Fast Content Placement and Alignment in 3D Scenes
We present a system for fast content placement on flat surfaces in 3D scenes that leverages plane detection and computer vision. For content creators, placing and aligning objects is often a time-consuming process. Especially with AR, this problem becomes more prevalent as virtual objects should be...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
LibraryPress@UF
2022-05-01
|
| Series: | Proceedings of the International Florida Artificial Intelligence Research Society Conference |
| Online Access: | https://journals.flvc.org/FLAIRS/article/view/130686 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | We present a system for fast content placement on flat surfaces in 3D scenes that leverages plane detection and computer vision. For content creators, placing and aligning objects is often a time-consuming process. Especially with AR, this problem becomes more prevalent as virtual objects should be aligned to the real world. For certain objects that should be aligned to planar regions, e.g, a virtual painting being placed on a real wall, the scene (wall) implies the rotation of the object. However, placing a virtual painting usually requires manually modifying its 3D coordinates and manually rotating it to exactly match the wall. In this work, we propose the use of plane detection to speed up the placement and alignment process for every object whose relation to surfaces has been defined, e.g., images or posters. We do this by detecting approximately planar regions in our scene and calculating planes for them. Then to place an object, one can simply click a point in the scene to select the location and with a raycast, the actual point in the 3d space is found. The object is then automatically positioned and aligned by moving it to the closest point on the plane of the region and matching the plane's rotation. Furthermore, we apply object detection and segmentation to the textured planes to gain a semantic understanding of the scene. This enables automatic placement suggestions and assistance when aligning objects to visual markers, e.g., snapping the corners of paintings to those of windows in AR. |
|---|---|
| ISSN: | 2334-0754 2334-0762 |