생몰정보
소속
직위
직업
활동분야
주기
서지
국회도서관 서비스 이용에 대한 안내를 해드립니다.
검색결과 (전체 1건)
원문 있는 자료 (1) 열기
원문 아이콘이 없는 경우 국회도서관 방문 시 책자로 이용 가능
목차보기더보기
Title Page
Abstract
Contents
Chapter 1. Introduction 10
1.1. Statement of Problem 10
1.2. Related Works 14
1.2.1. Motion Pattern Analysis Using Trajectory 14
1.2.2. Motion Pattern Analysis Using Local Motions 17
1.3. Contributions 19
1.4. Thesis Organization 19
Chapter 2. Preliminaries 21
2.1. Latent Dirichlet Allocation (LDA) 21
2.1.1. Probabilistic Graphical Model 21
2.1.2. LDA Property & Formulation 27
2.2. Inference of LDA 35
2.2.1. Collapsed Gibbs Sampling 36
2.2.2. Variational Inference 48
Chapter 3. Proposed Approach 61
3.1. Probabilistic Inference Model 62
3.2. Model Learning 70
3.2.1. Online Trajectory Clustering 72
3.2.2. Spatio-Temporal Dependency of Activities 76
3.2.3. Velocity Learning 77
3.3. Anomaly Detection 78
3.4. Summary of the Proposed Method 81
Chapter 4. Experiments 83
4.1. Result of Traffic Pattern Understanding 84
4.2. Applications in Anomaly Detection 95
4.3. Prediction Task 107
4.4. Comparison with Sampling 109
Chapter 5. Conculsion 111
5.1. Concluding Remarks 111
5.2. Future Works 112
Bibliography 113
초록 121
Figure 1.1 Examples of various traffic scenes. 11
Figure 1.2. An example of motion pattern analysis. 13
Figure 1.3. An example of perspective projection distortion. 16
Figure 2.1. (a) An example of a graphical models. 22
Figure 2.2. Probability density function of beta distribution. 26
Figure 2.3. An example of probabilistic generative process for LDA. 28
Figure 2.4. Examples of θd drawn by Dirichlet distributions for various settings of the...(이미지참조) 31
Figure 2.5. Graphical representation for latent Dirichlet allocation and summary of no-... 34
Figure 2.6. Relation among the log marginal probability logp(χ), KL-divergence... 50
Figure 2.7. Graphical representation of the original LDA model and approximated... 55
Figure 3.1. Overall scheme of the proposed method. 62
Figure 3.2. Example of a single trajectory corresponding with a set of cells. 63
Figure 3.3. Synthetic trajectory with marked points and relative vectors from origin... 64
Figure 3.4. Graphical representation of the state transition model. 65
Figure 3.5. Graphical representation of the trajectory pattern (topic) generative model. 66
Figure 3.6. Graphical representation of the proposed model. 68
Figure 3.7. Three sub-models for two-stage learning. 73
Figure 4.1. Typical patterns and their spatio-temporal relationship for the WI video... 85
Figure 4.2. Omitted typical patterns to facilitate display of trajectory patterns. 86
Figure 4.3. The process of online inference-(1). 87
Figure 4.4. The process of online inference-(2). 88
Figure 4.5. Typical patterns and their spatio-temporal relationship for the QMUL video... 89
Figure 4.6. Typical patterns and their spatio-temporal relationship for the MIT video... 90
Figure 4.7. The example of merging two typical patterns. Adjacent two patterns (each... 91
Figure 4.8. The example of splitting two typical patterns. One typical pattern is split... 92
Figure 4.9. Trajectory patterns when K=6 93
Figure 4.10. The result of parameters {mn|n=1,..,S} according to variation of S.(이미지참조) 94
Figure 4.11. Error rate of state estimation in the WI dataset and comparison with the... 95
Figure 4.12. Examples of anomaly detections related to the first require ment (semantic... 96
Figure 4.13. Examples of anomaly detections related to the second re quirement (Speed... 97
Figure 4.14. Comparison of motion likelihoods between the proposed model(actual ve-... 99
Figure 4.15. Examples of anomaly detections related to the third requirement (spatial ... 101
Figure 4.16. Scenario of a traffic animation to simulate a trouble of a traffic control... 102
Figure 4.17. State transition probability owing to the trouble of traffic signal. 103
Figure 4.18. Tracking failure case of the object based multi-target tracking method in a... 103
Figure 4.19. Examples of anomaly detections related to the fourth requirement (robust... 104
Figure 4.20. Video animation of a reversible lane 105
Figure 4.21. Examples of anomaly detections related to the fifth requirement (online... 106
Figure 4.22. Process of trajectory pattern adaptation. 106
Figure 4.23. Comparison of average accuracy on a prediction. X-ais indicates number... 107
Figure 4.24. Qualitative comparison of proposed method and sampling based learning. 109
Figure 4.25. Quantitative comparison with online Gibbs Sampling (Canini et al., 2009)... 110
원문구축 및 2018년 이후 자료는 524호에서 직접 열람하십시요.
도서위치안내: / 서가번호:
우편복사 목록담기를 완료하였습니다.
* 표시는 필수사항 입니다.
* 주의: 국회도서관 이용자 모두에게 공유서재로 서비스 됩니다.
저장 되었습니다.
로그인을 하시려면 아이디와 비밀번호를 입력해주세요. 모바일 간편 열람증으로 입실한 경우 회원가입을 해야합니다.
공용 PC이므로 한번 더 로그인 해 주시기 바랍니다.
아이디 또는 비밀번호를 확인해주세요