Thinking outside the box: Generation of unconstrained 3D room layouts
We propose a method for room layout estimation that does not rely on the typical box approximation or Manhattan world assumption. Instead, we reformulate the geometry inference problem as an instance detection task, which we solve by directly regressing 3D planes using an R-CNN. We then use a varian...
| প্রধান লেখক: | , , |
|---|---|
| বিন্যাস: | Conference item |
| প্রকাশিত: |
Springer
2019
|
| _version_ | 1826312946324602880 |
|---|---|
| author | Howard-Jenkins, J Li, S Prisacariu, V |
| author_facet | Howard-Jenkins, J Li, S Prisacariu, V |
| author_sort | Howard-Jenkins, J |
| collection | OXFORD |
| description | We propose a method for room layout estimation that does not rely on the typical box approximation or Manhattan world assumption. Instead, we reformulate the geometry inference problem as an instance detection task, which we solve by directly regressing 3D planes using an R-CNN. We then use a variant of probabilistic clustering to combine the 3D planes regressed at each frame in a video sequence, with their respective camera poses, into a single global 3D room layout estimate. Finally, we showcase results which make no assumptions about perpendicular alignment, so can deal effectively with walls in any alignment. |
| first_indexed | 2024-03-06T21:02:52Z |
| format | Conference item |
| id | oxford-uuid:3b72f44d-0cdd-4d3e-ae07-6cc897d19acf |
| institution | University of Oxford |
| last_indexed | 2024-09-25T04:03:16Z |
| publishDate | 2019 |
| publisher | Springer |
| record_format | dspace |
| spelling | oxford-uuid:3b72f44d-0cdd-4d3e-ae07-6cc897d19acf2024-05-16T12:07:04ZThinking outside the box: Generation of unconstrained 3D room layoutsConference itemhttp://purl.org/coar/resource_type/c_5794uuid:3b72f44d-0cdd-4d3e-ae07-6cc897d19acfSymplectic Elements at OxfordSpringer2019Howard-Jenkins, JLi, SPrisacariu, VWe propose a method for room layout estimation that does not rely on the typical box approximation or Manhattan world assumption. Instead, we reformulate the geometry inference problem as an instance detection task, which we solve by directly regressing 3D planes using an R-CNN. We then use a variant of probabilistic clustering to combine the 3D planes regressed at each frame in a video sequence, with their respective camera poses, into a single global 3D room layout estimate. Finally, we showcase results which make no assumptions about perpendicular alignment, so can deal effectively with walls in any alignment. |
| spellingShingle | Howard-Jenkins, J Li, S Prisacariu, V Thinking outside the box: Generation of unconstrained 3D room layouts |
| title | Thinking outside the box: Generation of unconstrained 3D room layouts |
| title_full | Thinking outside the box: Generation of unconstrained 3D room layouts |
| title_fullStr | Thinking outside the box: Generation of unconstrained 3D room layouts |
| title_full_unstemmed | Thinking outside the box: Generation of unconstrained 3D room layouts |
| title_short | Thinking outside the box: Generation of unconstrained 3D room layouts |
| title_sort | thinking outside the box generation of unconstrained 3d room layouts |
| work_keys_str_mv | AT howardjenkinsj thinkingoutsidetheboxgenerationofunconstrained3droomlayouts AT lis thinkingoutsidetheboxgenerationofunconstrained3droomlayouts AT prisacariuv thinkingoutsidetheboxgenerationofunconstrained3droomlayouts |