Welcome, visitor! [ Register | Loginrss

Disney Software Creates 3D Models from Any Photo

| 3D Scanners, News | July 25, 2013

Disney 3D Photo Software

The algorithm analyses the depth of objects in a photo using light rays

Researchers at Disney in Zurich have developed computer software which can create high resolution 3D models from digital photographs.

A specially designed algorithm is able to analyze the light and shade areas of an image to calculate the depth of the objects and creates a 3D reconstruction of the scene. This software is ideally suited to be used in Disney movies and video games but the resulting models can also be used to 3D print extremely accurate objects or scenes.

“Densely sampled light fields in the order of 10^9 light rays allow us to capture the real world in unparalleled detail… …an algorithm that leverages coherence in massive light fields by breaking with a number of established practices in image-based reconstruction.” Disney research paper.

The research is published in a paper called Scene Reconstruction from High Spatio-Angular Resolution Light Fields by authors Changil Kim, Henning Zimmer, Yael Pritch, Alexander Sorkine-Hornung and Markus Gross.

The software uses ‘a method for scene reconstruction of complex, detailed environments from 3D light fields’ using light rays to capture the ‘real world in unparalleled detail. Several high-resolution photos are required, taken from different viewpoints of the same scene, and the Disney software scans the images to analyze the different light rays around objects to work out their respective depths. Once the photos in a scene have been scanned they can be used to model lifelike scenes for use in movies.

Disney 3D Photo Software Process

After a photo is scanned and analyzed a 3D representation is produced

Previous software has not been successful at recreating an accurate scene from this data, and the Disney program is able to produce a more accurate model that laser scanning. the team says this is due to the software being able to make ‘reliable depth’ predictions around the outline of objects, as opposed to the object itself. If there is poor contrast between the light and shaded areas of an object the software can look in closer detail, pixel-by-pixel, using a ‘fine-to-coarse’ process. This makes the reconstructions highly detailed and starting the process with higher quality images leads to a more accurate reconstruction.

No Tags


Leave a Comment

Facebook RSS

Follow our 3D Printer Blog