1 / 16

REU 2013 Report 3

Alla Petrakova. REU 2013 Report 3. Last week recap. Trajectory Clustering TRACLUS UCF Motion Pattern Algorithm. Quality of Clusters. Attempt to find a Generally Accepted Quantiative Measure. Approaches to Evaluating Quality of Clusters. qualitative. quantitative.

brandy
Télécharger la présentation

REU 2013 Report 3

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Alla Petrakova REU 2013 Report 3

  2. Last week recap • Trajectory Clustering • TRACLUS • UCF Motion Pattern Algorithm

  3. Quality of Clusters Attempt to find a Generally Accepted Quantiative Measure

  4. Approaches to Evaluating Quality of Clusters qualitative quantitative Correct Clustering Rate Sum of Squared Error Accuracy Measure ± Error or Noise Penalty • Ground truth • Visual inspection • Synthetic datasets • Comparison to another algorithm

  5. SSE J. gil Lee and J. Han. Trajectory clustering: A partition-and-group framework. In Proceedings of the ACM International Conference on Management of Data (SIGMOD), Beijing, China, pages 593–604, 2007. Cited by 357

  6. Sum of Squared Error N denotes the set of all noise line segments.

  7. Correct Clustering Rate B. Morris and M. Trivedi, “Learning Trajectory Patterns by Clustering: Experimental Studies and Comparative Evaluation,” Proc. IEEE Conf. Computer Vision and Pattern Recognition, pp. 312- 319, June 2009.

  8. Correct Clustering Rate • Find one-to-one mapping between the ground truth and clustering labels which maximized the number of matches. • where N is the total number of trajectories and pc denotes the total number of trajectories matched to the c-th cluster.

  9. Accuracy Measure • IN – total number of clusters • bi = the number of labeled trajectories that are most frequent in a given cluster • Bi = the total number of trajectories in a cluster

  10. Testing

  11. Vehicle Motion Patterns • Dataset:

  12. TRACLUS results

  13. UCF Motion Pattern results

  14. Australian Sign Language Dataset Used in Following Papers: • M. Vlachos, G. Kollios, and D. Gunopulos, “Discovering Similar Multidimensional Trajectories,” Proc. Int’l Conf. Data Eng., pp. 673- 684, 2002. (cited by 631) • Lei Chen, M. Tamer Özsu, and Vincent Oria. 2005. Robust and fast similarity search for moving object trajectories. In Proc. of the 2005 ACM SIGMOD int’l conf. on Management of data (SIGMOD '05). ACM, New York, NY, USA, 491-502. DOI=10.1145/1066157.1066213 (Cited by 395) • A. Naftel and S. Khalid, “Motion Trajectory Learning in the DFT- Coefficient Feature Space,” Proc. IEEE Int’l Conf. Computer Vision Systems, pp. 47-47, Jan. 2006. (cited by 26) • W. Hu, X. Li, G. Tian, S. Maybank, and Z. Zhang, ” An Incremental DPMM-Based Method for Trajectory Clustering, Modeling, and Retrieval”, IEEE Transactions on Pattern Analysis and Machine Intelligence, VOL. 35, NO. 5, MAY 2013 • Tsumoto, S., Hirano, S.: Detection of risk factors using trajectory mining. J. Intell. Inf. Syst. 36(3), 403–425 (2011) (cited by 15)

  15. Australian Sign Language Dataset

  16. ASL Testing • No meaningful results • Separating out individual trajectories

More Related