Elimination of Image Saturation Effects on Multifractal Statistics Using the 2D WTMM Method

Area of Science:

• Medical Imaging
• Computational Analysis
• Fractal Geometry

Background:

• Image saturation is a common artifact that hinders computational analysis of medical images.
• Multifractal analyses, crucial for characterizing complex surfaces, are typically invalidated by image saturation.
• Excluding saturated regions reduces the statistical power of clinical analyses.

Purpose of the Study:

• To develop a robust multifractal analysis method capable of handling image saturation in medical imaging.
• To adapt the 2D Wavelet Transform Modulus Maxima (WTMM) method for artifact-resilient analysis.
• To maintain the statistical power of clinical analyses despite image saturation.

Main Methods:

• Adapted the 2D Wavelet Transform Modulus Maxima (WTMM) multifractal analysis.
• Developed a strategy to partition images based on localized responses to saturated regions.
• Excluded contributions from saturated regions in partition function calculations.

Main Results:

• Successfully estimated multifractal statistics in the presence of image saturation.
• The adapted method remains accurate with image saturation levels up to 20% of total pixels.
• Preserved the integrity of multifractal analysis for rough surfaces with artifacts.

Conclusions:

• The adapted WTMM method effectively overcomes limitations imposed by image saturation artifacts.
• This approach enhances the reliability of computational and clinical analyses of medical images.
• Enables more robust multifractal characterization of medical images with saturation.