Richardson-Lucy deconvolution as a general tool for combining images with complementary strengths

Area of Science:

• Microscopy
• Image Processing
• Computational Imaging

Background:

• Modern fluorescence microscopy generates complex datasets requiring advanced image combination techniques.
• Multiview light-sheet, localization, structured illumination (SIM), image scanning microscopy (ISM), and gated stimulated-emission depletion (STED) microscopy present unique challenges for image merging.
• Existing methods for merging images with varying point-spread functions or signal levels can be mathematically complex or limited in scope.

Purpose of the Study:

• To demonstrate the versatility and effectiveness of Richardson-Lucy (RL) deconvolution for merging diverse microscopy image datasets.
• To showcase RL deconvolution's ability to combine images with different point-spread functions, noise levels, and illumination patterns.
• To provide a simpler, yet powerful, alternative to existing algorithms for image reconstruction in advanced microscopy.

Main Methods:

• Application of Richardson-Lucy (RL) deconvolution algorithm to simulated microscopy image data.
• Testing RL deconvolution on datasets simulating multiview light-sheet microscopy, localization microscopy, SIM, ISM, and gated STED microscopy.
• Comparison of RL deconvolution results with standard inversion algorithms for ISM data.

Main Results:

• RL deconvolution successfully merged images with highly variable point-spread functions, preserving resolution.
• The method effectively combined high-resolution, high-noise images with low-resolution, low-noise images.
• Reconstructions for ISM using RL deconvolution were comparable in quality to standard methods but simpler to implement.
• RL deconvolution achieved high-quality merges for gated STED microscopy data, even when non-iterative algorithms are unknown.

Conclusions:

• Richardson-Lucy (RL) deconvolution is a highly versatile and effective tool for merging diverse microscopy image datasets.
• The technique offers significant advantages in simplicity and performance across various advanced microscopy modalities.
• RL deconvolution provides a robust solution for image fusion, enhancing resolution and data quality in challenging imaging scenarios.