Centering Noisy Images with Application to Cryo-EM.

Area of Science:

• Image analysis
• Computational biology
• Biophysics

Background:

• Estimating the center of mass in noisy 2D images is challenging.
• Standard methods like direct computation and geometric median are vulnerable to noise.
• Accurate centering is crucial for downstream analysis, particularly in cryo-electron microscopy (cryo-EM).

Purpose of the Study:

• To develop a novel, noise-robust method for estimating the center of mass in 2D images.
• To provide a mathematical and numerical analysis of the proposed surrogate function.
• To demonstrate the method's effectiveness in single-particle cryo-electron microscopy (cryo-EM) for improved translational alignment.

Main Methods:

• Definition of a novel surrogate function for the center of mass.
• Mathematical and numerical analysis of the proposed function's properties.
• Application of the centering method to single-particle cryo-electron microscopy (cryo-EM) data.

Main Results:

• The novel method outperforms existing techniques for center of mass estimation in noisy images.
• The approach demonstrates superior performance in realistic scenarios.
• Improved translational alignment of molecule images in cryo-EM data was achieved.

Conclusions:

• The developed centering method offers a robust solution for noisy image analysis.
• It significantly enhances the translational alignment of images in cryo-EM.
• This facilitates reconstruction, streamlines the cryo-EM pipeline, saves computational time, and supports resolution enhancement.