2019 AI Review: 3D Ken Burns Effect.
Vladimir Alexeev
Autor, Forscher, Künstler, Speaker, KI-Berater (Generative KI). Digital Experience Specialist - @ DB Schenker. OpenAI Community Ambassador. Digital Resident. Ich erforsche kreative Mitarbeit von Mensch + Maschine
What’s this about: this model, developed by Simon Nikolaus converts single images into tracking shots. The model recognizes the backgrounds, simulates depth, fills the missing areas with content-sensitive inpainting, adds new angles — short: with a single image, you can produce a spatial 3D video footage.
Simply put: 3D movement video from a single image.
Try it out: Colab Notebook by Manu Romero.
Here are some of my experiments (in twitter thread):
I wrote in details about this topic:
Index of Series "2019 AI Review".
Full essay "14 Deep and Machine Learning Uses that made 2019 a new AI Age."