Abstract:Single-molecule localization microscopy techniques, like stochastic optical reconstruction microscopy (STORM), visualize biological specimens by stochastically exciting sparse blinking emitters. The raw images suffer from unwanted background fluorescence, which must be removed to achieve super-resolution. We introduce a sparsity-based background removal method by adapting a neural network (SLNet) from a different microscopy domain. The SLNet computes a low-rank representation of the images, and then, by subtracting it from the raw images, the sparse component is computed, representing the frames without the background. We compared our approach with widely used background removal methods, such as the median background removal or the rolling ball algorithm, on two commonly used STORM datasets, one glial cell, and one microtubule dataset. The SLNet delivers STORM frames with less background, leading to higher emitters' localization precision and higher-resolution reconstructed images than commonly used methods. Notably, the SLNet is lightweight and easily trainable (<5 min). Since it is trained in an unsupervised manner, no prior information is required and can be applied to any STORM dataset. We uploaded a pre-trained SLNet to the Bioimage model zoo, easily accessible through ImageJ. Our results show that our sparse decomposition method could be an essential and efficient STORM pre-processing tool.

What problem does this paper attempt to address?

### Problems Addressed by the Paper This paper aims to address the issue of background fluorescence signals affecting high-resolution imaging in single-molecule localization microscopy techniques (such as STORM). Specifically, during STORM imaging, non-specific fluorescence background interference can affect the precise localization of emitter centers, thereby reducing image resolution and reconstruction quality. Therefore, removing background signals is a key preprocessing step to achieve high-precision and high-resolution STORM image reconstruction. ### Background and Methods Traditional background removal methods, such as median background removal, rolling ball algorithm, or image filtering, although effective to some extent, have issues with poor generality or high computational cost. These methods usually assume that the background is uniform, whereas in reality, the background in fluorescence microscopy is non-uniform and heterogeneous. To address this, the authors propose a background removal method based on sparse decomposition, using an adaptive neural network (SLNet) to calculate and remove background signals. ### Main Contributions 1. **Sparse Decomposition Method**: This method uses sparse low-rank decomposition techniques to decompose the original image into low-rank components (representing the background) and sparse components (representing the sample signal). By subtracting the low-rank components from the original image, a sparse image with the background removed is obtained. 2. **SLNet Neural Network**: SLNet is a lightweight convolutional neural network that can quickly compute low-rank representations and is trained through self-supervised learning. This network does not require prior information during training and is applicable to any STORM dataset. 3. **Performance Evaluation**: The authors compared the SLNet method with commonly used background removal methods (such as median background removal and rolling ball algorithm). The results show that the SLNet method excels in background removal, improving emitter localization accuracy, and enhancing image resolution. ### Experimental Results 1. **Qualitative Analysis**: By comparing STORM images processed by different methods, the images generated by the SLNet method have less background, better separation of emitter center positions, and clearer images. 2. **Quantitative Analysis**: Using FWHM (Full Width at Half Maximum) and SQUIRREL analysis methods for quantitative evaluation, the results show that the SLNet method outperforms other methods on two commonly used datasets, with lower FWHM values and higher linearity. ### Conclusion This paper proposes a background removal method based on sparse decomposition, effectively addressing the background interference problem in STORM imaging through the adaptive SLNet neural network. This method not only has high computational efficiency but also excels in background removal, improving emitter localization accuracy, and enhancing image resolution, making it suitable for detailed and overlapping structure sample studies.

Sparsity-based background removal for STORM super-resolution images

Accurate background reduction in adaptive optical 3D-STED nanoscopy by dynamic phase switching

Sub-Diffraction Imaging with Confocal Fluorescence Microscopy by Stochastic Photobleaching

Untrained Neural Network Enabling Fast and Universal Structured-Illumination Microscopy

Universal and High-Fidelity Resolution Extending for Fluorescence Microscopy Using a Single-Training Physics-Informed Sparse Neural Network

BGnet: Accurate and rapid background estimation in single-molecule localization microscopy with deep neural nets

STORM image denoising and information extraction

Deep Learning for Super-Resolution Localization Microscopy

Trade-offs between spatial and temporal resolutions in stochastic super-resolution microscopy techniques

Enhancing brain image quality with 3D U-net for stripe removal in light sheet fluorescence microscopy

Super-resolution Image Reconstruction for High-Density 3D Single-Molecule Microscopy.

Learning to Reconstruct Confocal Microscopy Stacks from Single Light Field Images

SparseSSP: 3D Subcellular Structure Prediction from Sparse-View Transmitted Light Images

Self-STORM: Deep Unrolled Self-Supervised Learning for Super-Resolution Microscopy

Accurate Background Reduction in Adaptive Optical Three-Dimensional Stimulated Emission Depletion Nanoscopy by Dynamic Phase Switching

Deep learning using a residual deconvolutional network enables real-time high-density single-molecule localization microscopy

Progress in Far-Field Super-Resolution Stochastic Optical Reconstruction Microscopy(storm)

Super-STORM: Molecular Modeling to Achieve Single-molecule Localization with STORM Microscopy.

Robust single-shot 3D fluorescence imaging in scattering media with a simulator-trained neural network

Rapid and Extreme Low-light Superresolution Imaging Via Artificial Intelligence

Spatial Covariance Reconstructive (SCORE) Super-Resolution Fluorescence Microscopy