Graph-cut has been proved to return good quality on the optimization of depth estimation. Leveraging the parallel computation has been proposed as a solution to handle the intensive computation of graph-cut algorithm. This pa-per proposes two parallelization techniques to enhance the execution time of graph-cut optimization. By executing on an Intel 8-core CPU, the proposed scheme can achieve an average of 4.7 times speedup with only 0.01% energy increase.
|主出版物標題||Advances in Intelligent Systems and Applications - Volume 2|
|主出版物子標題||Proceedings of the International Computer|
|編輯||Chang Ruay-Shiung, Peng Sheng-Lung, Lin Chia-Chen|
|出版狀態||Published - 28 6月 2013|
|名字||Smart Innovation, Systems and Technologies|