1
Similarity searching in medical image databases using R*-tree | |
Author | Utomo, Darmawan |
Call Number | AIT Thesis no. CS-99-6 |
Subject(s) | Image processing Imaging systems in medicine |
Note | A thesis submitted in partial fulfillment of the requirements for the degree of Master of Engineering, School of Engineering and Technology |
Publisher | Asian Institute of Technology |
Abstract | The R-tree data structure (Guttman, 1984) successfully handles approximate searching by image content in medical image databases (Petrakis, 1997). The image contents are mapped to points in a multidimensional space, and are indexed using the R-tree method. The method can retrieve all images up to a prespecified degree of similarity (tolerance), has no "false dismissals" (all images qualifying against the query selection criteria are retrieved), and is much faster than sequential scanning for searching in main memory. However, the R-tree structure pursues optimization based on minimizing area, which have been shown to be sub-optimal. The R *-tree data structure (Beckman, 1990), incorporates a combined optimization of area, margin and overlap, and so outperforms the R-tree data structure. This work describes several experiments designed to compare the response times, space utilization, split times and insertion times between R-tree and R *-tree similarity searching in medical image databases. The R*-tree outperforms than R-tree in terms of response time for about 41.7 % of the searches with one unlabeled object, and for about 36.6 % of searches with two unlabeled object, in presumably 20-50 retrieved images. The R*-tree smart deferred splitting displays space utilization 65 % higher than the R-tree, with ten or twelve minimum nodes filled. |
Year | 1999 |
Type | Thesis |
School | School of Engineering and Technology (SET) |
Department | Department of Information and Communications Technologies (DICT) |
Academic Program/FoS | Computer Science (CS) |
Chairperson(s) | Sadananda, Ramakoti |
Examination Committee(s) | Davison, Andrew;Yulu, Qi |
Scholarship Donor(s) | Oikumenisches Studientwerk e. V. German |
Degree | Thesis (M. Eng.) - Asian Institute of Technology, 1999 |