1 / 18

The DR-tree: A Main Memory Data Structure for Complex Multi-dimensional Objects

The DR-tree: A Main Memory Data Structure for Complex Multi-dimensional Objects. YOUNG-JU LEE , CHIN-WAN CHUNG. Seung-Hyun Ji Graphics Application Lab. Contents. Introduce Index Structure. Problem of Index Structure. Related Work(TR*-Tree). Introduce DMBR and DR-Tree.

merv
Télécharger la présentation

The DR-tree: A Main Memory Data Structure for Complex Multi-dimensional Objects

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. The DR-tree: A Main Memory Data Structure for Complex Multi-dimensional Objects YOUNG-JU LEE , CHIN-WAN CHUNG Seung-Hyun Ji Graphics Application Lab

  2. Contents • Introduce Index Structure. • Problem of Index Structure. • Related Work(TR*-Tree). • Introduce DMBR and DR-Tree. • Compare to state-of-the-art index structure(GENESYS).

  3. Main Memory Data Structure Original Data Secondary Storage Main Memory Data Structure

  4. Index Structure • Index structure for complex object. • MBR • Smallest aligned n-dimensional rectangle enclosing and object. • LSD-Tree, R*-Tree, X-Tree • Region decomposition • Divided into sub-region until a region obtains a desired simple component. • PM quadtree, TR*-Tree

  5. Index structure Problem • MBR • `False hit’ • False hit candidate. • Refinement step • refinement step is very costly. • Region decomposition • Simple component • Quadrants, trapezoid, line segment. • Number of decomposed components could result in a storage and query processing overhead.

  6. Related Work(1/2) • TR*-Tree • Improve R*-Tree • Represent exact geometry spatial attributes • Reduce memory operations • Store components of 1 decomposed object • Internal node • Pointer child node • Minimum bounding rectangle of trapezoids in child • Leaf node • Trapezoids

  7. Related Work(2/3) R1 1 • TR* Tree A 2 3 B 4 5 6 C 7 9 8 D 10 11 E R2 F 15 12 13 14 R1 R2 A E F B C D 1 3 8 11 2 9 12 7 10 13 14 15 4 5 6

  8. Related Work(3/3) • TR* Tree

  9. DR-Tree(1/3) • DMBR • Decomposition Method For multi-dimension complex object. • Extend to MBR. • Additional Constraint. • Accuracy of the Decomposition(AOD). • split permit above a threshold.

  10. DR-Tree(2/3) • Example of DMBR • AOD(2) : 1/4 • 2D Object • 3D Object

  11. DR-Tree(3/3) • Construction DR-Tree a c b e d

  12. Two-Step Index Structure • Original Object • R*-Tree • Decomposition • DR-Tree

  13. Query Processing • Query Processing • Point Query • Filter Step : R* Tree search algorithm. • Refinement Step : use DR Tree . • Region Query • Filter Step : R* Tree search algorithm. • Traditional decomposition methods not support efficient performance.(number of component) • Small number of components.(DMBR) • Spatial Join Query

  14. State of the art • Genesys index structure • Original Data • Use R*-Tree • Decomposition Method • Use TR* Tree

  15. Performance Analysis(1/3) • Performance • Using real geometric data(park,map,lake,state). • Compare to Genesys(TR* Tree). Query processing time for various spatial queries. IO-time and CPU time

  16. Performance Analysis(2/3) • Performance Storage requirements (saving 71%) Preprocessing cost

  17. Performance Analysis(3/3) • Performance Query processing time and storage requirement for TIGER/Line files.

  18. Conclusion • Proposed a main memory data structure for complex multi-dimensional object. • Extension of an existing index structure • Reduce processing time. • Reduce the amount of storage. • Easier to implement and applicable to various spatial data.

More Related