概要：1. Convolutional neural networks (CNN)show very good performance and achieve impressive results in biomedical image denoising.Nevertheless CNN-based methods strongly depend on the used training set and even small differences in the input data can cause output disturbance.Thus new more reliable hybrid denoising methods were suggested.They include combinations of CNN and «classical»algorithms like Non-Local Means,BM3D,Bilateral,Anisotropic diffusion,Total Variation(TV),etc.However hybrid methods need non-reference automatic parameters estimation for classical algorithms.In this papaer we presenta a hybrid DnCNN+BM3D method with automatic choice of the strength parameter for BM3D method.To contorol biomedical image structures by multiscale ridge difference between noisy and filtered images.Test results for different image datasets show practical applicability of the method. Some possible applications of he hybrid methods for blinking fluorescene imaging enhancement will be presented.
2. A probabilistic approach for super-resolution of blinking fluorescence microscopy will be presented.Its performance will be compared with modern blinking fluorescence image enhancement alorithms,namely SOFI,MUSICAL and SPARCOM in different conditions.The comparison will be performed usin both synthetic and real experimental data.The possible future work on blinking fluorescence imaging enhancement will be discussed.
関連リンク(大阪理研)