Enhancing performance of electron holography with mathematical and machine learning-based denoising techniques

Area of Science:

• Materials Science
• Physics
• Image Processing

Background:

• Electron holography analyzes material properties like electromagnetic fields and strains.
• Performance is limited by 'shot noise' in electron micrographs (holograms).
• Advanced denoising methods offer potential to overcome these limitations.

Purpose of the Study:

• To provide an overview of advanced denoising techniques for electron holography.
• To explain the principles and usage of sparse coding, wavelet hidden Markov model, and tensor decomposition.
• To evaluate the denoising performance of these methods on simulated and experimental holograms.

Main Methods:

• Review of sparse coding principles and applications in electron holography.
• Explanation of the wavelet hidden Markov model for noise reduction.
• Description of tensor decomposition techniques for hologram analysis.
• Application and evaluation of these methods on simulated and real electron holograms.

Main Results:

• Denoising methods can extract signals previously obscured by noise.
• Evaluation results demonstrate the effectiveness of reviewed techniques.
• Comparison clarifies the impact of denoising on electron holography research.

Conclusions:

• Advanced denoising techniques are crucial for improving electron holography.
• Understanding the principles and careful application of these complex methods are necessary.
• Denoising significantly impacts the analysis of functional properties in materials and devices.