A Study of Scalable Images Search Using Binarized SIFT Features on Hbase
| Autor: | Ren-Yi Wang, 王人儀 |
|---|---|
| Rok vydání: | 2015 |
| Druh dokumentu: | 學位論文 ; thesis |
| Popis: | 103 The Scale Invariance Feature Transform (SIFT) image descriptors have been widely used in the image recognition applications. The feature keypoints extracted by SIFT are invariant to scaling, rotation, affine distortion, and illumination changes. So the SIFT descriptors are very reliable for near duplicate image search. However, the traditional SIFT descriptors comparison is based on Euclidean distance of their feature vectors and thus is not scalable when the number of images increases. In this paper, we present a novel way to store the SIFT image descriptors in Hbase that facilitates constant (O(1)) descriptor lookup time irrelevant to the number of images stored in the databases. Our experiments in the scale of 10M images support that constant time SIFT descriptor lookup is achievable by our proposed method. |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|