Sparse Poisson Denoising of Microbial Images Based on Non-local Principal Component Analysis
-
摘要: 针对微生物显微图像去噪,提出一种基于图像稀疏块表示和字典学习的泊松去噪算法。根据微生物图像内在相关性进行分块处理,采用Poisson K-均值法对图像块进行聚类;运用主成分分析法实现非局部稀疏字典表示,完成簇内去噪;经融合重建,获得完整去噪图像。结果表明:通过稀疏块表示和字典学习直接对泊松噪声去噪,可减少噪声模型转换误差;改进的分块和聚类方法可提高去噪图像的信噪比;与其他去噪算法对比,本文算法不仅取得更好的去噪效果,且可改善去噪后图像模糊现象,最大程度地保留图像细节信息。
-
关键词:
- 微生物 /
- 图像去噪 /
- 主成分分析 /
- 稀疏表示 /
- Poisson K-均值聚类
Abstract: For the denoising of micro image of microorganism, a Poisson denoising algorithm based on image sparse block representation and dictionary learning was proposed. Firstly, according to the intrinsic correlation of microbial images, the image blocks were clustered by Poisson K-means method; Secondly, the non-local sparse dictionary representation was realized by principal component analysis to complete the intra cluster denoising; Finally, the complete denoised image was obtained by fusion reconstruction. The results show that through sparse block representation and dictionary learning, Poisson can be denoised directly, which can reduce the error of noise model transformation;The improved block and clustering methods can greatly improve the signal-to-noise ratio of denoised image. Compared with other denoising algorithms, the proposed method not only achieves better denoising effect, but also significantly improves the image blur quality, and retains more image details to the maximum extent. -
-
[1] 武治, 孙照旋, 周芳. 基于改进萤火虫算法的多阈值微生物图像分割[J]. 安徽工业大学学报(自然科学版), 2020, 37(1):46-52, 59. [2] AZZARI L, FOI A. Variance stabilization for noisy+estimate combination in iterative Poisson denoising[J]. IEEE Signal Processing Letters, 2016, 23(8):1086-1090.
[3] ZHANG M H, ZHANG F Q, LIU Q G, et al. VST-Net:variance-stabilizing transformation inspired network for Poisson denoising[J]. Journal of Visual Communication and Image Representation, 2019, 62:12-22.
[4] 徐苏, 周颖玥. 基于图像分割的非局部均值去噪算法[J]. 计算机应用,2017, 37(7):2078-2083. [5] 曹硕, 黄利萍, 侯倍倍, 等. 基于模糊边缘补足的自适应非局部均值图像去噪算法[J]. 激光与光子学进展, 2018, 55(1):011003. [6] 张俊, 马明溪, 宁成臻, 等. 分数阶整体变分泊松去噪模型的分裂Bregman方法[J]. 南昌工程学院学报, 2018, 37(4):57-60. [7] 张峥嵘, 刘红毅, 韦志辉. 欧拉弹性正则化的图像泊松去噪[J]. 电子学报, 2017, 45(1):181-191. [8] 汪浩然, 夏克文, 任苗苗, 等. 结合PCA及字典学习的高光谱图像自适应去噪方法[J]. 计算机应用, 2016, 36(12):3411-3417, 3422. [9] XU J, ZHANG L, ZHANG D. External prior guided internal prior learning for real-world noisy image denoising[EB/OL]. 2017:arXiv:1705.04505[cs.CV]. https://arxiv.org/abs/1705.04505.
[10] 印佳, 杜战战. 基于主成分分析的高光谱遥感图像非局部去噪[J]. 现代电子技术, 2015, 38(11):70-72. [11] SALMON J, HARMANY Z, DELEDALLE C A, et al. Poisson noise reduction with non-local PCA[J]. Journal of Mathematical Imaging and Vision, 2014, 48(2):279-294.
[12] BANERJEE A, MERUGU S, DHILLO N, et al.Clustering with bregman divergences[J]. J Mach Learn Res, 2005, 6:1705-1749.
[13] ROY N, GORDON G, THRUN S.Finding approximate POMDP solutions through belief compression[J]. Journal of Artificial Intelligence Research, 2005, 23:1-40.
[14] HARMANY Z F, MARCIA R F, WILLETT R M.This is SPIRAL-TAP:sparse poisson intensity reconstruction algorithms:theory and practice[J]. IEEE Trans Image Process, 2012, 21(3):1084-1096.
[15] DABOV K, FOI A, KATKOVNIK V, et al. Image denoising by sparse 3-D transform-domain collaborative filtering[J]. IEEE Transactions on Image Processing, 2007, 16(8):2080-2095.
[16] SALMON J, STROZECKI Y. Patch reprojections for non local methods[J]. Signal Processing, 2012, 92(2):447-489.
[17] 杨淑莹. 图像识别与项目实践:VC++、MATLAB技术实现[M]. 北京:电子工业出版社, 2014:34-35.
计量
- 文章访问数: 65
- HTML全文浏览量: 0
- PDF下载量: 2