Wouter Deenik

Email: w.deenik[at]student.utwente.nl

Final project assignment

Title:Intelligent interactive evolution of abstract animations
Completed: No
Mentor:Mannes PoelDennis Reidsma
Generative art algorithms are normally constructed with certain rules and constraints in mind that are carefully tuned to create an algorithm that generates pleasing art. This method ensures that the algorithm will not produce unpleasant art, but also reduces the freedom the algorithm has in the styles of art it generates; often most of the generated art looks alike.

This assignment is about finding out if it is possible to create an algorithm that has lots of freedom in the art it generates, while also avoiding the generation of unpleasant art. A system will be developed that can generate abstract animations and several methods of generation and selection will be applied and tested. Some of these methods will be tested in an actual user test to get a good idea about the effects these methods have on the quality of the generated art.

Capita selecta and Research Topics assignment

Title:Generative art
Completed: Yes
Mentor:Dirk Heylen
Generative art or content is used often in a range of fields. However, generative algorithms are often predetermined and do not have a lot of freedom, to make sure the generated content is good and fits certain requirements. This of course limits the amount of new content that may be generated by the algorithm. It would be very interesting if the generative system can figure out for itself whether or not the generated content is good or not and use this knowledge for generating new content.

For my Capita Selecta I will research the first step in this direction: figure out what characteristics of abstract animations determine people's appreciation of these animations. Method: After a small initial (literature) research, a range of animations will be made covering several combinations of key characteristics of abstract animations. Test subjects will be asked to 'rate' these animations.