Title:
Robust approaches and optimization for 3D data

dc.contributor.advisor Isbell, Charles L.
dc.contributor.author Sawhney, Rahul
dc.contributor.committeeMember Boots, Byron
dc.contributor.committeeMember Vela, Patricio A.
dc.contributor.committeeMember Christensen, Henrik I.
dc.contributor.committeeMember Li, Fuxin
dc.contributor.department Interactive Computing
dc.date.accessioned 2019-05-29T13:56:50Z
dc.date.available 2019-05-29T13:56:50Z
dc.date.created 2018-05
dc.date.issued 2018-04-06
dc.date.submitted May 2018
dc.date.updated 2019-05-29T13:56:50Z
dc.description.abstract We introduce a robust, purely geometric, representation framework for fundamental association and analysis problems involving multiple views and scenes. The framework utilizes surface patches / segments as the underlying data unit, and is capable of effectively harnessing macro scale 3D geometry in real world scenes. We demonstrate how this results in discriminative characterizations that are robust to high noise, local ambiguities, sharp viewpoint changes, occlusions, partially overlapping content and related challenges. We present a novel approach to find localized geometric associations between two vastly varying views of a scene, through semi-dense patch correspondences, and align them. We then present means to evaluate structural content similarity between two scenes, and to ascertain their potential association. We show how this can be utilized to obtain geometrically diverse data frame retrievals, and resultant rich, atemporal reconstructions. The presented solutions are applicable over both depth images and point cloud data. They are able to perform in settings that are significantly less restrictive than ones under which existing methods operate. In our experiments, the approaches outperformed pure 3D methods in literature. Under high variability, the approaches also compared well with solutions based on RGB and RGB-D. We then introduce a robust loss function that is generally applicable to estimation and learning problems. The loss, which is nonconvex as well as nonsmooth, is shown to have a desirable combination of theoretical properties well suited for estimation (or fitting) and outlier suppression (or rejection). In conjunction, we also present a methodology for effective optimization of a broad class of nonsmooth, nonconvex objectives --- some of which would prove problematic for popular methods in literature. Promising results were obtained from our empirical analysis on 3D data. Finally, we discuss a nonparametric approach for robust mode seeking. It is based on mean shift, but does not assume homoscedastic or isotropic bandwidths. It is useful for finding modes and clustering in irregular data spaces.
dc.description.degree Ph.D.
dc.format.mimetype application/pdf
dc.identifier.uri http://hdl.handle.net/1853/61103
dc.language.iso en_US
dc.publisher Georgia Institute of Technology
dc.subject 3D
dc.subject Geometry
dc.subject Robust
dc.subject Optimization
dc.subject Association
dc.subject Retrieval
dc.subject Robust loss
dc.subject Viewpoint invariance
dc.subject RGB-D
dc.subject Depth image
dc.subject Point cloud
dc.subject Geometric description
dc.subject Matching
dc.subject Correspondence
dc.subject Registration
dc.subject Reconstruction
dc.subject Surface
dc.subject Patch
dc.subject Segment
dc.subject Superpixel
dc.subject Majorization minorization
dc.subject Outlier rejection
dc.subject Model fitting
dc.subject Estimation
dc.subject Nonlinear least absolute deviations
dc.subject Mean shift
dc.subject Mode seeking
dc.subject Segmentation
dc.subject Hierarchical
dc.subject Geometric diversity
dc.subject Nonsmooth
dc.subject Nonconvex
dc.subject Edit distance
dc.subject Damerau Levenshtein
dc.subject Proximal algorithms
dc.subject Determinantal point process
dc.subject Fisher vector
dc.subject Feature space
dc.subject Variational factorization
dc.subject M - estimation
dc.subject Structured estimation
dc.title Robust approaches and optimization for 3D data
dc.type Text
dc.type.genre Dissertation
dspace.entity.type Publication
local.contributor.advisor Isbell, Charles L.
local.contributor.corporatename College of Computing
local.contributor.corporatename School of Interactive Computing
relation.isAdvisorOfPublication 3f357176-4c4b-402c-8b61-ec18ffb083a6
relation.isOrgUnitOfPublication c8892b3c-8db6-4b7b-a33a-1b67f7db2021
relation.isOrgUnitOfPublication aac3f010-e629-4d08-8276-81143eeaf5cc
thesis.degree.level Doctoral
Files
Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
SAWHNEY-DISSERTATION-2018.pdf
Size:
74.72 MB
Format:
Adobe Portable Document Format
Description:
License bundle
Now showing 1 - 1 of 1
No Thumbnail Available
Name:
LICENSE.txt
Size:
3.87 KB
Format:
Plain Text
Description: