PMID- 26579597
OWN - NLM
STAT- MEDLINE
DCOM- 20161101
LR  - 20181202
IS  - 1361-6560 (Electronic)
IS  - 0031-9155 (Linking)
VI  - 60
IP  - 24
DP  - 2015 Dec 21
TI  - High performance volume-of-intersection projectors for 3D-PET image 
      reconstruction based on polar symmetries and SIMD vectorisation.
PG  - 9349-75
LID - 10.1088/0031-9155/60/24/9349 [doi]
AB  - For high-resolution, iterative 3D PET image reconstruction the efficient 
      implementation of forward-backward projectors is essential to minimise the 
      calculation time. Mathematically, the projectors are summarised as a system 
      response matrix (SRM) whose elements define the contribution of image voxels to 
      lines-of-response (LORs). In fact, the SRM easily comprises billions of non-zero 
      matrix elements to evaluate the tremendous number of LORs as provided by 
      state-of-the-art PET scanners. Hence, the performance of iterative algorithms, 
      e.g. maximum-likelihood-expectation-maximisation (MLEM), suffers from severe 
      computational problems due to the intensive memory access and huge number of 
      floating point operations. Here, symmetries occupy a key role in terms of 
      efficient implementation. They reduce the amount of independent SRM elements, 
      thus allowing for a significant matrix compression according to the number of 
      exploitable symmetries. With our previous work, the PET REconstruction Software 
      TOolkit (PRESTO), very high compression factors (>300) are demonstrated by using 
      specific non-Cartesian voxel patterns involving discrete polar symmetries. In 
      this way, a pre-calculated memory-resident SRM using complex 
      volume-of-intersection calculations can be achieved. However, our original 
      ray-driven implementation suffers from addressing voxels, projection data and SRM 
      elements in disfavoured memory access patterns. As a consequence, a rather 
      limited numerical throughput is observed due to the massive waste of memory 
      bandwidth and inefficient usage of cache respectively. In this work, an 
      advantageous symmetry-driven evaluation of the forward-backward projectors is 
      proposed to overcome these inefficiencies. The polar symmetries applied in PRESTO 
      suggest a novel organisation of image data and LOR projection data in memory to 
      enable an efficient single instruction multiple data vectorisation, i.e. 
      simultaneous use of any SRM element for symmetric LORs. In addition, the 
      calculation time is further reduced by using simultaneous multi-threading (SMT). 
      A global speedup factor of 11 without SMT and above 100 with SMT has been 
      achieved for the improved CPU-based implementation while obtaining equivalent 
      numerical results.
FAU - Scheins, J J
AU  - Scheins JJ
AD  - Institute of Neuroscience and Medicine, Forschungszentrum Julich GmbH, Germany.
FAU - Vahedipour, K
AU  - Vahedipour K
FAU - Pietrzyk, U
AU  - Pietrzyk U
FAU - Shah, N J
AU  - Shah NJ
LA  - eng
PT  - Journal Article
DEP - 20151118
PL  - England
TA  - Phys Med Biol
JT  - Physics in medicine and biology
JID - 0401220
SB  - IM
MH  - *Algorithms
MH  - Brain/*diagnostic imaging
MH  - Data Compression
MH  - Humans
MH  - Image Interpretation, Computer-Assisted/*methods
MH  - Image Processing, Computer-Assisted/methods
MH  - Imaging, Three-Dimensional/*methods
MH  - Pattern Recognition, Automated/methods
MH  - Positron-Emission Tomography/*instrumentation/*methods
MH  - Reproducibility of Results
MH  - Software
EDAT- 2015/11/19 06:00
MHDA- 2016/11/02 06:00
CRDT- 2015/11/19 06:00
PHST- 2015/11/19 06:00 [entrez]
PHST- 2015/11/19 06:00 [pubmed]
PHST- 2016/11/02 06:00 [medline]
AID - 10.1088/0031-9155/60/24/9349 [doi]
PST - ppublish
SO  - Phys Med Biol. 2015 Dec 21;60(24):9349-75. doi: 10.1088/0031-9155/60/24/9349. 
      Epub 2015 Nov 18.