Low-frequency background estimation and noise separation from high-frequency for background and noise subtraction

Area of Science:

• Microscopy
• Image Processing
• Biophotonics

Background:

• Background blur and noise limit signal-to-noise ratio (SNR) in fluorescence microscopy.
• Sources include spontaneous sample fluorescence and out-of-focus light.
• Noise comprises Gaussian and Poisson components.

Purpose of the Study:

• To develop a novel algorithm for simultaneous background blur subtraction and denoising.
• To improve the quality of wide-field fluorescence images.

Main Methods:

• Introduced the low-frequency background estimation and noise separation from high-frequency (LBNH-BNS) algorithm.
• Integrated low-frequency background features with noise separation.
• Disentangled noise from the desired signal.

Main Results:

• LBNH-BNS effectively eliminated noise and background blur.
• Demonstrated significant improvements in peak signal-to-noise ratio (PSNR) compared to existing methods.
• Achieved substantial visual enhancements in fluorescence images.

Conclusions:

• LBNH-BNS offers a powerful solution for background removal and denoising in fluorescence microscopy.
• The algorithm has high potential to advance wide-field fluorescence imaging performance and quality.