Tracking and data association
IEEE Transactions on Pattern Analysis and Machine Intelligence
Computer Vision: A Modern Approach
Computer Vision: A Modern Approach
Data association methods with applications to law enforcement
Decision Support Systems
Clean Answers over Dirty Databases: A Probabilistic Approach
ICDE '06 Proceedings of the 22nd International Conference on Data Engineering
Data association for topic intensity tracking
ICML '06 Proceedings of the 23rd international conference on Machine learning
Principles of Constraint Programming
Principles of Constraint Programming
Efficient query evaluation on probabilistic databases
The VLDB Journal — The International Journal on Very Large Data Bases
Databases with uncertainty and lineage
The VLDB Journal — The International Journal on Very Large Data Bases
Ranking queries on uncertain data: a probabilistic threshold approach
Proceedings of the 2008 ACM SIGMOD international conference on Management of data
Online Filtering, Smoothing and Probabilistic Modeling of Streaming data
ICDE '08 Proceedings of the 2008 IEEE 24th International Conference on Data Engineering
Top-k queries on uncertain data: on score distribution and typical answers
Proceedings of the 2009 ACM SIGMOD International Conference on Management of data
Probabilistic Databases
Performance evaluation of object detection and tracking in video
ACCV'06 Proceedings of the 7th Asian conference on Computer Vision - Volume Part II
| Hi-index | 0.00 |
The Probabilistic Data Association (PDA) problem involves identifying correspondences between items over data sequences on the basis of similarity functions. PDA has long been a topic of interest in many application areas such as real-time tracking and surveillance. Despite its significance, however, it has largely been ignored by the event-processing community. Our work rectifies this situation by studying PDA in the context of continuous event processing. Specifically, we formulate PDA as a continuous probabilistic ranking problem with constraints and efficiently solve it using fast constraint resolution. Our solutions are built on a top-k approximation to the problem guided by resource-aware optimization techniques that adaptively utilize the available resources to produce real-time results. User-defined data association constraints are used to restrict the solution space and quickly eliminate inconsistent solution candidates. We also derive the runtime complexity of our solutions and experimentally evaluate these solutions using a prime PDA application: real-time tracking of moving objects within a camera network. Our evaluation results demonstrate the superiority of our solutions over traditional constraint programming formulations.