[Progress in filters for denoising cryo-electron microscopy images].

Area of Science:

• Structural Biology
• Biophysics
• Microscopy

Background:

• Cryo-electron microscopy (cryo-EM) bridges cellular and molecular biology, revealing macromolecular and cellular complex structures.
• Despite breakthroughs like direct detection devices, radiation damage limits resolution, resulting in noisy images with low signal-to-noise ratios (SNR).
• Effective noise reduction is essential for improving cryo-EM resolution and enabling accurate 3D reconstruction and in situ analysis.

Purpose of the Study:

• To systematically review and compare robust filters for noise reduction in cryo-electron microscopy (cryo-EM).
• To evaluate filter performance in single-particle analysis (SPA) and cryo-electron tomography (cryo-ET) applications.
• To identify methods that balance noise smoothing with preservation of fine structural details.

Main Methods:

• Review of conventional spatial filters: Gaussian, median, and bilateral filters.
• Analysis of wavelet transform methods for joint spatial and frequency domain noise reduction.
• Comparison of filter efficacy in managing low SNR cryo-EM images and their impact on resolution.

Main Results:

• Gaussian filters blur fine features; median filters risk aliasing.
• Bilateral filters maintain edges but are limited at very low SNR.
• Wavelet transforms offer superior noise reduction and detail preservation, outperforming conventional spatial filters.
• A modified wavelet shrinkage filter demonstrated significant improvements in image quality.

Conclusions:

• Noise reduction remains a critical challenge in cryo-EM for achieving higher resolution.
• Wavelet-based methods show significant potential for enhancing cryo-EM image quality and structural interpretation.
• This review provides insights into filter selection for optimizing cryo-EM data analysis and biological structure determination.