Hi All,

I wish to buy some services from a programmer comfortable
with programming graphics.

I require a routine for a security application I am writing.  My
application requires users to choose a password which consists
of 2 things:
1. a traditional alphanumeric secret password
2. a chosen picture

The picture they select will be one small image out of a grid of
similar images.  They will be required to remember which picture they
chose, so when they next log in, they must pick the correct picture
again.

FYI: This makes it difficult for them to tell anyone their password,
and since the pictures will be individually generated based in part
upon their password, this hinders spoofing and password theft.

I require the routine which generates the images.  This routine will
be called many times successively to generate multiple images.

The input to the routine will be: (1) an arbitrary lengthed string of
pure random data, (2) a width, (3) a height, and (4) an image-type
specifier (a small integer)

The output from the routine will be a picture (in any simple 24bit raw
format - it will be combined with other images by my code to be output
eventually as a GIF, PNG, JPG, BMP, or whatever).

The picture should be the height and width specified.

The picture should depend on the random data - every picture produced
with different random data must look different.  Every picture
produced using the same random data as an earlier picture must look
identical to that earlier one.

I am open to suggestion as to exactly what the pictures look like; but
it would be desirable if they appeared to be pictures of peoples
faces.  For example, I might want to produce a 4 x 4 grid of different
peoples faces.  I will call your routine 16 times, each time supplying
different random data.  You will use the data to "draw a face", and
return each picture in turn.  You might choose to implement it like
this:-

Data consumed       Purpose
-------------   ------------------------------------------------------
First 5 bits    Selects the hair color of the persons face (from 32
                different possible colors)
next 5 bits     Selects the length of their hair
next 5 bits     Selects the style of their hair
next 5 bits     Selects the shape of their face
next 4 bits     Selects color of their skin
next 3 bits     Selects color of their eyes
next 3 bits     Selects the gender of the person
next X bits     Selects the proportions of the layout of their eyes,
                nose, mouth, etc
next 5 bits     Selects their "fatness"
... and so on: this is just a general guide to help you understand
that the generated picture should be based upon the random data: any
other way of doing this is also fine.

I can provide the random data either expanded or contracted to suit
your needs - use as much as of it as you can.

The point of doing this is to produce a grid which contains a small
number (eg: 16) of different pictures with each one being
sufficiently unique that a person is going to remember which one they
chose as their password - days, weeks, or months later.  FYI: When the
grid is displayed on subsequent runs, the 16 pictures will be the
same, except jumbled up.  My code jumbles them though - all you need
to concentrate on is the production of single images.

The "image type specifier" input data is intended to determine what
kinds of pictures to produce.  Only 1 kind is required for this
project, but thought should be put into this so that different kinds
of images can be produced in the future.

I have just described the use of "peoples faces" as one kind of "image
type specifier".

A different kind of "image type specifier" might be "cartoon
characters" instead of peoples faces.  Yet another type might be
"dogs" or "cats" or "monsters" or "random" (where "random" means
either dog or cat or face or monster or cartoon etc...)

You will be required to enter into a confidentiality agreement before
commencing work, must agree to destroy all your source after delivery
and never to create derivative or competitive works for anyone besides
me in future.  The intellectual property in this project remains my
property, and you acknowledge that US and international patents exist
protecting this intellectual property.

My main requirement is that the pictures output by your code look
impressive.  I want everyone who encounters my password system to be
impressed by its quality.  If you can put some thought into your
submission and explain why your pictures will look better than anyone
elses, this will be my primary concern when choosing a winner.  Note
that your code must draw the pictures - it will probably not be
acceptable for the pictures to be saved in files and simply returned,
because insufficient variety of different images would result.

I would also like to know your proposed programming language (or list
if you are competent to develop in more than 1 language) and if
additional tools will be required (eg: ray-tracing engines or graphics
libraries) a list of what you need and details about any costs or
licencing issues if relevant.

If you wish to bid - please email details and amount+currency to me
at this address:-

cnd@geek

adding ".net.au" onto the above partial (non-spambot-readable-I-hope)
address.

If you are familiar with Elance - the project is there as well:-
http://www.elance.com/c/rfp/main/jobDesc.pl?jobId=8739625

Chris.

