Thread: Calling Any Java Programmers - Computer Vision Project

unfortunately I've put this project on hold for a bit. I'm currently working on tomtom (pretty much done) and on a new ibus program 'cos I'm getting very annoyed with what I can't do in vb!

I haven't forgotten this, I will resume at some point. I wish I didn't have to go to work!

ah this is so cool. I have been looking all over the place for people doing this kind of stuff and in java too. although i never considered this general appoach for object recognition. i was trying to develope a approch based more on 3d. but this stuff is so cool.

im feeling a little brave with my C++ knowledge. i will join this task force. Sama, ill im you Merry christmas!

Originally Posted by sama

this is exactly my aim, to try to get a framework going so that people with little or no knowledge of vision algorithms can build applications, which in turn would drive the development of algorithm plugins to use.

my maths isn't the best either, but there are plenty of resources out there.

the current download offers a fairly simple motion detection algorithm. it learns the background by continously averaging pixel values over time. It detects foreground objects by subtracting the current frame from the learnt background.

the problems currently with it are that any major change in lighting will render the background model useless, since it no longer applies to the new lighting conditions. a way of detecting this would be useful, so a reset can be issued to the background learning loop. This infers that there would be a window where the foreground is not determined whilst the background is being learnt again.

Perhaps another motion detection algorithm can kick in at that stage, something like differential analysis (subtracting previous frame from current). hopefully, by using knowledge prior to the light change, the regions of last known moving objects can be applied to the differntial algorithm.

Im' thinking out loud here on how the impact of lighting changes can be minimised. you may need to study the code a little to to see what I'm gettin at.

Perhaps using the background + a variable that represents intensity of the whole image. This way you can learn the background and keep it if all pixels change in the same way (intensity of whole image changes)... don't know if that's possible but it would be the most logical way of doing this.
On the other hand, it will only work within a certain dynamical range, as long as everything in the backgound remains visible (distinguishable pixels).

Good luck with this project!

Joachim

Very interesting project, wanted to bring it out of hibernation after finding it on Google.
Any updates? I'm using the JMF myself to develop a new front end but won't be incorporating video streaming until all the audio is done.

Hi,

Im very interested in your project.
What's available?

M

Originally Posted by sama

I am very much into AI. There's a book called Vision by a genius called David Marr. I highly recommend this (if you can find it). He delves into explaining how human vision is a process, and not some magic, and he mixes neuropsychology with computer programs. I've been trying to locate a copy ever since I read a part of it. I may have to buy it for the full price of $300-400 (out of print now). I will look into the Jeff Hawkins stuff too.

Sama,

Hope all is well, I have not posted on this forum in far too long.. I am taking a grad course in AI and I am hoping to get approval to work on a similar project to this one. In preparing I remembered this thread and stopped back to re-read it. I also wanted to let you know in case you were still looking that mit press re-released Marr's book and I got a copy. You can get it on amazon here

Hope all is well,

Brian