78
Adaptive
Storage and Archiving Database and Compression System--Physical
Optics Corporation, Engineering & Products Division, 20600 Gramercy Place,
Building 100, Torrance, CA 90501;
310-320-3088
Dr.
Michael Piliavin, Principal Investigator,
Mr.
Gordon Drew, Business Official, gdrew@poc.com
DOE
Grant No. DE-FG03-01ER83293
Amount:
$99,996
The
immense volume of data generated by nuclear physics experimentation, and the
unlikelihood that it could ever be regenerated, make efficient and secure
archiving extremely important. Current
data compression methods, whether lossy or not, make use of a priori knowledge to compress the input data.
This limits their capability to optimally perform compression as the type
of content in the data streams changes. In
addition, only certain types of data such as video and text are highly optimized
as separate input streams. This
project will develop an artificial intelligence/fuzzy logic algorithm to test
the characteristics of input data streams and select the optimal compression
method. The algorithm will allow for complex (e.g., a combination of
audio and video) data streams. Phase
I will investigate the available algorithms for lossless and lossy compression.
Also, a corpus will be selected for use in evaluating the performance of
these algorithms. A new AI/fuzzy
logic engine will be created and its performance demonstrated.
Commercial Applications And Other Benefits as described by awardee: The market for data compression is growing fast, driven by the proliferation of multimedia, which brings together video, sound, and text in a single data stream. Applications for this technology are in industries as diverse as finance, entertainment media, and science.