X-Message-Number: 7516
Date: Thu, 16 Jan 1997 09:45:00 -0800 (PST)
From: Joseph Strout <>
Subject: Re: slow-motion detector

On Thu, 16 Jan 1997, Kennita Watson <> wrote:

> How would it tell whether or not you're at home (when considering
> reference images, include changes in lighting and poor housekeeping)?

It could not, of course.  That's probably the biggest problem; you'd have
to always remember to turn it off when you leave, or rig up some
automatic system.  Or you could only use it at night, if the risk of
dying is greater in your sleep than when wandering about (I have no idea
whether this is the case).

> Could I use a sensor small enough to bond to the back of my Alcor
> bracelet to change the program mode?

Not a sensor, but a transmitter.  Hmm, I'm more of a software than
hardware person, so I'm not sure.  Perhaps.

> How would you connect multiple cameras? ...
> or even multiple running instances of the program; multiple computers
> is right out.

Well, each camera connects to the computer via the serial port (printer or
modem port).  You could probably rig up an automated switchbox that would
switch to the next camera periodically, and coordinate this with the
software.  But again, that's a hardware problem, so I'm not very equipped
to deal with it.  OTOH, if you only run it at certain times (e.g., at
night while sleeping), then one camera might be OK.

> The idea could be turned around, or added to, such that the system
> (assuming it _could_ tell when you were home) could call your pager
> (or wherever) if the picture _did_ change.  You could even save such
> anomalous photos for possible evidence.

Yes, someone else pointed out that you could make a great security system
out of something like this.  That's true, but it wasn't really the point
of my post.  The issues involved are different.

> When you say you don't need a very powerful computer,
> what do you mean?  Would a Mac Plus work? 

No, you need to be running at least System 7.0, and I suspect you need at
least 4-bit monochrome.

> >(It was pointed out that live people move in their sleep with fair
> 
> BTW -- I think that's more like once every five to ten minutes than
> once every minute

May be.  I haven't actually tested it on sleepers yet, but of course the
time interval is very easy to change.

> -- and I'd imagine some of the motions involve, say,
> scratching the nose and putting the hand back where it was (to within
> the resolution of a QuickCam, anyway, especially if the motion is
> under a thick blanket).

The system is VERY sensitive -- when I purposely tried to move a hand and
return it to the exact same position, it almost always caught me.  A
blanket might make it even easier, since it will have many subtle changes
in wrinkles etc. from any motion.  But as I said, I haven't tested it (not
expecting to die in my sleep any time soon, and if I do, my wife is sure
to notice by morning!).

>  How light-sensitive is a QuickCam?

You can adjust the exposure to rather long values, at which point (like
any CCD) it becomes very sensitive.  If you have a digital clock next to
the bed, it will probably cast enough light for the camera to see by.
Similarly, if your window blinds don't block out all the street light.

> Hmm -- fortunately I own no animals.

That would be another major hitch for some users.

,------------------------------------------------------------------.
|    Joseph J. Strout           Department of Neuroscience, UCSD   |
|               http://www-acs.ucsd.edu/~jstrout/  |
`------------------------------------------------------------------'


Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=7516