Inferring the Existence of Geometric Primitives to Represent Non-Discriminable Data
In this thesis, we set out to find an algorithm that uses only geometric primitives to represent an input pointcloud. In addition to the problems faced in general primitive fitting, non-discriminable data presents additional data association challenges. We propose to address these challenges by esti...
Main Author: | |
---|---|
Other Authors: | |
Format: | Thesis |
Published: |
Massachusetts Institute of Technology
2022
|
Online Access: | https://hdl.handle.net/1721.1/139159 |
_version_ | 1811097034580557824 |
---|---|
author | Peraire-Bueno, James A. |
author2 | Roy, Nicholas |
author_facet | Roy, Nicholas Peraire-Bueno, James A. |
author_sort | Peraire-Bueno, James A. |
collection | MIT |
description | In this thesis, we set out to find an algorithm that uses only geometric primitives to represent an input pointcloud. In addition to the problems faced in general primitive fitting, non-discriminable data presents additional data association challenges. We propose to address these challenges by estimating the existence rather than parameters of geometric primitives, and explore various options to do so. We first explore a sampling-based Markov-Chain Monte-Carlo approach together with a ray likelihood model. We then explore a neural network approach and finish by presenting a method to make the Chamfer distance differentiable with respect to primitive existence. |
first_indexed | 2024-09-23T16:53:16Z |
format | Thesis |
id | mit-1721.1/139159 |
institution | Massachusetts Institute of Technology |
last_indexed | 2024-09-23T16:53:16Z |
publishDate | 2022 |
publisher | Massachusetts Institute of Technology |
record_format | dspace |
spelling | mit-1721.1/1391592022-01-15T03:02:08Z Inferring the Existence of Geometric Primitives to Represent Non-Discriminable Data Peraire-Bueno, James A. Roy, Nicholas Massachusetts Institute of Technology. Department of Aeronautics and Astronautics In this thesis, we set out to find an algorithm that uses only geometric primitives to represent an input pointcloud. In addition to the problems faced in general primitive fitting, non-discriminable data presents additional data association challenges. We propose to address these challenges by estimating the existence rather than parameters of geometric primitives, and explore various options to do so. We first explore a sampling-based Markov-Chain Monte-Carlo approach together with a ray likelihood model. We then explore a neural network approach and finish by presenting a method to make the Chamfer distance differentiable with respect to primitive existence. S.M. 2022-01-14T14:53:37Z 2022-01-14T14:53:37Z 2021-06 2021-06-16T13:26:57.960Z Thesis https://hdl.handle.net/1721.1/139159 In Copyright - Educational Use Permitted Copyright MIT http://rightsstatements.org/page/InC-EDU/1.0/ application/pdf Massachusetts Institute of Technology |
spellingShingle | Peraire-Bueno, James A. Inferring the Existence of Geometric Primitives to Represent Non-Discriminable Data |
title | Inferring the Existence of Geometric Primitives to Represent Non-Discriminable Data |
title_full | Inferring the Existence of Geometric Primitives to Represent Non-Discriminable Data |
title_fullStr | Inferring the Existence of Geometric Primitives to Represent Non-Discriminable Data |
title_full_unstemmed | Inferring the Existence of Geometric Primitives to Represent Non-Discriminable Data |
title_short | Inferring the Existence of Geometric Primitives to Represent Non-Discriminable Data |
title_sort | inferring the existence of geometric primitives to represent non discriminable data |
url | https://hdl.handle.net/1721.1/139159 |
work_keys_str_mv | AT perairebuenojamesa inferringtheexistenceofgeometricprimitivestorepresentnondiscriminabledata |