Motion Structures: about | gallery | software


'Motion structures' is an approach to explore and interact with time-based media, such as film, video, and motion graphics in a different way. We want to represent spatial and temporal transformations in a moving image sequence as a 3D shape. The transformation from a 2D image sequence to a 3D shape are done automatically by a custom software we wrote.

Such a shape encodes the changes in the positions of the objects in a frame, their movements, and spatial and temporal relations. Our goals are both to contribute to ongoing research in the artistic possibilities of 3D computer generated forms, and also offer a novel way to visualize moving image sequences. The outcome of the process are 3D digital objects which can be represented as images and also as 3D printed forms.

We want to thank Lev Manovich for his valuable suggestions and Gunalan Nadarajan for his comments at re-new 2013.


Motion structure generated from Charles Csuri’s “Hummingbird” (1967). 02:10. minutes, 1295 frames. This is a forward angle perspective captured from ImageJ 3D Viewer

Motion structure generated from Peter Weibel’s “Endless Sandwich” (1969). 00:38 seconds, 378 frames. This is a forward angle perspective captured from ImageJ 3D Viewer

Motion structure of Bill Viola’s “Acceptance” (2008), generated in 2013. 02:03 minutes, equivalent to 1231 frames. This is a reverse angle perspective captured from ImageJ 3D Viewer, which means the first figure to the left is the last frame shown in the original video

In 2012 we have also started to 3D print motion structures, which would be called 'motion object'. The first of this series was created from the main title sequence of The Game of Thrones TV series.

The average brightness of the 110 frames that constitue the 5-second fragment is:
A perspective of the 'motion structure' and the resulting 'motion object':

More images are available and updated from the Flickr album and the Flickr group.


We build on top of the open-source software ImageJ (developed by Wayne Rasband at the National Institute of Mental Health, Maryland), which is mainly used for scientific imagery practices. We have created an ImageJ plugin that performs a series of operations that take as input a folder containing a sequence of frames. The script then automatically converts the images into 8-bit format, subtracts background, and finally runs the 3D viewer. From the 3D Viewer it is possible to save the result as static image, as 360-degree rotation movie, and to export it as a mesh surface.

Download our ImageJ macro here:
motionStructure-v2.txt (v. 0.2)
imageToObject.txt (v. 0.1)
Right-click the hyperlink and save it your hard drive. Then you can run it from ImageJ, from the menu Plugins / Macros / Run...

Project started in 2011 by Everardo Reyes-Garcia Paris, 2016