- International Journal of Applied Mathematics Electronics and Computers
- Special Issue: (2016) Special Issue
- Parallelization of a Hierarchical Graph-Based Image Segmentation using OpenMP
Parallelization of a Hierarchical Graph-Based Image Segmentation using OpenMP
Authors : Ali Sağlam, Nurdan Akhan Baykan
Pages : 6-11
Doi:10.18100/ijamec.271038
View : 11 | Download : 7
Publication Date : 2016-12-01
Article Type : Research
Abstract :In many image-processing applications, image segmentation is an essential stage. In this stage, an image is partitioned into several regions according to the similarity of its pixels. In addition to the accuracy of the image segmentation, the speed is also very important for real-time image processing applications. Many computer applications take advantages of the multi-processor architecture to up to their running performance. However, to run an algorithm as parallel is very difficult in many cases. Due to using the same memory blocks, many conflicts might be happened between the processors. Moreover, each process of one processor may depend on those of another processor. For this reason, the algorithm to be parallelized must be suitable to parallel. In addition, the processing traffic that is pursued by the processors must be controlled within some parallel directives. In this paper, we provide a parallel implementation to a hierarchical graph-based image segmentation method by using its hierarchical processing steps. To achieve this goal, we utilize the OpenMP (Open Multi-Processing) Library to run the segmentation process as parallel on images of different sizes from the INRIA Holidays dataset. The experimental results show that the parallel implementation of the algorithm is more effective than the serial type according to processing time.Keywords : Parallel programming, Image segmentation, Graph, OpenMP