Structure assisted compressed sensing reconstruction of undersampled AFM images

Area of Science:

• Atomic Force Microscopy (AFM)
• Image Reconstruction
• Compressed Sensing

Background:

• Atomic Force Microscopy (AFM) generates high-resolution surface images.
• Compressed sensing (CS) offers potential for faster AFM imaging and reduced probe-specimen interaction.
• Current CS methods for AFM can be limited in reconstruction quality.

Purpose of the Study:

• To investigate the inherent structure within AFM images to improve compressed sensing reconstruction.
• To develop a novel structure model and modify existing algorithms for enhanced AFM image reconstruction.
• To quantitatively and qualitatively demonstrate the improvements offered by the proposed structured model.

Main Methods:

• Analysis of structure in discrete cosine transform (DCT) coefficients of typical AFM images.
• Development of a generic support structure model based on observed AFM image properties.
• Modification of established iterative thresholding algorithms to incorporate the proposed structure model.

Main Results:

• Demonstrated that inherent structure in AFM images can significantly improve reconstruction quality.
• The proposed structure model and modified algorithms enhance the fidelity of reconstructed AFM images.
• The new method achieves comparable or better reconstruction quality with fewer measurements than traditional methods.

Conclusions:

• Leveraging inherent image structure through a novel model significantly advances compressed sensing in AFM.
• The developed algorithm provides a more efficient and effective approach to AFM image reconstruction.
• This work paves the way for faster, higher-detail AFM imaging with reduced experimental burden.