基于并行计算优化的WindSTORM PLUS算法  

WindSTORM PLUS Algorithm with Parallel Computing Optimization

作　　者：肖文 吴天琦 李仁剑 唐黎 陈玲玲[1,2] Xiao Wen;Wu Tianqi;Li Renjian;Tang Li;Chen Lingling(College of Physics and Optoelectronic Engineering,Key Laboratory of Optoelectronic Devices and Systems of Ministry of Education and Guangdong Province,Shenzhen University,Shenzhen,Guangdong 518060,China;College of Health Science and Environmental Engineering,Shenzhen Technology University,Shenzhen,Guangdong 518118,China)

机构地区：[1]深圳大学物理与光电工程学院,光电子器件与系统广东省/教育部重点实验室,广东深圳518060 [2]深圳技术大学健康与环境工程学院,广东深圳518118

出　　处：《中国激光》2020年第6期277-285,共9页Chinese Journal of Lasers

基　　金：广东省普通高校特色创新类项目(2018KTSCX349);深圳市科创委学科布局项目(JCYJ 20150525092941015)。

摘　　要：随机光学重建显微镜(STORM)关键技术包括大量随机闪烁图像的数据定位与重建算法,而现有的常用开源算法在大数据量情况下存在用时过长或内存受限等限制,影响了STORM技术的进一步推广应用。基于MATLAB和并行计算的方法构建了WindSTORM PLUS开源算法,采用该算法进行单分子定位数据处理。在模拟数据集下,对比WindSTORM和ThunderSTORM,WindSTORM PLUS算法的处理速度提高了1000%,且对比WindSTORM,内存需求降低了60%。此外,搭建了easySTORM系统,在实验数据的处理耗时对比中,WindSTORM PLUS只有WindSTORM和Gauss-WLS的9%,验证了其在超大数据集下处理速度的优越性。WindSTORM PLUS开源算法为超分辨图像处理提供了一个新的高速处理方案。The major stochastic optical reconstruction microscope(STORM)techniques include data localization and reconstruction algorithms with respect to a large number of images.However,the existing open source algorithms are limited by their slow speed or computer memory when processing an ultralarge dataset,restricting the extensive application of STORM techniques.Therefore,we propose WindSTORM PLUS for performing single-molecule localization data processing using MATLAB and parallel computation.The outcomes obtained using the simulated large datasets demonstrate that the data processing speed of WindSTORM PLUS is greater than those of the existing WindSTORM and ThunderSTORM by 1000%.Furthermore,the memory requirements are reduced by 60%compared with those in case of WindSTORM.In addition,we establish an easySTORM system and conduct a test using some samples to verify the superiority of our algorithm.The time-consuming of WindSTORM PLUS is only 9%of WindSTORM and Gauss-WLS.We believe that this open source algorithm can provide a novel high-speed STORM data processing approach.

关 键 词：显微 随机光学重建显微镜 并行计算 超分辨成像 图形处理器 

分 类 号：O439[机械工程—光学工程]