Thursday, July 12, 2007

Matlab week? sic!

Who said matlab is easier to debug? Jumping gods, I was really frustrated with this one. It seems that I finally completed the AdaBoost training code, the classifiers are trained, and I'm getting some positive results. That's mostly thanks to the revised version of Viola & Jones paper that I found yesterday evening, hiding on my hard drive actually! I trained 15 classifiers and put them into 3 stages of a cascade, and I'm getting this (!!!) :



Wow, so I'm back on top of it, now rewriting to C! :-)

5 comments:

Marc-André Lureau said...

Indeed, VJ paper is really good. What kind of weak classifier/feature extraction are you using? What is your learning algorithm?

I would be glad to help you implement this in a gnomish-way (ex: embeddable in gegl or gstreamer).

I have been working on this methods in the past (with AB/GA). Are you going to GUADEC?

apart said...

Hi, I'll post the algorithm to my webpage soon, I'm currently tuning some stuff - it seems that my error rate is decreasing a bit too fast ;)
Please contact me via email, (iapart at gmail) I'll probably have some questions :-)
Unfortunately I'm not going to GUADEC, and I have very little time to finish this :/ so any help apreciated!

Anonymous said...

Hey! Matlab works really great with wine! Have a try!

Will Dwinnell said...

Who said matlab is easier to debug?

Easier than... what? At least most of the loops, pointer stuff, etc. goes away with MATLAB.


-Will

Unknown said...

Hiiii

I am doing research on multiuser CDMA; I have simulated MUD's in both AWGN and Rayleigh fading channel. I need help in Multistage Successive interference Cancellation. Kindly if anyone have the m code for this, please share.

For others, I can get you papers from most of the journals as I have access to them. Just ask and I will send you the paper.

I will be happy to share all the .m code along with the SIC code once I finish the multistage as I have already finished single stage ordered SIC .
I just need help for multistage.