Post-focusing digital photography is nothing new for modern electronics, it is quite accessible to owners of a number of flagship smartphones - if they want to master it. Formally, the development of the University of California team belongs to the same area, but in practice - they create images that cannot be obtained with a classic camera. What they are immensely proud of, although they find it difficult to explain how to apply it in practice.
Computational Zoom technology allows you to manipulate all objects in the frame across the entire depth in violation of the laws of optics, because in a purely digital world they lose their power. Initially, the system creates a series of shots with a constant approach to the subject without correcting the focal length. Then the "3D-reconstruction" begins, the result of which is a virtual three-dimensional model of the shot scene with the calculation of a complete depth map for all objects in the frame.
And with the digital model, you can work as you please, creating collages from multi-perspective photographs. For example, sharply focusing on a person in front of an ancient temple while simultaneously showing the entire building behind him using wide-angle photography. Two mutually exclusive modes, if you used a conventional, albeit digital camera, an impossible thing until recently. And an ordinary function for Computational Zoom technology.
The technique is not ideal to combine two such different angles, the system corrects each layer in the image, which causes parasitic artifacts and the pros will easily calculate the "deception". But that is why it is a research technology - the University of California, together with Nvidia engineers, are just testing new image processing algorithms. It is assumed that a plug-in for popular software may appear on the basis of Computational Zoom in the future, but the developers are not sure that users will understand and love it.