HTML and forms libraries. It uses the task library for grid inte-
gration, which is based on the Java CoG Kit abstractions.
8.2 Face Recognition Using the Multi-Feature DPR Ap-
proach
With the advent of Internet connectivity, the outgrowth of images pro-
duced and posted on the Web is tremendous in numbers. The nature of the
images is diverse. Some images are taken by normal or highly specialized cam-
eras and some, such as MRI and seismic images, are sophisticated and multi-
dimensional. Images are used in a variety of fields, including arts, engineering,
and sciences. Images are commonly used to recognize or detect objects, such
as a human face.
Face recognition is a well-known application that implements pattern recog-
nition concepts and approaches. Numerous computational designs and imple-
mentations have been proposed in the literature. Each of these approaches
contributes to accuracy and/or e
ciency of the application. In this section,
we will present a distributed approach for face recognition using the multi-
feature DHGN algorithm. Our intent is not to outweigh the capabilities of
current and more establish facial recognition schemes, but to give an insight
into how this complex operation on Internet-scale data can be performed in
a distributed manner. Further details on the analysis of the recognition accu-
