Wednesday, September 30, 2020 Home   | Search
Viewing Abstract

Page: 38-44 ------------------------------> Last_modified :7/11/2017 12:11:00 PM

Title: Distributed visual enhancement on surveillance video with Hadoop Mapreduce and performance evaluation in pseudo distributed mode
Authors: Mohd Fikri Azli Abdullah; Md. Shohel Sayeed; Anang Hudaya Muhammad Amin; Nazrul Muhaimin Ahmad; Ibrahim Yusof; Liew Tze Hui; Housam Khalifa Bashier
DOI:
Aff: Faculty of Information Science and Technology, Multimedia University, Jalan Ayer Keroh Lama, 75450 Melaka, Malaysia
Author Email:
Keywords: Distributed; Visual Enhancement; Histogram Equalization; Hadoop MapReduce
URLs: ABSTRACT-HTML  | FULLTEXT-PDF  | 
Abstract:Background: A large amount of surveillance video is generated due to the improvement in camera performance and the low-priced light-weighted cameras availability. Also, there are limits to the single computer to process large-scale data, such as video analysis. Thus, the advantages of parallel distributed processing of a video database by using the computational resources of a cloud computing environment should be considered. Objective: This work implements the distributed visual enhancement using histogram equalization algorithm on image database from surveillance cameras. The experiment is conducted in pseudo distributed mode under Hadoop MapReduce architecture. The database images were processed with different number of map tasks and the performance is analysed. Results: The result shows that the increases of map tasks will increase the total time because the processing time of histogram equalization is very small. Main part of the total time is contributed by the transferring time of images among map tasks. Conclusion: We argue that the distributed infrastructure implemented with Hadoop MapReduce is suitable for more complex problems which require higher processing capability.d