THE DVIP
LAB
 
 

IMAGES IN MOTION

 
 
BLUR 
IS
GOOD!
 
 
 Image Sequences
 
 
"Close" surveilance
MD 
 (representative frames)
"Eavesdrop" sequence
NMD
  (representative frames)
"Block" 
(representative frames)
md.zip  md.tar.gz
nmd.zip  nmd.tar.gz
nblock.tar.gz
 
NEW MATLAB TOOLS
showdvf.m
pgmmovie.m
bounds.m

Intelligence
Project Horatio

1. Image data in PGM format.
Do not download the animated gif of these sequences for the purposes of this assignment. It is for aesthetic purposes only.  The image sequence data is contained in .zip or .gz files below the gif.  Frame numbers for each sequence begin with 0.
 
2. If you plan to consider image sequence processing (including other motion estimation approaches) as your final project, this would be a good time to develop some tools to assist in your effort.

3. On viewing difference images:
Remember that viewing images using imagesc(f) scales the entire dynamic range of the image to use the full range of the image gray map. For this reason, imagesc will not give an accurate account of the magnitude of the difference.  Another approach would take the difference image add 128 and bound the values between 0 and 255 without scaling thus preserving the relative magnitude of the differences.

 
PROJECT DUE: Monday, April 13, 1998 before 4:00pm in Rm. 3537,
Agent Jean Stover's Office.