the easiest way of doing this would be by using the ‘jit.change’ object.
That passes through a matrix if that matrix changed more or equal to a
this makes it easy to detect motion, but could make things harder to make
the difference if the visitor has left or when he just stands still.
For that matter you could work with background substraction, where you
‘calibrate’ your system by making a picture when setting up the
installation. You then substract each newly captured frame from the
‘reference frame’ (= the picture you took). This gives a measure for change
in the room and is done ustin the ‘jit.op’ object.
you can re-calibrate your system by taking a fresh picture once in while
when you’re sure there’s no-one around >> this is interesting to incalculate
changes in light etc. in the room.
These are basic techniques, but i think you can start with those to see if
they deliver what you need. If this all doesn’t sound satisfying you could
go have a look at the cv.jit objects which offer a nice load of computer
vision techniques, but require some more in-depth knowledge of how max and
On 10/25/06, luckyforest wrote:
> I am building a project that must trigger the effect when someone goes
> into the camera area,when the man gets away, the screen comes back to normal
> – the effect is over.
> could somebody know how to catch digits from the live video, and get it
> changes along with the movement of the video?