Adaptive Least-Squares Approximation for Noisy MR Image Interpolation
0
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
In this paper, we present a data-adaptive least-squares method for interpolating noisy magnetic resonance (MR) images. The proposed method generates a denoised high-resolution output from a low-resolution MR image contaminated by additive Gaussian noise. To achieve both noise suppression and the preservation of important image features, such as edges and fine structures, we first construct least-squares approximation operators that incorporate penalties based on task-specific distance metrics, and then adaptively combine these operators according to the characteristics of image regions. Our algorithm can estimate unknown values at arbitrary query points, making it particularly useful for complex restoration tasks in medical imaging. The effectiveness of the proposed method in maintaining image quality while reducing noise and artifacts is demonstrated through experiments on simulated brain MR images.
Abstract
In this paper, we present a data-adaptive least-squares method for interpolating noisy magnetic resonance (MR) images. The proposed method generates a denoised high-resolution output from a low-resolution MR image contaminated by additive Gaussian noise. To achieve both noise suppression and the preservation of important image features, such as edges and fine structures, we first construct least-squares approximation operators that incorporate penalties based on task-specific distance metrics, and then adaptively combine these operators according to the characteristics of image regions. Our algorithm can estimate unknown values at arbitrary query points, making it particularly useful for complex restoration tasks in medical imaging. The effectiveness of the proposed method in maintaining image quality while reducing noise and artifacts is demonstrated through experiments on simulated brain MR images.
참고문헌 (15)
- Greenspan H, "Super-resolution in medical imaging", Comput J, 52, pp.43~63 (2009)
- Irani M, "Improving resolution by image registration", CVGIP, Graph Models Image Process, 53, pp.231~239 (1991)
- Marquina A, "Image super-resolution by TV regularization and Bregman iteration", J Sci Comput, 37, pp.367~382 (2008)
- Brudfors M, "Medical image understanding and analysis. MIUA" (2018), Springer
- Shi F, "LRTV : MR image super-resolution with low-rank and total variation regularizations", IEEE Trans Med Imaging, 34, pp.2459~2466 (2015)
- Shi J, "MR image super-resolution via wide residual networks with fixed skip connection", IEEE J Biomed Health Inform, 23, pp.1129~1140 (2019)
- Zhang Y, "MR image super-resolution with squeeze and excitation reasoning attention network", 2021 IEEE, 2021, 13420~13429
- de Leeuw den Bouter ML, "Deep learning-based single image super-resolution for low-field MR brain images", Sci Rep, 12, pp.6362~ (2022)
- Ko K, "MRIFlow : Magnetic resonance image super-resolution based on normalizing flow and frequency prior", J Magn Reson, 352, pp.107477~ (2023)
- Levin D, "The approximation power of moving least-squares", Math Comput, 67, pp.1517~1531 (1998)
- Buades A, "A non-local algorithm for image denoising", IEEE Int’l Conf. on Computer Vision and Pattern Recognition, 2, pp.60~65 (2005)
- Manjón JV, "Non-local MRI upsampling", Med Image Anal, 14, pp.784~792 (2010)
- Boyd S, "Distributed optimization and statistical learning via the alternating direction method of multipliers", Found Trends Mach Learn, 3, pp.1~122 (2011)
- Zhao X, "Channel splitting network for single mr image super-resolution", IEEE Trans Image Process, 28, pp.5649~5662 (2019)
- "BrainWeb: Simulated Brain Database" (2024)[URL]