Post a reply to this message

On Sun, 20 Jul 2003 21:01:31 +1000, " !" <ban### [at] bangbangom> wrote:
> I wish to buy some services from a programmer comfortable
> with programming graphics.

See point 5 or 6 at http://www.povray.org/supporting-povray.html if your wish
fits. Use email listed there.

ABX

Post a reply to this message

Ta - good idea!

"ABX" <abx### [at] abxartpl> wrote in message
news:vp0nhvotooe2s61no86ee6rvn776pmu1jg@4ax.com...
> On Sun, 20 Jul 2003 21:01:31 +1000, " !" <ban### [at] bangbangom> wrote:
> > I wish to buy some services from a programmer comfortable
> > with programming graphics.
>
> See point 5 or 6 at http://www.povray.org/supporting-povray.html if your
wish
> fits. Use email listed there.
>
> ABX

Post a reply to this message

Nobody replied.

Is anyone here interested?

Post a reply to this message

Well, it sounds like a fun thing to do, and it
isn't than complicated really. To take a string,
divide it into different parts, then use these
as base for the rand()-Stream, that's not
very difficult. Actually, I've done that for
years. But it would involve POV-Ray
as the raytracing engine.

I see a problem in the precalculation of
images. One could easily create tons of
abstract images based on a random-stream.
But generating images, like faces (though
cartoonish like ones would probably be
sufficient), isn't so easy to do. My main
problem is that there would have to be a routine
which will hand the string to the image-file.

For example, I could easily create some scene
which will take a certain file (like "keyword.inc")
which has a string inside (with quotes at beginning
and end, e.g. "thisisthekeyword") and render an
image from that. I could even take the keyword,
joggle it around and render a sequence of random
images directly (if file I/O is supported so that I can
keep an ever evolving random-number somewhere).

Another problem lies in the speed of rendering
with POV-Ray, which would mean that there'd
be either waiting when loading the password-screen,
or precalculation of images involved. You can
email me for details if you like, but I can't make promises
just yet.

Regards,
Tim

-- 
Tim Nikias v2.0
Homepage: http://www.digitaltwilight.de/no_lights


> Nobody replied.
>
> Is anyone here interested?
>
>


---
Outgoing mail is certified Virus Free.
Checked by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.502 / Virus Database: 300 - Release Date: 18.07.2003

Post a reply to this message

Regarding your massive multi-posting in several other groups (which has been
removed), please read:

Message-ID: <3AB04992.6DAA8B43@pacbell.net>
Date: Wed, 14 Mar 2001 20:48:18 -0800
From: Ken <tyl### [at] pacbellnet>
Subject: News Server Acceptable Use Policy
Newsgroups: povray.announce.frequently-asked-questions
Xref: news.povray.org povray.announce.frequently-asked-questions:46

    Thorsten


____________________________________________________
Thorsten Froehlich
e-mail: mac### [at] povrayorg

I am a member of the POV-Ray Team.
Visit POV-Ray on the web: http://mac.povray.org

Post a reply to this message

OK - I've read it - it doesn't say anyplace that I'm not allowed to
cross-post to the (few) places that I considered relevant for my
specific request - what exactly are you objecting to, and where in
the AUP should I be reading???


"Thorsten Froehlich" <tho### [at] trfde> wrote in message
news:3f1ecf38$1@news.povray.org...
> Regarding your massive multi-posting in several other groups (which has
been
> removed), please read:
>
> Message-ID: <3AB04992.6DAA8B43@pacbell.net>
> Date: Wed, 14 Mar 2001 20:48:18 -0800
> From: Ken <tyl### [at] pacbellnet>
> Subject: News Server Acceptable Use Policy
> Newsgroups: povray.announce.frequently-asked-questions
> Xref: news.povray.org povray.announce.frequently-asked-questions:46
>
>     Thorsten
>
>
> ____________________________________________________
> Thorsten Froehlich
> e-mail: mac### [at] povrayorg
>
> I am a member of the POV-Ray Team.
> Visit POV-Ray on the web: http://mac.povray.org

Post a reply to this message

Hi Tim,

Thanks for taking the time to share your thoughts
with me & the group!

