Information Technology Reference
In-Depth Information
in a TV article prefixed by the phrase: “Is this some kind of hellish nightmare?” This
is surely an over reaction, but it serves to highlight the public's perceived fear over
the automation of human abilities, particularly in creative domains such as painting.
In response to this, we argue that established artists have more to fear from the latest
batch of art school graduates than from computers, because there will always be a
premium in art for human involvement.
It does not diminish the potential for computer generated art to engage audiences
in meaningful dialogues if we point out that many people appreciate art entirely
because of the human aspect: art lovers want to get to grips with the mind and
mood of the human behind the artwork. Hence, computer generated art may well
occupy different niches to that produced by people, and so there is little to worry
about in the automation of painting processes. More interestingly, the news team
described above also interviewed Ralph Rugoff—director of the Hayward Gallery
in London—and asked for his response to the notion of computer generated art. He
pointed out that while software is good at playing games with fixed rules, such as
chess, it is less obvious that computer programs can be playful in an artistic sense,
where there are no such rules and where cultural knowledge plays an important
role. Moreover, James Faure-Walker (another artist at the exhibition) pointed out
that most of the research in non-photorealistic graphics was essentially photograph
based, i.e., images are turned into painterly renditions. He added that this is rather a
naive approach, and noted that an idea rather than an image should be the motivation
for a piece of art. The issues raised by Rugoff and Faure-Walker led us to address the
(lack of) imaginative aspects of the software, and ultimately provided the inspiration
for the projects described under the scene invention and collage generation sections
below.
1.4.2 Emotional Modelling
Human emotion plays an enormous role in the visual arts. Often, paintings are pro-
duced in order to convey the emotion of the painter, or to evoke a particular emotion
in the viewer. In many cases, an important aspect of appreciating an artwork boils
down to understanding the emotions at play. When building an automated painter,
we have two choices with respect to emotional modelling. We could simply admit
that computers are not human, and therefore any attempt for the software to sim-
ulate emotions or model the emotions of viewers would be facile and doomed to
failure. In this case, we can emphasise that computer generated paintings can still
evoke emotions in viewers without necessarily modelling human emotions, and that
there are many other dialogues one can have with a painting other than trying to
understand the emotional state of the painter who produced it. Given that we argue
in the guiding principles given above that we should celebrate the difference be-
tween computers and people, it is certainly a defensible option to ignore emotion.
However, this would miss an opportunity to use pioneering work from the field of
affective computing, as described in Picard (
2002
), where software has been built to
