home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!pipex!warwick!warwick!not-for-mail
- From: esrgl@csv.warwick.ac.uk (Mr S J Sheppard)
- Newsgroups: comp.ai.neural-nets
- Subject: Re: Image Input for a Neural Net
- Date: 13 Nov 1992 09:53:49 -0000
- Organization: Computing Services, University of Warwick, UK
- Lines: 42
- Message-ID: <1dvtvdINNfgm@violet.csv.warwick.ac.uk>
- References: <1992Nov12.140317.7643@debbie.cc.nctu.edu.tw> <1992Nov12.181625.4135@newssun.med.miami.edu>
- NNTP-Posting-Host: violet.csv.warwick.ac.uk
-
- In article <1992Nov12.181625.4135@newssun.med.miami.edu> dbrown@newssun.med.miami.edu (Daniel Brown) writes:
- >I'm looking for information about preprocessing for image
- >inputs into a neural-net.
- >
- >Our project requires the use of a 480x480 24-bit color
- >image file in the tiff format. Even after converting to
- >a greyscale image, that's over 65,000 inputs if used
- >directly. I don't know too many systems that can handle
- >that much information; certainly ours can't.
- >
- >I've looked a couple of compression techniques: i) summing
- >and averaging over 100 pixels at a time (a coarse
- >version of the former image), ii) coarse coding similar
- >to the latter), and iii) sampling of a FFT transform of
- >the input data.
- >
- >The obvious goal is to use as much important information
- >without overloading the system. Does anyone have any other
- >suggestions?
- >
-
- Try Principal Component Analysis!
- I've been working on applying Neural Networks to the Diagnosis of
- Medical Images and have also been working with large input images.
- (Although perhaps not as large as these). Using PCA's I've managed
- to reduce the dimensionality of my data by upto 5x, but still being
- able to keep over 98% of the information.
-
- Good Luck
-
- Steve Sheppard
-
- ************************************************************************
- ** **
- ** Steve Sheppard **
- ** Research Postgraduate E-mail: esrgl@uk.ac.warwick.csv **
- ** Dept of Engineering Tele : 44 - 203 - 523523 extn 2654 **
- ** Warwick University Fax : 44 - 203 - 418922 **
- ** Coventry **
- ** CV4 7AL **
- ** **
- ************************************************************************
-