Call number:
S 99.0139(356)
In:
Wissenschaftliche Arbeiten der Fachrichtung Vermessungswesen der Universität Hannover
Type of Medium:
Series available for loan
Pages:
x, 111 Seiten
,
Illustrationen, Diagramme
ISSN:
0174-1454
Series Statement:
Wissenschaftliche Arbeiten der Fachrichtung Vermessungswesen der Universität Hannover Nr. 356
Language:
English
Note:
Dissertation, Gottfried Wilhelm Leibniz Universität Hannover, 2020
,
Abstract
Zusammenfassung
Acknowledgments
Definition, Acronyms and Symbols
1 Introduction
1.1 Motivation
1.2 Person Re-Identification
1.3 Problem statement and research objective
1.4 Contribution
1.5 Outline of this thesis
2 Related work
2.1 Scope
2.2 Historical overview
2.3 Terminology and strategies
2.4 Handcrafted feature extraction methods
2.5 Data-driven feature extraction methods
2.6 Person view specific methods
2.7 Re-Ranking based methods
2.8 Domain adaptation methods
2.9 Discussion
3 Fundamentals
3.1 Fisheye camera geometry and projection model
3.2 Feature extraction
3.2.1 GOG/XQDA -
a handcrafted feature extraction method
3.2.2 TriNet and SRNN -
two data-driven feature extraction methods ....
4 A new approach for person re-identification
4.1 General overview
4.2 Input and assumptions
4.3 Projection alignment
4.4 View classification and sampling
4.5 Per-view matching
4.6 Fusion
4.7 Discussion of the approach
5 Experimental evaluation
5.1 General structure of this chapter
5.2 Multi-view investigations
5.2.1 Datasets
5.2.2 Training and inference procedure
5.2.3 Evaluation and discussion
5.3 Bird's eye view investigations
5.3.1 Datasets
5.3.2 Training and inference procedure
5.3.3 Evaluation and discussion
5.4 Influence of data
5.4.1 Datasets
5.4.2 Training and inference procedure
5.4.3 Evaluation and discussion
5.5 Fisheye investigations
5.5.1 Datasets
5.5.2 Training procedure
5.5.3 Projection alignment
5.5.4 Person view classification
5.5.5 Assessment of PRID results
5.5.6 Comparison with a contemporary approach
5.5.7 Qualitative comparison
6 Conclusions and future work
A Datasets
A.l Our novel datasets
A.2 Public datasets
References
,
Sprache der Zusammenfassungen: Englisch, Deutsch
Location:
Lower compact magazine
Branch Library:
GFZ Library
Permalink
