GiRR3D
GiRR3D (Generic iterative Retrospective Reconstruction in 3D) technology employs strong know-how of statistical priors of the CT noise distribution along with a state-of-the-art method for SNR enhancement to restore the image detail and quality of low-dose CT scans.
SafeCT® is based on a unique image processing technology that allows a dramatic enhancement of Signal-to-Noise Ratio (SNR) of low-dose CT studies. This capability can be leveraged to acquire CT studies at much lower doses. The low-dose, poor-quality scans are then processed to generate new CT images of diagnostic quality and visual acuity typically obtained with standard-dose protocols. The result is high-quality scans acquired utilizing only a fraction of the routine dose.
SafeCT’s image processing technology consists of a proprietary patented iterative image post-processing algorithm, GiRR3D (Generic iterative Retro Reconstruction in 3DTM), which dramatically enhances the SNR of low-dose CT images generated by conventional scanners. SafeCT has the potential for 80% dose reduction with no trade-off in terms of image quality and diagnostic value.
GiRR3D is an iterative algorithm for volumetric retrospective image reconstruction of low-dose CT scans, significantly reducing noise and restoring the fine details underneath the noise. GiRR3D begins with the low-dose poor-quality images as input; it employs strong know-how of statistical priors of the CT noise distribution along with a state-of-the-art method for SNR enhancement, preserving image detail, edges, and partial volume effects.
The SafeCT image below (right) shows the result of applying GiRR3D image processing to a noisy, low-dose CT image (below, left) of a phantom test target. Note the superior visual acuity of the SafeCT image in comparison with the conventional images of the phantom scanned at low dose, or even at full-dose (below, middle).
 |
 |
 |
|
Phantom scanned at 60mAs |
Phantom scanned at 300mAs |
SafeCT output for the 60mAs phantom scan (80% dose reduction)
|
GiRR3D is extremely powerful and robust, and applicable to all types of CT protocols. The algorithms are implemented on dedicated off-the-shelf scalable computing hardware, consisting of multiple Graphical Processing Units (GPU’s), enabling fast parallel iterative image reconstruction with a runtime of 1.5 minutes per full study.
Clinical results from phantom images, clinical images, and radiologist evaluations clearly demonstrate the power and quality of SafeCT.
|
