(Thursday, 7/18/2019) 7:30 AM - 9:30 AM
Room: 221AB
Purpose: To meet the clinical demand for low image noise in many CT applications, a thick image slice is often required. However, a thick image slice suffers from partial volume effect and obscures key anatomy and pathology. Here, we develop a prior-knowledge-aware-iterative-denoising (PKAID) technique which utilizes the spatial data redundancy in the slice increment direction to reduce image noise, while preserving resolution, noise texture and slice sensitivity profile. We demonstrate this technique in various clinical CT applications including non-contrast (NC) head and high-resolution wrist imaging.
Methods: A flexible image-domain denoising technique was developed which incorporated a low-noise prior image at the same location but thicker slice thickness to reduce noise in a thinner slice image. Phantom experiments were performed on a clinical CT using NC head protocol. Images were reconstructed with clinically-used thickness (5mm) and reduced thickness (2mm). The 2mm images were processed using PKAID with 5mm image as prior. The modulation transfer function (MTF), noise power spectra (NPS) and slice profile were derived. In vivo NC head images and thin-slice (0.25mm) high-resolution wrist images acquired on clinical CT and/or a whole-body photon-counting-detector (PCD)-CT were also used to demonstrate this technique.
Results: MTF and NPS showed that PKAID preserved image resolution and noise texture compared to original images. Slice profiles demonstrated that the original slice thickness was retained. In vivo brain images showed that PKAID reduced noise of 2mm image to a similar level as 5mm image, while better revealing the underlying pathology. In vivo wrist images on PCD-CT showed that PKAID allowed thin-slice high-resolution wrist imaging with lower noise compared to clinical CT.
Conclusion: We demonstrate a denoising technique that simultaneously meets the clinical demands for low noise, reduced partial volume effect and/or high imaging resolution, and can improve the diagnostic quality of clinical CT images in various areas.
Keywords
Taxonomy
IM- CT: General (Most aspects)
Contact Email
