[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [help-GIFT] 'Re: Ups.

From: risc
Subject: Re: [help-GIFT] 'Re: Ups.
Date: Thu, 13 Apr 2006 10:19:42 -0500
User-agent: Mutt/1.4.1i

On Thu, Apr 06, 2006 at 06:15:01PM +0100, David Squire wrote:
> Wolfgang Müller wrote:
> >>i packaged it as a fork for a reason. i had to change generally EVERYTHING
> >>about the coding style, to get gcc to auto-vectorize. there are (almost) 
> >>no
> >>global variables anymore, the gabor filters are in one array, not four.. 
> >>it
> >>is truely unholy what i've done to your poor program.
> >>but, it works. ;)
> >>    
> >
> >I would suggest that you and David fix a close deadline for David to look 
> >at it more closely. He knows the code best. I have barely touched it since 
> >it came out. David, if you do not feel you have the time to commit, please 
> >tell us.
> >
> >  
> I should be able to look at it some time in the next week. I have a more 
> pressing deadline with coding I am doing for Damocles (which has "real 
> users") that I need to focus on first. I'm afraid that I can't be more 
> precise than that.

Hi David, its been a week. ;)

> You know, I am amazed that that feature extraction code has survived as 
> long as it has. It was only ever done really as a very quick and dirty 
> proof of concept for the SCIA'99 paper :) I guess it was "good enough", 
> so it lasted.

looking through the code, i'm not too supprised. i've re-written most of 
it, and i still dont quite understand some things. for instance, why
are the gabor waves stored in four different arrays in your origional program?
I think that confusion may have helped when optimizing, however. ;)

> I agree that I would prefer to see ImageMagick retained, though perhaps 
> by linking to their libraries rather than making system calls.

the only reason i went with netPBM was one of my initial goals was to run 
against a whole one of my images, in native resolution. i've read that PPM 
has changed formats, and it looked to me like the PPM code in GIFT wouldnt
handle any images greater than 256x256.

> Cheers,
> David
> -- 
> Dr David McG. Squire,  Senior Lecturer, on sabbatical in 2006
> Caulfield School of Information Technology, Monash University, Australia
> CRICOS Provider No. 00008C
Look forward to hearing from you,

Julia Longtin <address@hidden>

reply via email to

[Prev in Thread] Current Thread [Next in Thread]