I tried to easily create a high-precision 3D image with one photo [0]. (Confirmed how to capture the space, put a net)
Overview
** One photo ** ** High-precision 3D image ↓ Easy creation **. github is below. https://github.com/vt-vl-lab/3d-photo-inpainting
Paper "3D Photography using Context-aware Layered Depth Inpainting" Meng-Li Shih1 and others are as follows. https://arxiv.org/pdf/2004.04727.pdf
Here, in order to confirm how to grasp the space,
- Calculate the depth of the original image
- From the state where the depth has already been calculated, put a ** grid ** in the original image and perform 3D processing.
The result is as follows. (The source of the still image before processing is https://pixabay.com/ja/)
⇒ ** I feel that the horizontal and vertical planes are well captured **.
Summary
** I think the point of the above github technology is in another place **, so this is a consideration of ** the base technology used there ** (beginners, interested in that) said).
The following are works that probably exceed the level of the author of github. A minor failure of technology was buried in the difficulty of separating pizza?
(Including the following related articles, the articles are listed in the order of [1] ⇒ [2] ⇒ [3] ⇒ 0.) If you have any comments, please let us know.
Relation
** Understand how to work in this article ** For more information, please refer to the following.
I tried to easily create a high-precision 3D image with one photo [1]. (Depth can now be edited in PNG.) I tried to easily create a high-precision 3D image with one photo [2]. (Try processing depth with numpy) I tried to easily create a high-precision 3D image with one photo [3]. Only depth without permission feat. Intel-isl.
Recommended Posts