TY - GEN
T1 - DM-SIRT
T2 - 15th International Symposium on Bioinformatics Research and Applications, ISBRA 2019
AU - Wang, Zihao
AU - Zhang, Jingrong
AU - Liu, Xintong
AU - Liu, Zhiyong
AU - Wan, Xiaohua
AU - Zhang, Fa
N1 - Publisher Copyright:
© 2019, Springer Nature Switzerland AG.
PY - 2019
Y1 - 2019
N2 - The ‘missing wedge’ of single tilt in electron tomography introduces severely artifacts into the reconstructed results. To reduce the ‘missing wedge’ effect, a widely used method is ‘multi-tilt reconstruction’, which collects projections using multiple different axes. However, as the number of tilt series increases, its computing and memory costs also rises. While the demand to speed up its reconstruction procedure grows, the huge memory requirement from the 3D structure and strong data dependencies from projections heavily limit its parallelization. In our work, we present a new fully distributed multi-tilt reconstruction framework named DM-SIRT. To improve the parallelism of the reconstruction process and reduce the memory requirements of each process, we formulate the multi-tilt reconstruction as a consensus optimization problem and design a distributed multi-tilt SIRT algorithm. To improve the reconstruction resolution, we applied a multi-agent consensus equilibrium (MACE) with a new data division strategy. Experiments show that along with the visually and quantitatively improvement in resolution, DM-SIRT can acquire a 5.4x speedup ratio compared to the raw multi-tilt reconstruction version. It also has 87% decrease of memory overhead and 8 times more scalable than the raw reconstruction version.
AB - The ‘missing wedge’ of single tilt in electron tomography introduces severely artifacts into the reconstructed results. To reduce the ‘missing wedge’ effect, a widely used method is ‘multi-tilt reconstruction’, which collects projections using multiple different axes. However, as the number of tilt series increases, its computing and memory costs also rises. While the demand to speed up its reconstruction procedure grows, the huge memory requirement from the 3D structure and strong data dependencies from projections heavily limit its parallelization. In our work, we present a new fully distributed multi-tilt reconstruction framework named DM-SIRT. To improve the parallelism of the reconstruction process and reduce the memory requirements of each process, we formulate the multi-tilt reconstruction as a consensus optimization problem and design a distributed multi-tilt SIRT algorithm. To improve the reconstruction resolution, we applied a multi-agent consensus equilibrium (MACE) with a new data division strategy. Experiments show that along with the visually and quantitatively improvement in resolution, DM-SIRT can acquire a 5.4x speedup ratio compared to the raw multi-tilt reconstruction version. It also has 87% decrease of memory overhead and 8 times more scalable than the raw reconstruction version.
KW - Consensus optimization
KW - Cryo-electron Tomography
KW - Multi-tilt reconstruction
KW - Parallel computing
KW - TxBR
UR - http://www.scopus.com/inward/record.url?scp=85066823026&partnerID=8YFLogxK
U2 - 10.1007/978-3-030-20242-2_19
DO - 10.1007/978-3-030-20242-2_19
M3 - Conference contribution
AN - SCOPUS:85066823026
SN - 9783030202415
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 220
EP - 231
BT - Bioinformatics Research and Applications - 15th International Symposium, ISBRA 2019, Proceedings
A2 - Li, Min
A2 - Cai, Zhipeng
A2 - Skums, Pavel
PB - Springer Verlag
Y2 - 3 June 2019 through 6 June 2019
ER -