Cyber Technology, Inc.

Image Acquisition, Reconstruction, and Presentation
  for 3-D Computed Tomography (CT)

Simulations for High Efficiency CT (HECT)™, in TEM

For performance evaluation a 256 x 256 x 64 voxel phantom was created. Some 100 cubes of size 6 x 6 x 6 voxels of density X = 5 were placed at random locations within that slab of density Y = 0.80. In order to simulate the quantum counting statistics of the transmission electron imaging system the projection
of
these cubes were superimposed by noise.  The rms-noise level in the 0-degree projection is about 25% of the intensity of projected cubes. A total of 49 views
at three degree increment were used for the reconstruction. These projection test data and the SIRT reconstruction were kindly provided by the National Institute
of Biomedical Imaging and  Bioengineering
(NIBIB / NIH) (generated by M. Aronova). HECT reconstruction was performed by Cyber Technology, Inc.

All renditions (C) Wolfram R. Jarisch, 2008

HECT versus SIRT CT for Transmission Electron Microscopy

Images

Original noisy projection data: three of 49 selected projections at -72, 0, and +72 degree tilt angles rendered with IMOD2.

+/- 72 degree range, 3 degree increment

Comparison of HECT reconstruction with SIRT. The residuals are shown for 15 iterations with SIRT (left) and a single iteration with HECT (right). The selected view is at 0 degree tilt. Observe that the residuals with HECT contain a much smaller trace of the cubes relative to the noise when compared to SIRT, implying the effective extraction of the information about the cube density and the surrounding matrix.

RESIDUALS: SIRT vs. HECT

IMOD slice-views through the 3-D HECT reconstruction of the slab.

IMOD slice of 3-D Reconstruction after 1 Iteration

2 Kremer J.R., D.N. Mastronarde and J.R. McIntosh, Computer visualization of three-dimensional image data using IMOD, J. Struct. Biol. 116 (1996) 71-76.


Home   |   About Us    |   Feedback Form   |   Site Map
Patent Pending