CPSC 815 Special Effects Production
Fall 2000
TTh 12:30--1:45 Rhodes 312
Project 4
<http://www.vr.clemson.edu/~andrewd/courses/cpsc815/fall00/hw/proj4.html>
- Objectives:
- To learn the Discrete Wavelet Transform (DWT) and explore some
issues involved in image compression.
- Motivation:
- This project involves programming the DWT (decomposition and
reconstruction).
Issues include pyramidal image representation and thresholding
of wavelet coefficients.
- Due dates:
- Project writeup (Image DWT tool): 10/31/00
- Description:
- Implement a DWT tool in your image manipulation program, including
image decomposition, coefficient filtering, and reconstruction.
The user should be able to:
- control the levels of decompositions (1 to log(n) for
n x n images).
- select the thresholding parameter for coefficient filtering
(0% - 100%). This is similar to a quality metric,
where 0% filtering implies perfect reconstruction of
the image, and 100% filtering gives the lowest level
of decomposition (or vice versa).
During interaction with the DWT tool, the user should be
presented with a temporary display of the transformed image.
The user should be able to save this image or simply dismiss it.
- Notes, suggestions, and additional requirements:
- For the program, as in the previous assignment, the use of a good
toolkit is recommended (such as
gtk).
The following interaction
mechanisms should be made available to the user:
- the user needs to be able to manipulate the level of
DWT decomposition
- the user needs to be able to adjust the level of
coefficient decimation (filtering)
- What to hand in:
- ``Professional-quality'' brief writeup, including:
- Cover page containing:
Course Id--Section No.
Name:
Assignment No.
- Assignment description
- Screenshots of the program, including the GUI control and
image windows
- Sample images
- Electronic submission of images: provide a URL (web address) to your
on-line presentation of the assignment. To keep the page size
manageable, please use reduced images on your web page, but provide
links to the full-scale example images.