.png){kind=tracking}
'The Bibites'
Network changes, and benchmarking optimization.
Made in Unity engine with C#.
Created as a client project for my masters thesis.
​
This is an artificial life simulation, where organisms are controlled by neural networks with augmenting topologies, over generations these organisms breed, creating new connections in the brain, and getting smarter over time, learning to navigate their environment with plenty of emergent behaviour.
​
This is still in completion as a part of my masters thesis.
Given the artificial life simulation, 'The Bibites', I have created a benchmarking system to automatically test different breeding and mutation systems of the organisms neural networks in order to have a system that has the fastest learning speed possible in order to make the simulation more viable as a commercial product.
​
​
​
Style Transfer - Style Selection.
Made with Python.
Created as my dissertation project.
​
Given a larger 'style set' (a collection of images of a particular style; IE a collection of Van Goughs artwork, or a selection of printscreens from a specific game), and an input content image, the system will find the style from that set which is the closest fit to your content image to apply the style transfer to it.
​
Choosing a style with similar detail levels, colour pallets, and objects in them to the given content image makes for a much more appealing and accurate style transfer
​
This builds upon existing style transfer methods by allowing it to be used in more commercial applications, such as apps where you make your photos look like they were set in your favorite game.
​
​
​
Pictobot.
Made with Phython.
Created as an artefact for a university project.
​
Created as a touchscreen application for the raspberry pi. For this it needed to be heavily optimized to run in real-time on the limited hardware.
The network was trained using Google's 'Quick, Draw!' database.
Users draw images into the application, which are compressed, converted, and fed into the network for it to guess what you have drawn.
​
​