Efficient Local Histogram Searching via Bitmap Indexing
Loading...
Date
2015
Journal Title
Journal ISSN
Volume Title
Publisher
The Eurographics Association and John Wiley & Sons Ltd.
Abstract
Representing features by local histograms is a proven technique in several volume analysis and visualization applications including feature tracking and transfer function design. The efficiency of these applications, however, is hampered by the high computational complexity of local histogram computation and matching. In this paper, we propose a novel algorithm to accelerate local histogram search by leveraging bitmap indexing. Our method avoids exhaustive searching of all voxels in the spatial domain by examining only the voxels whose values fall within the value range of user-defined local features and their neighborhood. Based on the idea that the value range of local features is in general much smaller than the dynamic range of the entire dataset, we propose a local voting scheme to construct the local histograms so that only a small number of voxels need to be examined. Experimental results show that our method can reduce much computational workload compared to the conventional approaches. To demonstrate the utility of our method, an interactive interface was developed to assist users in defining target features as local histograms and identify the locations of these features in the dataset.
Description
@article{10.1111:cgf.12620,
journal = {Computer Graphics Forum},
title = {{Efficient Local Histogram Searching via Bitmap Indexing}},
author = {Wei, Tzu-Hsuan and Chen, Chun-Ming and Biswas, Ayan},
year = {2015},
publisher = {The Eurographics Association and John Wiley & Sons Ltd.},
DOI = {10.1111/cgf.12620}
}