PORTR(AA) ~ luluxxx



a curated semi-generative , neural-network assisted , automatic portrait generator.

The system inputs a set of photographic images , which are abstracted locally through various digital techniques while being transformed through a neural network using style-transfer techniques. Then it goes through a super-resolution and color-grading process to generate a final output. (read more about the project)

The final collection consists of 2000 pictures at 1600x2365 pixels resolution. You can use the [random] button to explore the collection

Trait : style

Style Transfer uses neural network technology and allows the creation of new images by using an algorithm to redraw one image using the stylistic elements of another (the style). Portr(aa) uses 7 styles. [ST0 .. ST6]. you can use the [same style] button to explore a particular style. Distribution of styles across the full collection is shown here. (read more)

style distribution

Trait : source

i'm using a curated set of images [sources] as inputs of the system. i wasn't sure at the beginning how much i would need to have enough variety in the output. finally there are 110 different sources used [SRC]. you can use the [same source] button to explore a particular source . Distribution of sources across the full collection is shown here. (read more)

source distribution

Trait : abstract

i experimented with various algorithms to generate abstraction in parts of the pictures. it includes directional and fragment blur , point based in painting , fractal generation , different triangulation techniques and also various local and global geometric transformations.i kept track of all parameters all along the experiment and tried to generate a global abstraction parameter for each results. all min/max significative parameters value were normalized between [0,100] then weighted together. finally the abstraction [ABS] ranges between 15 to 87. you can use the [similar abstract] button to look for more or less abstraction. Distribution of abstraction across the full collection is shown here. (read more)

abstract distribution

Trait : cartouche

I wanted to introduce a purely random trait. so at the very end of the pipeline i added this little box under the picture , print a color palette from the picture and the values of the different traits. black and white backgrounds are random while orange are a selection of early results that i reincorporated in the collection near the end [1039-1552]. Distribution of cartouche across the full collection is shown here. (read more)

cartouche distribution
DigitalOcean Referral Badge