Keywords:
Radiation physics, Image manipulation / Reconstruction, Technology assessment, Image verification
Authors:
T. Higaki1, E. Nishimaru1, Y. Nakamura1, F. Tatsugami1, Z. Yu2, J. Zhou2, A. Prabhu Verleker3, N. Akino3, K. Awai1; 1Hiroshima/JP, 2Vernon Hills, IL/US, 3Otawara/JP
DOI:
10.1594/ecr2018/C-1656
Methods and materials
Figure 1 shows outline of the DLR technique. Filtering processes which remove image noise and artifacts to improve signal reliability are placed to image reconstruction processes. The DCNNs were trained by pairs of dataset,
one of the pairs was the teaching data acquired under ideal condition,
and the other was the training data including noise and artifacts. The teaching dataset were acquired with high tube current,
large number of views,
small cone beam angle,
and environment with less beam hardening effect.
We scanned 200 mm of acrylic cylinder phantom using 320-row detector CT with 120 kV of tube voltage. The referenced tube current was determined to 100 mA using automatic tube current modulation. To evaluate the capability of the DLR for radiation dose reduction,
the phantom was scanned with the tube current of from 10 to 100 mA at an interval of 10 mA. Image dataset were reconstructed by hybrid iretative reconstruction (hybrid IR) (AIDR 3D, Canon Medical Systems Corporation) and the DLR. We evaluated noise characteristics of the dataset using noise power spectrum (NPS) and standard deviation (SD) of CT number.