Sign In

Communications of the ACM

ACM TechNews

Researchers Teach AI to See Depth in Photos, Paintings

View as: Print Mobile App Share:
A classic image, top, and a depth map created from that image.

Researchers in the Computational Photography Lab of Canada's Simon Fraser University are teaching artificial intelligence how to determine depth from a single photograph.

Credit: Simon Fraser University

Researchers at Simon Fraser University (SFU) in Canada and Adobe Research used a process called monocular depth estimation to teach artificial intelligence to determine depth from a single photograph.

SFU's Sebastian Dille said, "We analyze the image and optimize our process by looking at the image content according to the limitations of current architectures. We give our input image to the network in many different forms, to create as many details as the model allows while preserving a realistic geometry."

SFU's Yagiz Aksoy added, "With the high-resolution depth maps that we are able to get for real-world photographs, artists and content creators can now immediately transfer their photograph or artwork into a rich 3D (three-dimensional) world."

From Simon Fraser University (Canada)
View Full Article


Abstracts Copyright © 2021 SmithBucklin, Washington, DC, USA


No entries found