Next: Full text index Contents: Conference
Page
Wavelet-Based 3D Compression Schemes for the Visible Human
Data Set and Their Applications
Abstract
Visualizing volume data is a task requiring great cost of time and
space. It is extremely expensive particularly when the volume data is very
large in size. This paper presents a method that produces effectively a
direct-volume-rendered image from compressed volume data. Contrary to most
previous efforts which tend to use high-end workstations or high performance
parallel computers, our intention is to propose a technique which visualizes
needed parts from compressed volume. For compression of the huge volume
data, we use an enhanced encoding scheme based on wavelets. The technique
does not only offer fairly good compression ratio, but also supports much
faster random access than our previous method. In this paper, we describe
an enhanced wavelet-based 3D compression scheme, called zerobit encoding,
and report experimental results on the Visible Human dataset. And we will
show that our compression schem is useful in developing various interactive
visualization systems on personal computers or workstations with limited
memory as well as parallel computers.
Keywords: wavelet, lossy compression, multiresolution, embedded
zero tree.
Next: Full text index Contents: Conference
Page
