Voronoi

A natural extension of the Bitscape project was to explore more methods of re-presenting image data, additional re-presentations of the pixels and colours.

Initially image segmentation proved tempting as a way to start extracting shape and possibly even semantic data from images, however before long it proved too complex and time consuming a task to execute robustly on the wide range of source material.

Since features were already a part of the data set it was a small jump to consider using them within a graph, specifically a Voronoi graph.

With a computed Voronoi graph based on extracted features this algorithm traverses the planes one at a time, filling each plane with a colour based on the average of 5 samples within that region.

No additional tooling was built for this algorithm, it was worked into the Javascript/Node/PHP application and provided a suitable launch pad for much future work using the Voronoi graph.

Series A - Nature

Series B - Fashion

Detail (1:1)

Dimensions

1916 x 944

FAQ

Can this exact image be re-produced? No, these images were prouduced with tooling no longer available.

Are there any other images on this website that were generated by this algorithm? No

Is this a limited edition print? No

How can I buy a print? Prints of these images are not available.