Impact of Filtering on Resolution and Signal-to-Noise Ratio in T2-Weighted

Abstract

Purpose: Image filtering affects image quality in magnetic resonance imaging (MRI). Different filtering techniques are applied in MRI, such as the K-space filter, signal filter, and pixel interpolation. This study aims to evaluate the effects of various filtering approaches on image resolution and signal-to-noise ratio (SNR).

Material and Methods: The modulation transfer function (MTF) for T2 weighted image (T2-w) is employed as a quantitative method to represent spatial resolution and signal-to-noise (SNR) ratio according to the National Electrical Manufacturers Association (NEMA). In this work, different filtering is applied prospectively. The ImageJ software employed to measures MTF, while the image QC determines SNR.

Results: the normal k-space without filter provides modulation transfer function at 10% (MTF10), the same as the raw and elliptical filters; the MTF10 was around 0.41. However, the MTF at 50% (MTF50) shows more differentiation between K-space filters. signal filters exhibit more difference at, MTF50, the scan without signal filter exhibit highest MTF50 that reach 0.29, while the MTF50 for B1 filter and pre-scan normalization around 0.16. The pixel interpolation mostly doesn’t affect MTF50 except for the no signal filter; the MTF50 exhibits pronounced reduction. For SNR, the elliptical filter exhibits lower SNR than the normal k-space, with the highest value equal to 123%. In comparison, the normal k-space without filter provides the highest value, equal to 130%, and the raw filter provides the highest SNR value, equal to 167%. For the signal filter effect on SNR, the B1 filter improves the SNR, while the broad range pre-scan normalisation exhibits less im-provement for SNR. For pixel interpolation, the SNR is mostly enhanced.

Conclusion: In conclusion, the study underscores the meticulousness required in selecting appropriate MRI filters. The significant influence of filters on image signal and resolution in MRI image processing cannot be excluded. Each filter choice must be carefully tailored to the specific imaging needs, considering the tradeoffs between resolution and SNR.