idose 4  (Philips Healthcare)

Journal: Journal of the Korean Society of Radiology

Article Title: Additional Dose Reduction Potential of Vendor-Agnostic Deep Learning Model: A Phantom Study

doi: 10.3348/jksr.2025.0047

Figure Lengend Snippet: A. CT images of a phantom were acquired at multiple doses using two CT machines equipped with their IR methods: ADMIRE from Siemens, and iDose 4 and IMR from Philips. A vendor-agnostic DLM was applied to each image, and we compared the original image with the DLM-applied image. B. Multiple scans were performed at each dose to obtain reliable results, and these were averaged into a single image. ADMIRE = advanced modeled iterative reconstruction, DLM = deep learning model, iDose 4 = iterative dose, IMR = iterative model reconstruction, IR = iterative reconstruction

Article Snippet: The acquired phantom images were reconstructed using three IR methods from two vendors: Advanced Modeled Iterative Reconstruction (ADMIRE, Siemens Healthineers), iDose 4 (Philips), and Iterative Model Reconstruction (IMR, Philips).

Techniques:

Journal: Journal of the Korean Society of Radiology

Article Title: Additional Dose Reduction Potential of Vendor-Agnostic Deep Learning Model: A Phantom Study

doi: 10.3348/jksr.2025.0047

Figure Lengend Snippet: ADMIRE = advanced modeled iterative reconstruction, DLM = deep learning model, D-ADMIRE 3 = DLM-applied ADMIRE level 3, D-ADMIRE 5 = DLM-applied ADMIRE level 5, D-iDose 4 4 = DLM applied iDose 4 level 4, D-iDose 4 6 = DLM applied iDose 4 level 6, D-IMR 1 = DLM applied IMR level 1, D-IMR 3 = DLM applied IMR level 3, iDose 4 = iterative dose, IMR = iterative model reconstruction

Article Snippet: The acquired phantom images were reconstructed using three IR methods from two vendors: Advanced Modeled Iterative Reconstruction (ADMIRE, Siemens Healthineers), iDose 4 (Philips), and Iterative Model Reconstruction (IMR, Philips).

Techniques:

Journal: Journal of the Korean Society of Radiology

Article Title: Additional Dose Reduction Potential of Vendor-Agnostic Deep Learning Model: A Phantom Study

doi: 10.3348/jksr.2025.0047

Figure Lengend Snippet: A. DRP of D-IRs categorized by inlet type and IR method. The x-axis represents different inlets with various IR methods applied, while the y-axis shows the DRP, measured as a percentage. B. DRP results based on target diameter for each inlet and IR method. The x-axis indicates target diameters of 1, 5, and 10 mm, while the y-axis represents the DRP, measured as a percentage. ADMIRE = advanced modeled iterative reconstruction, D-ADMIRE 3 = DLM-applied ADMIRE level 3, D-ADMIRE 5 = DLM-applied ADMIRE level 5, D-iDose 4 4 = DLM applied iDose 4 level 4, D-iDose 4 6 = DLM applied iDose 4 level 6, D-IMR 1 = DLM applied IMR level 1, D-IMR 3 = DLM applied IMR level 3, DLM = deep learning model, DRP = dose reduction potential, HU = Hounsfield unit, iDose 4 = iterative dose, IMR = iterative model reconstruction, IR = iterative reconstruction

Article Snippet: The acquired phantom images were reconstructed using three IR methods from two vendors: Advanced Modeled Iterative Reconstruction (ADMIRE, Siemens Healthineers), iDose 4 (Philips), and Iterative Model Reconstruction (IMR, Philips).

Techniques:

Journal: Journal of the Korean Society of Radiology

Article Title: Additional Dose Reduction Potential of Vendor-Agnostic Deep Learning Model: A Phantom Study

doi: 10.3348/jksr.2025.0047

Figure Lengend Snippet: A. The CNR increment was positive for all inlet types and IR methods. The x-axis represents the percentage of dose relative to the reference dose, and the y-axis shows the percentage increase in CNR of the DLM-applied image compared with the original image at the reference dose. B. The x-axis represents the percentage of dose relative to the reference dose, and the y-axis shows the percentage increase in TTF of the DLM-applied image compared with the original image at the reference dose. ADMIRE = advanced modeled iterative reconstruction, CNR = contrast-to-noise ratio, D-ADMIRE 3 = DLM-applied ADMIRE level 3, D-ADMIRE 5 = DLM-applied ADMIRE level 5, D-iDose 4 4 = DLM applied iDose 4 level 4, D-iDose 4 6 = DLM applied iDose 4 level 6, D-IMR 1 = DLM applied IMR level 1, D-IMR 3 = DLM applied IMR level 3, DLM = deep learning model, HU = Hounsfield unit, iDose 4 = iterative dose, IMR = iterative model reconstruction, TTF = task-based transfer function

Article Snippet: The acquired phantom images were reconstructed using three IR methods from two vendors: Advanced Modeled Iterative Reconstruction (ADMIRE, Siemens Healthineers), iDose 4 (Philips), and Iterative Model Reconstruction (IMR, Philips).

Techniques:

Journal: Radiographics

Article Title: State-of-the-Art Deep Learning CT Reconstruction Algorithms in Abdominal Imaging

doi: 10.1148/rg.240095

Figure Lengend Snippet: Precise Image reconstruction algorithm. Axial contrast-enhanced CT images of the using FBP (A) , iDose 4 hybrid IR technique (Philips Healthcare) (B) , and DLR with Precise Image AI (C) show an increasing SNR and CNR from A to C . Zoomed inset images of two hypoattenuating liver lesions show dramatically improved lesion delineation using the Precise Image DLR technique (C) .

Article Snippet: Axial contrast-enhanced CT images of the using FBP (A) , iDose 4 hybrid IR technique (Philips Healthcare) (B) , and DLR with Precise Image AI (C) show an increasing SNR and CNR from A to C . Zoomed inset images of two hypoattenuating liver lesions show dramatically improved lesion delineation using the Precise Image DLR technique (C) .

Techniques: