• "squish" all consecutive identical rows into only one row.
• Do the same for columns

my $IMAGE = <<END:
0000000000000000000000
0000000001111100000000
0000000001111100000000
0000000001111100000000
0000000001111100000000
0000011111111111110000
0000011111111111110000
0000011111111111110000
0000011111111111110000
0000000001111100000000
0000000001111100000000
0000000001111100000000
0000000001111100000000
0000000000000000000000
END

sub transpose {
    my @matrix = map [ split // ], split /\n/, shift;
    my $dim    = $#{ $matrix[0] };

    @matrix = map { my $x=$_; [ map { $matrix[$_][$x] } 0..$#matrix ] 
+}
              0..$dim;
    return join "", map { join("", @$_) . "\n" } @matrix;
}

sub squish {
    my $str = shift;
    $str =~ s/^(.*\n)\1+/$1/gm;
    return $str;
}

print transpose squish transpose squish $IMAGE;

__OUTPUT__
00000
00100
01110
00100
00000
[download]

To match the exact output you wanted, you would also have to trim all leading and trailing trailing all-zero rows & columns.

This doesn't answer your question at all but it's still kind of neat. It transforms your example input into the example output by removing repeated lines, vertically and horizontally. It also removes "empty" lines from the edges.

$image = <<IMAGE;
0000000000000000000000
0000000001111100000000
0000000001111100000000
0000000001111100000000
0000000001111100000000
0000011111111111110000
0000011111111111110000
0000011111111111110000
0000011111111111110000
0000000001111100000000
0000000001111100000000
0000000001111100000000
0000000001111100000000
0000000000000000000000
IMAGE

$result =
    transpose(
        trim( vertical_squish( transpose(
            trim( vertical_squish( $image ) ) ) ) ) );

$result eq <<EXPECT or die "Failed.\n$result\n";
010
111
010
EXPECT

sub trim {
    local $_ = shift;

    s/\A(?:^0+[\r\n]+)+//m;
    s/(?:^0+[\r\n]+)+\z//m;

    return $_;
}

sub vertical_squish {
    local $_ = shift;

    s/(^[^\r\n]+[\r\n]+)\1+/$1/gm;

    return $_;
}

use Algorithm::Loops 'MapCarU';
sub transpose {
    local $_ = shift;

    return
        join( "",
              &MapCarU(
                  sub { join( "", @_ ) . "\n" },
                  map( [ split // ],
                       split( /[\r\n]+/ ) ) ) );
}
[download]

Found via Googling "fractal image comparison" - a paper from the Journal of Information Science and Engineering:

... However, when users obtain an image from a newspaper or magazine, they want to retrieve related information about this type of image through a search engine, traditional search engines are incapable of meeting their needs. To help solve this problem, this paper provides an Image Comparison Search Engine (ICSE), and makes use of ˇ§Fractal Image Processingˇ¨ to create a database using image eigenvalues. When users input an image query, this system will generate image eigenvalue data, compare this with the data in the database of image eigenvale, and output the results. Based on our experimental results, ICSE can not only find the exact input image for the source image, but also find the ˇ§right imageˇ¨ when the source image is rotated and misty, and contains impurities...

Disclaimer: this is so not my field, so I've no idea how good / feasible this may be for you.

There's quite a lot of technical literature on pattern recognition. (a searh for "pattern recognition" on ACM's website gave me about 4500 hits). I know you can google at least as well as I can, so I'm not going to just give you some random links.

I think that MIT's AI lab has done quite a bit of work on pattern recognition. Looking there would probably be a good place to start (and some of their papers are probably classics).

It may be interesting (and challenging) to try to detect steganography in images, too.

Yes I can google. There is an abundant amout of information and I should have mentioned that. I was more wondering if someone else was interested in this, and perhaps if anyone had some "real world" expereince/ideas to share. I did google for manything but often google is bane in that it returns tons of useless stuff too. BTW what is ACM? Anf if you've done anything with pattern recognition do you have specific links that are good? I'm off to try and find MIT AI papers now. Thanks.


___________
Eric Hodges

ACM is the Association for Computing Machinery.

---

Jeff japhy Pinyan, P.L., P.M., P.O.D, X.S.: Perl, regex, and perl hacker
How can we ever be the sold short or the cheated, we who for every service have long ago been overpaid? ~~ Meister Eckhart

01110
10001
10001
11111
10001
10001
[download]