國立高雄大學圖資館 |

Language: English

Back

Hybrid discriminative-generative met...

Dilsizian, Mark.

Hybrid discriminative-generative methods for human pose reconstruction from monocular imagery.

Record Type:	Electronic resources : Monograph/item
Title/Author:	Hybrid discriminative-generative methods for human pose reconstruction from monocular imagery.
Author:	Dilsizian, Mark.
Published:	Ann Arbor : ProQuest Dissertations & Theses, 2016
Description:	113 p.
Notes:	Source: Dissertation Abstracts International, Volume: 78-04(E), Section: B.
Notes:	Adviser: Dimitris Metaxas.
Contained By:	Dissertation Abstracts International78-04B(E).
Subject:	Computer science.
Online resource:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=10291949
ISBN:	9781369350999

Hybrid discriminative-generative methods for human pose reconstruction from monocular imagery.
Dilsizian, Mark.

Hybrid discriminative-generative methods for human pose reconstruction from monocular imagery. - Ann Arbor : ProQuest Dissertations & Theses, 2016 - 113 p.

Source: Dissertation Abstracts International, Volume: 78-04(E), Section: B.

Thesis (Ph.D.)--Rutgers The State University of New Jersey - New Brunswick, 2016.

Estimating 3D human pose from monocular images is an important and challenging problem in computer vision with numerous applications including human-computer interaction, human activity recognition, biomechanical analysis, and security. Existing state-of-the-art methods utilize statistical learning models that are inherently limited because they require sufficient training data that does not often include uncommon pose articulations or subject proportions. In addition, these methods often return global average case results and cannot easily leverage anthropomorphic, kinematic, and other physics-based constraints. However, a prior model-based search can be computationally prohibitive due to the combinatorially large set of plausible joint combinations. We combine statistical learning-based approaches with a prior part-based model into a hybrid constrained optimization that leverages strengths of both approaches. The method guarantees a plausible human pose while also resolving local ambiguities among body parts. Qualitative evaluation of the proposed methods on human pose datasets show improvement in reconstruction accuracy compared to current state-of-the-art methods.

ISBN: 9781369350999Subjects--Topical Terms:

199325
Computer science.

Hybrid discriminative-generative methods for human pose reconstruction from monocular imagery.
LDR:02172nmm a2200301 4500 001 502185
005 20170619070735.5
008 170818s2016 ||||||||||||||||| ||eng d
020 $a 9781369350999
035 $a (MiAaPQ)AAI10291949
035 $a AAI10291949
040 $a MiAaPQ $c MiAaPQ
100 1 $a Dilsizian, Mark. $3 766257
245 1 0 $a Hybrid discriminative-generative methods for human pose reconstruction from monocular imagery.
260 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2016
300 $a 113 p.
500 $a Source: Dissertation Abstracts International, Volume: 78-04(E), Section: B.
500 $a Adviser: Dimitris Metaxas.
502 $a Thesis (Ph.D.)--Rutgers The State University of New Jersey - New Brunswick, 2016.
520 $a Estimating 3D human pose from monocular images is an important and challenging problem in computer vision with numerous applications including human-computer interaction, human activity recognition, biomechanical analysis, and security. Existing state-of-the-art methods utilize statistical learning models that are inherently limited because they require sufficient training data that does not often include uncommon pose articulations or subject proportions. In addition, these methods often return global average case results and cannot easily leverage anthropomorphic, kinematic, and other physics-based constraints. However, a prior model-based search can be computationally prohibitive due to the combinatorially large set of plausible joint combinations. We combine statistical learning-based approaches with a prior part-based model into a hybrid constrained optimization that leverages strengths of both approaches. The method guarantees a plausible human pose while also resolving local ambiguities among body parts. Qualitative evaluation of the proposed methods on human pose datasets show improvement in reconstruction accuracy compared to current state-of-the-art methods.
590 $a School code: 0190.
650 4 $a Computer science. $3 199325
650 4 $a Statistics. $3 182057
650 4 $a Electrical engineering. $3 454503
690 $a 0984
690 $a 0463
690 $a 0544
710 2 $a Rutgers The State University of New Jersey - New Brunswick. $b Graduate School - New Brunswick. $3 492824
773 0 $t Dissertation Abstracts International $g 78-04B(E).
790 $a 0190
791 $a Ph.D.
792 $a 2016
793 $a English
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=10291949