CT Reconstruction Video Tutorials

This video series demonstrates the use of the CT Reconstruction plug-in. The videos cover all topics necessary to reconstruct image data from projections, and includes both cone beam and parallel beam acquisitions, preprocessing filters and geometry calibration/correction to obtain the best possible image quality.

You can also view these videos and others on our YouTube channel (https://www.youtube.com/@ObjectResearchSystems).

This introductory tutorial demonstrates how to reconstruct cone-beam projection data.

Cone Beam CT Reconstruction (22:40)

Topics Covered in the Video

The following topics are discussed in the video. Links to Help topics with further information are also provided.

  • Loading the CT image files (00:28). See also Importing Image Files.

    Make sure that the XY plane is the detector plane and the rotation axis is along the Y axis.

    Note The steering arm dataset is shared by Laboratoire sur les alliages à mémoire et systèmes intelligents (LAMSI), École de technologie supérieure.

  • Opening the CT Reconstruction plug-in and selecting the loaded image files as the 'projection dataset' (04:32).

  • Selecting 'Cone Beam' as beam type and entering the acquisition parameters. You can import metadata files from different vendors, or manually enter the required parameters (04:59).

    Note Example metadata files from Nikon, YXLON, Skyscan-Bruker, and KA Imaging are shown.

  • Selecting the reconstruction engine (09:10).

  • Reconstructing, computing previews, as well as loading and saving the reconstruction (12:08).

This introductory tutorial demonstrates how to reconstruct parallel-beam projection data.

Parallel Beam CT Reconstruction (10:32)

Topics Covered in the Video

The following topics are discussed in the video. Links to Help topics with further information are also provided.

  • Opening the CT Reconstruction plug-in and selecting the loaded image files as the 'projection dataset' (02:10).

  • Selecting 'Parallel Beam' as beam type and entering the acquisition parameters (04:59).

  • Selecting pre-processing filters.

  • Reconstructing the data (06:26).

This tutorial demonstrates how to find the rotation center and tilt angle of CT datasets. Two examples are presented, one for parallel beam and the other for cone beam.

Geometry Calibration in CT Reconstruction (35:07)

Topics Covered in the Video

The following topics are discussed in the video. Links to Help topics with further information are also provided.

  • A ceramic composites dataset is used to demonstrate the rotation center finder algorithms exclusively for parallel beam dataset (00:49). The dataset was shared by Aly Badran from University of Colorado (https://link.springer.com/article/10.1007/s10853-020-05148-7).

    • Tomopy Nghia Vo’s method (02:06).

    • Tomopy image entropy error method (02:23).

    • Tomopy phase correlation method (02:30).

  • An engine part dataset is reconstructed with beam hardening correction (11:21). Thanks to Denis Cormier from Rochester Institute of Technology for sharing this dataset.

    • Manual range (12:50).

    • Image metric method (13:48).

    • Sum of projection method (in tilt angle finder) (27:00). This method can be used for determining both the rotation center offset and the tilt angle.

      Reference: Meng, Y., et al. Online Geometric Calibration of Cone-Beam Computed Tomography for Arbitrary Imaging Objects. IEEE transactions on medical imaging 32(2): 278-288, 2013.

This tutorial demonstrates how to use preprocessing filters to remove artifacts and improve image quality of the reconstructed images.

Pre-Processing Filters (34:19)

Topics Covered in the Video

The following topics are discussed in the video. Links to Help topics with further information are also provided.

  • A ceramic composites dataset is reconstructed with different ring artifact removal filters and median filter. The dataset was shared by Aly Badran from University of Colorado (https://link.springer.com/article/10.1007/s10853-020-05148-7).

    • Flat field correction (02:54).

    • Wavelet-Fourier stripe removal (07:18).

    • Titarenko stripe removal (10:43).

    • Median filter (12:42).

  • A battery dataset is presented (16:20) to show the multi-point piecewise flat field correction (17:53) and defective pixel correction (19:48). The data is shared by Pratt & Whitney Canada.

  • An engine part dataset is reconstructed with beam hardening correction (26:08). Thanks to Denis Cormier for sharing this dataset.

  • Introduction of the phase retrieval filter (31:55) and the log filter (33:00).