Ever since I created the pixel portrait of Kylie by combining 4 existing pictures, I have wanted to try and automate that process using code, to see how that would work. This is the first experiment to use the processing language to try and achieve that.
I thought it would be fun to use a different model for every experiment I tried, so I collected images of model Candice Swanepoel to use as source material for this one.
I wrote a program that would randomly select 4 images from the collection, crop the middle of the image to a square, resize them all to the same size, then start swapping the pixels in the same square grid as used for the Kylie portrait: top-left a pixel from the first image, top-right a pixel from the second, bottom-left a pixel from the third image, and bottom right a pixel from the fourth image, then repeat for the next 2 x 2 block of pixels.
Because the images were in no way lined up, as they are in my photoshop-based versions of these pixel portraits, the result is unexpected and less coherent. I generated 1000 portraits this way and selected these 8 as the resulting portrait series. I chose 8 portraits because 8 bits is a byte, I decided to call these computer-code aided portraits ‘bytes’.
The result here is pretty far removed from the Kylie example, they look surreal / dreamlike and not as portraits as all. Still I like how, in the same way as the Kylie portrait does, you see different faces of Candice at the same time, resulting in an overall general idea that is more comprehensive than the individual pictures.