Information Technology Reference
In-Depth Information
the logical next step is the development, through collaboration, of AJSs that en-
compass the metrics used by the different research groups. These could lead, for
instance, to an international research project where several research groups build a
common AJS. Some of the groups could propose metrics, others design the classi-
fier, and so on. Using the validation approaches proposed in this chapter (and future
research in this area) it becomes possible to validate the classifier and compare the
results with previous approaches. Moreover, due to the numerical nature of the val-
idation approach, it is possible to identify relevant metrics in the classifier for the
tasks considered.
AJSs can be valuable for real life applications, including:
•
Image Classification
—e.g., discriminating between professional and amateur
photos, paintings and photos, images that are interesting to a particular user, etc.
•
Image Search Engines
—which could take into account user preference, or stylis-
tic similarity to a reference image or images.
•
Online Shopping
—the ability to recognise the aesthetic taste of the user could be
explored to propose products or even to guide product design and development.
The development of AJSs can also play an important role in the study of aesthet-
ics, in the sense that the ability to capture aesthetic preferences of individuals and
groups may promote a better understanding of the phenomena influencing aesthetic
preferences, including cultural differences, training, education, trends, etc.
More importantly, the creation of systems able to perform aesthetic judgements
may prove vital for the development of computational creativity systems. For in-
stance, the development of an AJS that closely matches the aesthetic preferences of
an individual would open a wide range of creative opportunities. One could use such
an AJS in conjunction with an image generation system to create custom made “ar-
tificial artists” that would be able to create artworks which specifically address the
aesthetic needs of a particular person. These systems could change through time,
accompanying the development of the aesthetic preferences of the individual and
promoting this development. They could also be shared between people as a way of
conveying personal aesthetics, or could be trained to match the aesthetic preferences
of a community in order to capture commonality. These are vital steps to accomplish
our long term goal and dream: the development of computational systems able to
create and feel their art and music.
Acknowledgements
The authors would like to thank the anonymous reviewers for their con-
structive comments, suggestions and criticisms. This research is partially funded by: the Span-
ish Ministry for Science and Technology, research project TIN2008-06562/TIN; the Portuguese
Foundation for Science and Technology, research project PTDC/EIA-EIA/115667/2009; Xunta de
Galicia, research project XUGA-PGIDIT10TIC105008-PR.
References
Arnheim, R. (1956).
Art and visual perception, a psychology of the creative eye
. London: Faber
and Faber.
