Help I accidentally a wigglegram

By leveraging perceptual hashing, a developer automated the process of finding similar photos in their library and stitching them into 3D wigglegrams. The tool is now open-source and available for anyone to try.
Do you know what a wigglegram is?
It is a kind of stereo image you make by looping frames together, like as a GIF.
The effect is quite convincing.
I am something of an indecisive photographer and when I like an angle I will take a lot of frames, from slightly different angles etc., looking for "the shot". And since I am also a bit of a hoarder I never clear out my camera roll.
"Same shot from different angles"? You know what that sounds a bit familiar.
Sure enough my phone is full of wigglegrams that I took by accident. Years' worth, waiting for me to sit down and stitch them together.
Or, perhaps, for something to stitch them together. It occurred to me last weekend that I can use perceptual hashing - what TinEye (et al.) uses for reverse image search - to try and find runs of similar images and pull them out from my library automatically. So I wrote a little script to hash all my pictures:
Hashing is quick but downloading photos from iCloud is not.
The result is a hash that - unlike a cryptographic function like sha1 - will share more bits with hashes of similar-looking images than with dissimilar ones. We can use that to calculate the hamming distance between pairs of images and find a threshold:
And extract pairs:
And hundreds of wigglegrams spew forth.
A few of them I am guilty of taking intentionally. But most are true accidents. As such many of them come out as less "stereoscopic" and more "kinescopic" - like little unintentional movies.
Animals are a natural fit for the concept, unpredictable as they are:
Design-work also. (I am always indecisive.)
And sculpture:
What fun. I have the script up on Github if you want to play with it - it'll work on your iCloud photos library if you're on a Mac, or you can point it at a directory of pictures otherwise.
Cheers~
spacer
more juicy text for you to read
Source: Hacker News












