How VEO helps Fukuda Art Museum create “moving paintings”

How VEO helps Fukuda Art Museum create "moving paintings"

Use of technology to explore art in new ways and expand its accessibility comes with a technical and curatorial challenge: how to move beyond the individual, fixed framework. Any static masterpiece, historical photograph or archive artifact not only has a captive moment, but the potential of a story that continues beyond the edge of the image. The recent initiatives from Google Arts & Culture focused on solving this through VEO – Google’s advanced video cleaning model – creating a framework for animation of static visual assets from moving archives with the Harley Davidson Museum and moving paintings with Fukuda Art Museum in Japan, freeing new storytelling options for curators.

The technical breakthrough is VEO’s ability to extrapolate plausible movement from a fixed composition. The model bridges the space between a static input and hundreds of video frames and create a temporal context that feels intentional.

Google Arts & Culture has developed two different operational conditions for this process, each designed to answer another kind of visual question:

1. Animation mode: reveals the narrative

This condition is driven by expert -defined input. Curators, in partnership with Google Teams, identify the implied energy within the stage – the falling rain, the passing traveler, the fluttering banner – to translate these into specific movement vectors. VEO then synthesizes this input to make a continuous high-definition sequence. The result is a controlled narrative that unfolds that transforms the implicit story of the composition into an explicit visual event inviting viewers to analyze the moment after The artist’s brush left the canvas.

2. Photorealistic state: Imagine the source

This condition addresses the question of what could actually have been here. It focuses purely on contextual and environmental plausibility. VEO uses the static image as a visual seed to generate a high-faith video that simulates the photo-realistic world that may have inspired the original view. This computer vision process predicts a stable, temporary coherent environment from a single-frame signal that essentially offers a digital window into the reality that preceded the artistic interpretation.

Using VEO, we have the prototype a way to increase digital archives for dynamic, analyzing-clear assets, tossing a new path to both preservation and visual storytelling.

Explore moving paintings at goo.gle/moving-males on Google Arts & Culture.