KSC 2018
Current Result Document : 2 / 2
ÇѱÛÁ¦¸ñ(Korean Title) |
A similar image search framework using Spark and HBase in cloud environment |
¿µ¹®Á¦¸ñ(English Title) |
A similar image search framework using Spark and HBase in cloud environment |
ÀúÀÚ(Author) |
Tri D.T. Nguyen
Tra My Pham Thi
Dong Kwan You
Dongyeong Son
Eui-Nam Huh
|
¿ø¹®¼ö·Ïó(Citation) |
VOL 45 NO. 02 PP. 0468 ~ 0470 (2018. 12) |
Çѱ۳»¿ë (Korean Abstract) |
|
¿µ¹®³»¿ë (English Abstract) |
The exponentially increasing social networks such as Facebook, Instagram and Flick have become one of the most challenging tasks of storing and retrieving desired images in content based image retrieval (CBIR) research nowadays. To enable the decrease in the speed of searching time on such large image datasets, we propose a parallel search framework using Spark and HBase. While HBase is an efficient method to store a large dataset in the distributed manner, Spark is a well-known framework to enhance the performance of retrieving similar images in CBIR system by using distributed computing paradigm. Additionally, we used a speeded-up robust features (SURF) method to extract local features of an image, then match relevant images based on Euclidean distance between two features. Simulation results show that our proposed framework can achieve good performance when working in a large dataset.
|
Å°¿öµå(Keyword) |
|
ÆÄÀÏ÷ºÎ |
PDF ´Ù¿î·Îµå
|