September 4, 2018

Animation created using artificial intelligence and deep learning in a technique called ‘artistic style transfer’.

Neural models of human visual perception are used to transfer the visual style of a painting or photograph onto another image.

In this video I’m taking visual styles such as computer fractals, abstract photography, sci fi art and HD wallpapers and transferring them onto repeating GIF loops – which are originally just simple 3D animations with no texture or color.

Everything was rendered from open source code in Google’s Colab using their GPU runtime support. GPU processing power is an essential requirement in all deep learning projects.

Artistic style transfer part of an exciting new branch of AI based art, in what could tentatively be called ‘neuralism’.

Animation & Music by Glenn Marshall

I’ve been further experimenting with neural style transfer techniques – and have always wanted do something with Buddhist art. I’ve always been uninspired and disappointed by modern Buddhist art, I guess though these are mere worldly expressions that don’t really aid one’s quest on the path. However, I feel obliged to try my hand. The ‘art’ in neural style transfer is finding the right source imagery to best suit / express want you want. In these examples – I used a flame fractal – full of repeating geometry, sweeping curves, and lots of recursion. When applied to some Buddhist statues and iconography – it gives it a certain sophisticated, scientific, mathematical feel – yet hand made with love and care. Buddhism is a science of the heart and mind.

Now, having been pleased with these results – I turned my mind to something quite random. I like subverting things, particularly imagery. I thought what would pornographic imagery look like through these filtered changes. So I applied the style transfer using the same flame fractal – and the result is kind of horrifying and creepy. And then I realised that this is very much connected to the Buddhist path – where it is a common meditation practice to contemplate the foulness and disgust of the human body in order to release one from desire and lust – more of a practice among dedicated monks than of lay people, understandably.
Anyway – it got me thinking about the neural theme throughout – the process is about understanding human perception, and changing it – very apt…

Neural Art

July 24, 2018

Ever restless and never wanting to keep doing the same kind of thing, I’m currently delving into ‘neural art’. I thrive on the new, and learning new things, doing new things, with new tools and new techniques – that, historically has been always been my inspiration.

By neural art, I mean machine learning, artificial intelligence and in my first quick experiments – exploring neural style transfer. Neural style transfer is an AI process of looking at one picture, and transferring its visual aesthetic style onto another picture.

There’s lots of open source code for this available on the web. I spent time finding the best code, installing it and setting up my laptop with GPU / CUDA enhanced acceleration (absolutely essential for this kind computationally intense work).

Here’s my first results, it takes time, like anything, tweaking parameters, seeing what works, why it works.. learn, process, see, repeat..

For the source images with which to ‘extract’ a visual style, I chose local Belfast artist John Luke (1906 – 1975) – and transferred his style onto photos of some iconic landmarks in the city – the whimsical notion of how a dead artist would new perceive and paint his modern day birth city if he were alive.


You can now purchase digital limited editions of my work online at

Starting with ‘Particle Man’, you can now own a special hi res 4k, digitally signed and authenticated version – visit sedition to view and buy the work –




My animation, Particle Man, has been nominated for The Lumen Prize.  There is a public vote in the online gallery. You can vote there, and need to register first.


If you have the time, and like my work, your vote would be very much appreciated!