[cross-posted to:bionet.neuroscience, bionet.software,
comp.sys.mac.scitech, alt.graphics.pixutils, alt.image.medical,
comp.graphics.visualization, sci.image.processing]
Hello All,
I have a colleague here in the Biology Department at Swarthmore College
who is beginning to use a PM 7100/AV to automate her observations of
animal behavior. She came to me for advice, but my experience has been
primarily in the application of digital image analysis to microscopic
anatomy. Though I was able to give general advice on the use of
digital image analysis in the laboratory, I could offer little specific
information regarding the use of digital video to automate behavioral
analysis.
The sort of investigations she is hoping to automate entail identifying
animals within the video frame, performing simple morphometric
measurements on these animals (e.g., the width, breadth, area, &
centroid of an arthropod thorax; the angle between a foreleg & thorax;
etc.) , then following the identified animals through a series of video
frames. Typical questions she plans to address are:
(1) given two animals of different size, how often does each animal
visit a food item within the video frame?
(2) when two or more animals are within a video frame, what body parts
do they use when initially contacting one another?
(3) in the course of exhibiting a stereotypical behavioral display, how
do an animal's body parts move in relation to one another?
I am aware that imaging programs (e.g., NIH-Image, IPLab-Spectrum,
etc.) support various frame-grabber boards. Which programs directly
support digital video input from AV Macs/PowerMacs? My colleague could
partially automate her behvioral analysis by using these general
purpose imaging tools, but are there specific software/hardware
solutions that would be more suited to the needs of motion analysis?
Most of her images are high contrast and of sufficient resolution to
easily identify the structural elements she is interested in
monitoring; therefore, the tools that would be of most use to her are
not the techniques commonly used to improve image resolution in digital
microscopy, but rather tools specifically designed to perform
time-series analysis on segmented objects within the video frame (e.g.,
location and postion of segmented objects, movements of several
segmented objects relative to one another, etc.). She is interested
both in ways that she might use these tools to analyze old data already
stored on video tape, as well as how such techniques could change the
way in which she performs future experiments.
Thank you all ahead of time for your assistance. Replies may be posted
to this newsgroup or mailed directly to me. If I receive bevvy of
useful replies, I'll will certainly summarize & post them.
Cheers,
Bill Bug
************************************************************************
* \ / Bill Bug
* \__/ bbug1 at swarthmore.edu
* {''} Swarthmore PA USA
* {}
* (My opinions are as much my own as anyone else's)
*-----------------------------------------------------------------------
*
* None is so poor that he need sit on a pumpkin.
*
* (from _Walden,_or_Life_In_The_Woods_ by Henry David Thoreau)
*
************************************************************************
