Stripe artifact elimination based on nonsubsampled contourlet transform for light sheet fluorescence microscopy

Area of Science:

• Microscopy
• Image Processing
• Biotechnology

Background:

• Stripe artifacts commonly degrade image quality in light sheet fluorescence microscopy (LSFM).
• These artifacts arise from light absorption or scattering within the illumination path.
• Both unidirectional and multidirectional LSFM systems are affected, hindering detailed biological imaging.

Purpose of the Study:

• To develop and validate an effective method for removing stripe artifacts in LSFM images.
• To provide a versatile solution applicable to both unidirectional and multidirectional LSFM.

Main Methods:

• A novel multidirectional stripe remover (MDSR) method was developed.
• The MDSR method utilizes the nonsubsampled contourlet transform (NSCT) domain.
• A fast Fourier transform (FFT) filter is integrated within the NSCT domain to reduce stripe components and noise.

Main Results:

• MDSR successfully eliminated significant stripe artifacts in both unidirectional and multidirectional LSFM images.
• The method was validated on images acquired from custom-made and commercial LSFM systems.
• Comparative analysis showed superior noise removal performance of MDSR over existing methods like variational stationary noise remover and wavelet-FFT.

Conclusions:

• The MDSR method offers an effective solution for stripe artifact removal in LSFM.
• Its multiscale and multidirectional properties ensure broad applicability across different LSFM configurations.
• MDSR significantly improves LSFM image quality, facilitating more accurate biological analysis.