Source Image: Sydney Long
Not the Mona Lisa
This piece requires some explanation.
It is an example of computer generative art from a program I have written called Darwin.
The image you see are 4,000 polygons arranged in a mathematical structure called a Voronoi tessellation. What this means is that this entire image is defined by only 4,000 points in the plane – not the millions of pixels you normally have when you take a photo of a scene.
The mathematics is very beautiful.
To create the image requires a clever computer algorithm. For this, I have developed a Genetic Algorithm which behaves very similar to Darwin’s theory of Evolution – continuous variation and selection of a population of individuals. Individuals are bred based on their relative fitness – the best survive to have children, the weakest die out. Over time the population gets stronger. Here – “fitness” is a mathematical score based on how different an individual is from the actual Mona Lisa image. I’m not the first to have used this approach.
The artistic idea I’m really exploring here are concepts around parsimony, minimalism, simplicity. Happiness for me as I grow older revolves around more simplicity, less complexity, being content with what I have. It is actually quite difficult to evolve an image that somehow captures the original aura of the Da Vinci masterpiece whilst constraining the number of allowed points in the plane. This image has 4,000 points. Images with less points require more work to be aesthetically interesting. Images with more points tend to look like the original image. I will be exploring this in more detail in time.
It’s also hard to avoid acknowledging the current heated debate around generative art provoked by tools such as Dall-E. I will make a few brief points, noting that this requires a longer discussion in a more suitable venue.
- the technology I have developed here is morally not the same as that used by Dall-E. I have not stolen large datasets of images nor I have I used massive amounts of computing power to build large AI models.
- more generally, generative art is just another medium to support artistic endeavours. When photography was first demonstrated nearly 200 years ago there was a similar outcry about the impacts to art and artists. It will have an important impact but it will settle into a cultural niche in society – it will not make void all other forms of artistic expression.
Kenya Woman
Photo Study
Stabilo pastel on Rivers paper
Merimbula NSW Winters Night Full Moon 2021
Winter, Moon and Water …
Black Mountain from Yaralumla Bay 2021
2020 Portrait Conte and Pastel