Pet Styklet Part 1
The desire to make a selfie Styklet stemmed from a hope that we could quickly and systematically produce customized, 3D-printable stickers. As I was printing out Eric’s iterations from his previous blog posts (you can read Part 1 and Part 2), I started thinking—is it something about depicting people in this way that makes it a bit terrifying, or is it really just the process of taking a 2D photo of a 3D thing and making it into a 2.5D monochrome sticker that makes anything scary in this way?
My boyfriend’s family has this extraordinarily handsome dog named Jet Li. Jet is an 8-year old German Shepard (a German-German Shepard, as his dad likes to emphasize), and I decided I’d work with a photo of him to try to create a sticker.
In his experiments, Eric had moved on to manually trace the contours of his face to create his 3D model, but I have significantly less patience and looking at the fur very quickly decided I would prefer not to do such a thing. Manual tracing wouldn’t scale for turning anyone’s pet photo quickly into a sticker anyways, so I was really hoping that I could just use some filters in GIMP (a free, opensource Photoshop alternative) and get a result to do this cute dog justice.
First, I took the photo of Jet in the park above, cropped it square, and turned it grayscale. Rhino uses black and white values to run the 3D displacement. Displacing the surface means moving the surface to create the texture. Essentially, in using Rhino’s displacement feature, I can assign a height to the white and black values (say, 0mm and 1mm), and the shades of grey will be elevated proportionately within those bounds. I then applied the image to two square plaques, one 1.5” by 1.5” and the other 2” by 2”. I decided to keep working with the 2” square, figuring that the image would have a better chance of being visible if printed bigger.
I also applied the edge detect and cartoon filters to the image, and used those images to displace the 2” by 2” plaque as well, hoping the defined edges would help the image turn out more crisply as well. Unfortunately, after printing, I could barely tell the two apart (see the image below). I guess the resolution of the printer is just too low, and so I went back to GIMP determined to make large enough features to print reliably.
After playing with more filters and the threshold settings, I got a more defined black and white image. I tried all the filters, from pixels to oilify to edge detect, and tried all kinds of combinations of all of them together, too. However, as soon as I used the images with the displacement map command in Rhino I could clearly see there was no hope to be had here. Poor Jet, he’s a large dog but with relatively small eyes, and in my experiments I realized they kept disappearing. The tongue also was proving problematic, coming out as an awkward, blank blob.
I felt the two circled ones might print okay, having larger element sizes than the first two squares I printed. You can see how all four prints turned out below:
I thought making the features larger would make for a more robust and defined print. Unfortunately, I guess they just came out looking smushed, so the first one without changes actually seems to be the most defined.
I gave it one more shot (shown below) before deciding to call it a day on this effort:
He came out looking a bit like a cartoon (it’s giving me Winnie the Pooh vibes), but it actually looks probably the closet to Jet! The tongue also at least has some dimension to it, too.
Ultimately, I think my conclusion from this experiment is that this isn’t an easy, automated way to take a photo and turn it into a cute 2.5D depiction. Though not necessarily scary, the prints just don’t look like much of anything. But, the last print gives me hope that if I worked manually, I could generate a cute Jet sticker. That’ll probably be another blog post, as well as maybe a future sticker for the shop!
Until next time - Alex