Congress:
ECR25
Poster Number:
C-20882
Type:
Poster: EPOS Radiologist (scientific)
Authorblock:
A. Van Der Most1, C. Balta2, K. Bishop2, J. Heemskerk1, A. J. Van Der Molen1, I. I. A. Dekkers1; 1Leiden/NL, 2Amstelveen/NL
Disclosures:
Arjen Van Der Most:
Nothing to disclose
Christiana Balta:
Employee: Canon Medical Systems Europe
Kelly Bishop:
Employee: Canon Medical Systems Europe
Jan Heemskerk:
Nothing to disclose
Aart J. Van Der Molen:
Nothing to disclose
Ilona I. A. Dekkers:
Grant Recipient: Canon Medical Systems Europe
Keywords:
Radioprotection / Radiation dose, CT, CT-High Resolution, Image manipulation / Reconstruction, Equipment, Physics, Quality assurance
Compared to our standard reconstruction method, Deep-learning-based iterative reconstruction algorithm (AiCE),the new super-resolution deep-learning image reconstruction (SL-DLR) algorithm PIQE can simultaneously improve noise characteristics and image sharpness of CT images at varying dose levels. For both AiCE and PIQE increasing the strength of the algorithm from Mild to Strong suppresses noise at slight cost of resolution, Therefore, for optimal image quality, it might be beneficial to increase the strength of AiCE and PIQE when moving towards lower dose values.