Multiframe denoising of high-speed optical coherence tomography data using interframe and intraframe priors

Area of Science:

• Biomedical Imaging
• Optical Engineering
• Medical Diagnostics

Background:

• Optical coherence tomography (OCT) is vital for subsurface microstructure imaging.
• High-speed OCT suffers from significant speckle noise, degrading image quality.
• Existing denoising methods struggle to balance noise reduction and detail preservation.

Purpose of the Study:

• To develop an advanced algorithmic method for denoising OCT volumes.
• To improve the imaging quality of high-speed OCT by reducing speckle noise.
• To preserve crucial microstructural details often lost during noise reduction.

Main Methods:

• A multiframe algorithmic approach was proposed for OCT volume denoising.
• A mathematical optimization model was formulated incorporating low-rank and sparse constraints.
• The model considered logarithmic image formation and nonuniform noise variance.
• A convex optimization algorithm using the augmented Lagrangian method was employed.

Main Results:

• The proposed method demonstrated superior speckle noise suppression compared to existing techniques.
• Crucial subsurface microstructural details were effectively preserved.
• The algorithm successfully addressed limitations of high-speed OCT imaging.

Conclusions:

• The multiframe algorithmic method offers a significant advancement in OCT image denoising.
• This technique enhances the diagnostic utility of high-speed OCT for biological tissues.
• The approach provides a robust solution for noise reduction while maintaining image fidelity.