K-means clustering for support construction in diffractive imaging

Area of Science:

• Diffractive imaging
• Computational optics
• Image reconstruction

Background:

• Object support construction is crucial in diffractive imaging.
• Traditional methods often require manual adjustment of unknown parameters.
• Existing techniques may necessitate prior knowledge of the object support.

Purpose of the Study:

• To introduce a new, automated method for constructing object supports in diffractive imaging.
• To eliminate the need for manual parameter tuning in support construction.
• To develop a support construction technique compatible with established algorithms like the Gerchberg-Saxton diagram.

Main Methods:

• Utilizing K-means clustering applied to the object-intensity distribution.
• Dynamically generating the object support based on clustering results.
• Integrating the K-means based support construction with the Gerchberg-Saxton algorithm.

Main Results:

• The proposed K-means clustering method effectively constructs object supports.
• The method successfully removes the requirement for initial parameter adjustments.
• Numerical simulations demonstrate the dynamic and automated nature of the support construction.
• The technique allows for support generation without requiring an initial prior support.

Conclusions:

• The K-means clustering approach offers an effective and automated solution for object support construction in diffractive imaging.
• This method simplifies the diffractive imaging workflow by removing critical parameter dependencies.
• The dynamic support generation is advantageous for applications requiring adaptive or iterative reconstruction processes.