Medical image resolution enhancement for healthcare using nonlocal self-similarity and low-rank prior
- 247 Downloads
Medical images have high information redundancy, which can be used to improve image analysis and visualization for purpose of healthcare. In order to recover a high-resolution (HR) image from its low-resolution (LR) counterpart, this paper proposes a resolution enhancement method by using the nonlocal self-similar redundancy and the low-rank prior. The proposed method consists of three main steps. First, an initial HR image is generated by nonlocal interpolation, which is based on the self-similarity of medical images. Next, the low-rank minimum variance estimator is exploited to reconstruct the HR image. At last, we iteratively apply the subsampling consistency constraint and perform the low-rank reconstruction to refine the reconstructed HR result. Experimental results conducted on MR and CT images demonstrate that the proposed method outperforms conventional interpolation methods and is competitive with the current stat-of-the-art methods in terms of both quantitative metrics and visual quality.
KeywordsResolution enhancement Low rank approximation Minimum variance estimation Nonlocal self-similarity Healthcare
This work is partially supported by National Natural Science Foundation (61572286, 61332015, and 61472220), Shandong Provincial Key Research and Development Plan (2017CXGC1504), Natural Science Foundation of Shandong Province (2016ZRB01143), and Fostering Project of Dominant Discipline an Talent Team of Shandong Province Higher Education. The authors also gratefully acknowledge the helpful comments and suggestions of the anonymous reviewers, which have improved the presentation significantly.
- 8.Dong W, Zhang L, Shi G, Wu X (2009) Nonlocal back-projection for adaptive image enlargement. In: Proceeding of IEEE International Conference on Image Processing, pp 349–352Google Scholar
- 11.Guo Q, Zhang C, Zhang Y, Liu H, Shen X (2015) Low-rank image denoising based on minimum variance estimator. J Comput-Aided Des Comput Graph 27(12):2237–2246. In ChineseGoogle Scholar
- 13.Guo Q, Gao S, Zhang X, Yin Y, Zhang C (2017) Patch-based image inpainting via two-stage low rank approximation. IEEE Trans Visualization and Computer Graphics, acceptedGoogle Scholar
- 15.He K, Sun J (2012) Computing nearest-neighbor fields via propagation-assisted kd-trees. In: Proceedings of IEEE International Conference on Computer Vision, pp 111–118Google Scholar
- 19.Hung KK, Siu Wc (2012) Single image super-resolution using iterative Wiener filter. In: Proceedings of IEEE International Conference on Acoustics Speech, Signal Processing, pp 1269–1272Google Scholar
- 22.Korman S, Avidan S (2011) Coherency sensitive hashing. In: Proceedings of IEEE International Conference on Computer Vision, pp 1607–1614Google Scholar
- 24.Larsen RM (1998) Lanczos bidiagonalization with partial reorthogonalization. DAIMI Rep Ser 537:1–101Google Scholar
- 31.Liu H, Geng F, Guo Q, Zhang C, Zhang C (2017) A fast weak-supervised pulmonary nodule segmentation method based on modified self-adaptive FCM algorithm. Soft Computer, acceptedGoogle Scholar
- 42.Tomasi C, Manduchi R (1998) Bilateral fitlering for gray and color images. In: Proceedings of IEEE International Conference on Computer Vision, pp 836–846Google Scholar