Saturday, July 29, 2006

The Divine Beauty Of Imaging Processing




Hallo!

The picture of a Swedish girl is the most applied to several tests related to image processing. But a thing that, maybe, people do not know is the history behind that picture, which was got from a 1972 Playboy issue.

The name of that girl is Lenna (or Lena) and there is the complete history of her famous picture inside image processing: The Lenna Story.

That link also tell us the meeting where Mrs Lenna met in 1997 with the imaging experts, when she attended the 50th Anniversary IS&T conference in Boston held in May, and this story was reported on Playboy and Wired. The IEEE Transactions on Image Processing and SPIE Journal Of Optical Engineering also had some stories about that picture and problems realted to copyrighted images.

Take a moment to read The Lenna Story, it is very very interesting.

If you would like to see the whole famous image of Lenna, click here (WARNING: She is completely nude!)

If you want to know how to seduce a woman like Lenna, please, see here. :)


Até Mais!

Marcelo

Wednesday, July 26, 2006

Tolstoy Knows Why!




Here you are words which some Evolutionary Computation researchers (and, also, researchers in general) should remember every time they get up from bed:



"I know that the majority of people, including those usually involved with problems of the greatest complexity, rarely accept even the simplest or the most obvious truth if this truth obliges them to admit the falseness of the conclusions which they got, perhaps, with great difficulty, because with these conclusions they delighted themselves when explaining them to their colleagues, they got proud when teaching them to others and that they had sewn, thread by thread, in the weaving of their lives."

Leo Tolstoy (September 9, 1828 - November 20, 1910)

Labels: , , , ,

Sunday, July 23, 2006

WCCI And CEC Blogging

Thursday, July 13, 2006

Interactive Evolutionary Computation




Hallo!

An interesting subject inside the Evolutionary Computation field is the so called Interactive Evolutionary Computation (IEC). According to Wikipedia, IEC "is a general term for methods of evolutionary computation that use human evaluation. Usually human evaluation is necessary when the form of fitness function is not known (for example, visual appeal or attractiveness) or the result of optimization should fit a particular user preference (for example, taste of coffee or color set of the user interface)".

Maybe, the main problem that IEC faces is the human fatigue. As computer evaluations of fitness functions are faster than those made by humans, that means that only few evaluations of the fitness function can be made through humans without causing to those the fatigue. Although there are some strategies which can be applied to avoid that, such as small number of fitness function evaluations and friendly human-computer interfaces.

Some examples of IEC are:



Karl Sims' Galápagos.

SBART developed by Tatsuo Unemi from Soka University, Japan.

Darwin Poetry.

Evolution through subjective selection by Michael Herdy from Technical University of Berlin (This last one is a flash animation which you can download to your computer and run in your browser.)



Até Mais!

Marcelo

Seekers + Solvers = InnoCentive



Hallo!

I found via Automating Invention an interesting post about "an exciting web-based community matching top scientists to relevant R&D challenges facing leading companies from around the globe. We provide a powerful online forum enabling major companies to reward scientific innovation through financial incentives". That community is called InnoCentive. See here: Inventing by and for the Masses.

Very Interesting!

Até Mais!

Marcelo

More Evolutionary Car Racing Videos




Hallo!

Julian Togelius posted new videos about his project on evolutionary car racing. This time he experimented a more complicated design for the track. The evolved car performed very well and could avoid all walls inside the track.

I only do not know the kind of Evolutionary Algorithm he applied. Last time he used Evolution Strategies.

It is very interesting Togelius' work on evolutionary cars!

Até Mais!!

Marcelo

Saturday, July 08, 2006

Professor Hans-Paul Schwefel talks to EvoNews




Hallo!

Some months ago I read an interesting interview with Professor Hans-Paul Schwefel (from the Chair of Systems Analysis (Ls11) - University of Dortmund) at EvoWeb. See here.

The interview has a historic, biographical, and interesting spirit related to Evolutionary Computation.

A small excerpt:

'Why,' he [Professor Schwefel] wonders, 'look for GA variants that provide a similar performance to ES, if Evolution Strategies have already provided such performance for more than twenty years?'


Well, that is a question which I also would like to find an answer! :)

In the picture above there are from left to right: Professor Xin Yao, Professor Hans-Paul Schwefel and Professor Zbigniew Michalewicz.

Até Mais!

Marcelo

Sunday, July 02, 2006

CMA-ES




Hallo!

First of all: N.F.L.!!!

Nikolaus Hansen from Bionics & Evolution Technique at TUB (Technical University of Berlin) reports the results achieved through a CMA-ES (Covariance Matrix Adaptation Evolution Strategy) in the 2005 IEEE Congress on Evolutionary Computation: Session On Real Paramater Optimization.

Here you have a complete list of all algorithms submitted.

The results are very interesting! :)

Até Mais!!

Marcelo
Charles Darwin Has A Posse Check Google Page Rank