I imagine your "main problem" (you mentioned below)
can be solved by simply making an intermediate program
which outputs a .POV input file - the "random" stuff can
then be hardcoded into the "POV" source, so there's no
need to work with complicated "rand()" formulas when
drawing stuff?

Anyhow - are you interested in having a go?  If you really
can't do "faces" - that's OK - so long as you can generate
"abstract" (or anything else) images which are sufficiently
distinct (given differing random input) that "Joe User" will
be able to choose one (out of 16 or so), and later remember
which one he chose - that's the essence of the system I'm
looking for.

As for rendering - the images will be presented to the user in
a grid (probably 16 at a time) - and will be small: postage-
stamp-sized, so I imagine a decent 3ghz pentium should be
able to draw these for moderate loads in near-real-time; and
if not, some kind of caching or pre-generation can easily be
worked out (or quality and image size could be reduced, or
more PCs thrown at the problem, or whatever...)

Chris.


"Tim Nikias v2.0" <tim### [at] gmxde> wrote in message
news:3f1ea8e6$1@news.povray.org...
> Well, it sounds like a fun thing to do, and it
> isn't than complicated really. To take a string,
> divide it into different parts, then use these
> as base for the rand()-Stream, that's not
> very difficult. Actually, I've done that for
> years. But it would involve POV-Ray
> as the raytracing engine.
>
> I see a problem in the precalculation of
> images. One could easily create tons of
> abstract images based on a random-stream.
> But generating images, like faces (though
> cartoonish like ones would probably be
> sufficient), isn't so easy to do. My main
> problem is that there would have to be a routine
> which will hand the string to the image-file.
>
> For example, I could easily create some scene
> which will take a certain file (like "keyword.inc")
> which has a string inside (with quotes at beginning
> and end, e.g. "thisisthekeyword") and render an
> image from that. I could even take the keyword,
> joggle it around and render a sequence of random
> images directly (if file I/O is supported so that I can
> keep an ever evolving random-number somewhere).
>
> Another problem lies in the speed of rendering
> with POV-Ray, which would mean that there'd
> be either waiting when loading the password-screen,
> or precalculation of images involved. You can
> email me for details if you like, but I can't make promises
> just yet.
>
> Regards,
> Tim
>
> -- 
> Tim Nikias v2.0
> Homepage: http://www.digitaltwilight.de/no_lights
>
>
> > Nobody replied.
> >
> > Is anyone here interested?
> >
> >
>
>
> ---
> Outgoing mail is certified Virus Free.
> Checked by AVG anti-virus system (http://www.grisoft.com).
> Version: 6.0.502 / Virus Database: 300 - Release Date: 18.07.2003
>
>

Post a reply to this message

" !" wrote:
> 
> OK - I've read it - it doesn't say anyplace that I'm not allowed to
> cross-post to the (few) places that I considered relevant for my
> specific request - what exactly are you objecting to, and where in
> the AUP should I be reading???

Actually i think Thorsten was more referring to the Cross-Posting
Guidelines:

Subject: Cross-Posting Guidelines for news.povray.org
Date: Sat, 22 Jan 2000 01:28:13 -0800
From: Ken <tyl### [at] pacbellnet>
Newsgroups: povray.announce.frequently-asked-questions

Christoph

-- 
POV-Ray tutorials, include files, Sim-POV,
HCR-Edit and more: http://www.tu-bs.de/~y0013390/
Last updated 17 Jun. 2003 _____./\/^>_*_<^\/\.______

Post a reply to this message

In article <3F27FED1.9CD4EC17@gmx.de> , Christoph Hormann 
<chr### [at] gmxde>  wrote:

> Actually i think Thorsten was more referring to the Cross-Posting
> Guidelines:
>
> Subject: Cross-Posting Guidelines for news.povray.org
> Date: Sat, 22 Jan 2000 01:28:13 -0800
> From: Ken <tyl### [at] pacbellnet>
> Newsgroups: povray.announce.frequently-asked-questions

Ups, you are of course right.  I grabbed the wrong message header.  Sorry
for the confusion.

    Thorsten

____________________________________________________
Thorsten Froehlich, Duisburg, Germany
e-mail: tho### [at] trfde

Visit POV-Ray on the web: http://mac.povray.org

Post a reply to this message