ICRA 2007

Today is saturday and I'm coming back from my presentation on ICRA 2007, in Rome, Italy. It was the last day of the full conference...

The main positive point of the conference was a very good atmosphere of scientific knowledge... Two lectures have called my attention: Alain Bertroz's one and Moran Triverdi's one...

Both were completely different but extremelly correlated. The first one was about some psychological researches conducted to map the areas of the human being's brain and algorithms related to vision and motion and the latter one was about semantic information in vision applied to increase safety in vehicles...

As I said, completely different lectures, but where is the relation between them? Well, I think there are many researchers, particularly in vision comunity, that are very worried about pixels and low level information... Low level information is important but, in my opinion, the end of this chain is how to apply everything for human being's progress... And human being understand most of all semantically... and we are the most complete machine that ever exists...

Then, trying to relate brain, vision and semantics, one may realize high level algorithms in order to do accurately tasks like: object recognition, motion recognition, tracking, etc....

More than investigating just the same techniques that everyone utilizes, the probabilistical and semantical meaning of these tasks can lead us to a successfull project, I think myself...

To conclude this brief brainstorm, I really found out, after this conference, that we have to follow this pyramid low to high level methods to succeed in artificial vision...

Some pictures about the conference: