Apple engineers have created a new AI tool called Depth Pro, which can turn flat 2D images into 3D by mapping their depth. They shared details about this tool in a research paper on the arXiv preprint server and on Apple’s Machine Learning Research page.
Humans naturally see depth by comparing the images from both eyes. Some video cameras do something similar to create 3D effects. But smartphones, which have only one camera, use extra hardware or software to add some sense of depth to pictures. Apple’s engineers have now developed a model that can create a complete depth map of an image using just the data from the picture itself, without needing extra camera details.
A depth map takes each pixel in an image and figures out how far it is from the object in the picture, helping to add a 3D effect. Apple’s team says their method creates clearer 3D images than traditional smartphone techniques.
They also claim that Depth Pro can create a depth map in just 0.3 seconds on a regular computer with a standard GPU. This fast processing could allow for real-time 3D imaging from a single-lens camera, which could be useful for robots, self-driving cars, and other applications that need quick mapping.
Source: techxplore
