ABSTRACT
This study aims to explore optimization methods for superpixel segmentation algorithms in large-scale image processing. Due to the significant time consumption of existing superpixel segmentation algorithms when dealing with large-scale image data, this study focuses on parallel optimization of the SLIC superpixel segmentation algorithm on the new-generation Sunway manycore processor SW26010-Pro. The research analyzes the characteristics of the SW26010-Pro heterogeneous manycore processor, existing superpixel image segmentation algorithms, and the SLIC superpixel segmentation algorithm. The study primarily analyzes hot-spot functions and improves the algorithm's running speed by 6-7 times through optimizing processes such as RGB-to-CIELab color space conversion, cluster updating, and seed point updating. It effectively presents a parallel execution scheme for the SLIC superpixel algorithm on the SW26010-Pro.
- Radhakrishna Achanta, Appu Shaji, Kevin Smith, Aurelien Lucchi, Pascal Fua, and Sabine Süsstrunk. 2012. SLIC Superpixels Compared to State-of-the-Art Superpixel Methods. IEEE Transactions on Pattern Analysis and Machine Intelligence, 34(11), 2274–2282. DOI:https://doi.org/10.1109/TPAMI.2012.120.Google ScholarDigital Library
- Quan Zhang, Baiyue Lin, Bo Yang, Bo Peng, Wei Zhang and Ran Tu. 2020. Parabolic Radon transform parallel algorithm for CPU-GPU heterogeneous platform. Oil Geophysical Prospecting 55, 06, 1263–1270+1163. DOI:https://doi.org/10.13810/j.cnki.issn.1000-7210.2020.06.012.Google ScholarCross Ref
- Yi Xu. 2022. Research on Parallel Optimization of BLAS Based on the new generation of Sunway many-core processor. Master's thesis. East China Normal University. DOI:https://doi.org/10.27149/d.cnki.ghdsu.2022.003511.Google ScholarCross Ref
- Yuedan Chen. 2020. Design and Implementation of Heterogeneous Parallel Algorithms on the Sunway TaihuLight. Ph.D. Dissertation. Hunan University. DOI:https://doi.org/10.27135/d.cnki.ghudu.2020.002006.Google ScholarCross Ref
- Guo-Liang Chen. 2002. The Design and Analysis of Parallel Algorithms. Higher Education Press, Beijing, China.Google Scholar
- Chunyao Wang, Junzhou Chen, and Wei Li. 2014. Review on superpixel segmentation algorithms. Computer Applications Research, 31(01), 6-12.Google Scholar
- Carl Yuheng Ren, Victor Adrian Prisacariu, and Ian D Reid. 2015. gSLICr: SLIC superpixels at over 250Hz. arXiv:1509.04232. DOI:https://doi.org/10.48550/arXiv.1509.04232Google ScholarCross Ref
- ITU-R. Parameter values for the hdtv standards for production and international programme exchange. Technical report, ITU-R Rec, BT.709-6, 2015.Google Scholar
- Xiaoqi Luo, Yuanjie Xing, and Senhai Xu. 2022. Parallel Optimization of Super Pixel Algorithm SLIC. In 2022 3rd International Conference on Information Science, Parallel and Distributed Systems (ISPDS), 0–0. DOI:https://doi.org/10.1109/ISPDS56360.2022.9874224Google ScholarCross Ref
- Chen, X., Zhang, R. 2016. Parallel Implementation of SLIC Superpixels Segmentation. Modern Computer (Professional Edition), 2016(29), 75-77.Google Scholar
- Zhu, S. X., Gu, T. X., & Liu, X. P. (2014). Minimizing synchronizations in sparse iterative solvers for distributed supercomputers. Computers & Mathematics with Applications, 67(1), 199-209.Google ScholarDigital Library
- T. Gu, X. Liu, "On the Memory Wall and Performance of Symmetric Sparse Matrix Vector Multiplications In Different Data Structures on Shared Memory Machines," 2015 IEEE 12th Intl Conf on Ubiquitous Intelligence and Computing and 2015 IEEE 12th Intl Conf on Autonomic and Trusted Computing and 2015 IEEE 15th Intl Conf on Scalable Computing and Communications and Its Associated Workshops (UIC-ATC-ScalCom), Beijing, China, 2015, pp. 1439-1444, doi: 10.1109/UIC-ATC-ScalCom-CBDCom-IoP.2015.259.Google ScholarCross Ref
Index Terms
- Parallel Optimization of SLIC Algorithm for New-Generation Sunway Processors
Recommendations
Texture-based superpixel segmentation algorithm for classification of hyperspectral images
To increase classification accuracy, a variety of feature extraction techniques have been presented. A pre-processing method called superpixel segmentation divides an image into meaningful sub-regions, which simplifies the image. This substantially ...
A novel method of cervical cell image segmentation via region merging and SLIC
SoICT '16: Proceedings of the 7th Symposium on Information and Communication TechnologyConsidering that the existing methods of cell segmentation are sensitive to noise, in this paper, a novel method was proposed to obtain the contour of cervical cell accurately and robustly. Firstly, the mean shift algorithm was utilized to smooth the ...
Improved SLIC imagine segmentation algorithm based on K-means
Dividing the image into superpixels contributes to further processing of the image. Simple linear iterative clustering (SLIC) algorithm achieves good segmentation result by clustering color and distance characteristics of pixels. However, finite ...
Comments