Local mean suppression filter for effective background identification in fluorescence images.

Area of Science:

• Microscopy and Image Analysis
• Computational Biology
• Biophotonics

Background:

• Accurate background identification is crucial for analyzing fluorescence microscopy images, especially those with dense or low-contrast features.
• Existing methods for background subtraction and foreground identification can be complex or computationally intensive.

Purpose of the Study:

• To introduce a user-friendly, nonlinear filter for reliable background identification in challenging fluorescence microscopy images.
• To provide a robust and efficient alternative to current state-of-the-art image processing techniques.

Main Methods:

• A pixel-wise nonlinear filtering approach comparing pixel intensity to local neighborhood mean intensity.
• Generation of multiple pixel labels by varying neighborhood sizes and accumulating results for final classification.
• Implementation in Python 3 for accessibility and ease of use.

Main Results:

• The developed filter demonstrates performance comparable to or exceeding state-of-the-art image processing, machine learning, and deep learning methods.
• Successful application in three distinct use cases, including multiplexed fluorescence imaging and image segmentation pre-processing.

Conclusions:

• The nonlinear filter offers an effective and accessible solution for background identification in fluorescence microscopy.
• Its versatility makes it suitable for various applications, including advanced imaging and segmentation tasks.