[Nets-seminars] Talk on Friday 11th December 3pm GS/306 Davide Anastasia

Richard G. Clegg richard at richardclegg.org
Mon Dec 7 10:10:31 GMT 2009


Davide from EE will be talking on ""Software Designs Of Image Processing 
Tasks With Incremental Refinement of Computation".

Note the change of time and venue to avoid clashing with the (now 
cancelled) Christmas party.

Software realizations of computationally-demanding image processing
tasks (e.g. image transforms and convolution) do not currently provide
graceful degradation when their clock-cycles budgets are reduced, e.g.
when delay deadlines are imposed in a multi-tasking environment to
meet throughput requirements. This is an important obstacle in the
quest for full utilization of modern programmable platforms"
capabilities, since (i) worst-case considerations must be in place for
reasonable quality of results, and (ii) throughput-distortion
tradeoffs are not possible for a variety of distortion-tolerant image
processing applications. In this talk we outline work on a proposed
(and make available online) platform-independent software designs
performing bitplane-based computation combined with an incremental
packing framework in order to realize block transforms, 2D convolution
and frame-by-frame block matching. The proposed framework realizes
incremental computation: the computation of the final result for each
input video frame is performed incrementally, with progressive
output-quality improvement when more system resources are utilized.
Comparisons with the equivalent non-incremental software realization
of each algorithm on a mainstream notebook computer and on OLPC's
subnotebook ("100$ laptop") reveal that, for the same precision of the
result, the proposed approach can lead to comparable or faster
execution, while it can be arbitrarily terminated and provide the
result up to the computed precision. Tests with task scheduling per
frame and energy-distortion scalability verify that our proposal
provides significant performance scalability with graceful
degradation. This is achieved in a straightforward manner and without
the need for system-specific customizations.



-- 
Richard G. Clegg,
Dept of Elec. Eng.,
University College London
http://www.richardclegg.org/



More information about the Nets-seminars mailing list