Multifractal based image processing for estimating the complexity of COVID-19 dynamics

Area of Science:

• Medical Imaging
• Computational Biology
• Fractal Geometry

Background:

• COVID-19 poses a global health and economic threat, significantly impacting human lungs.
• Differentiating COVID-19 from other respiratory diseases using chest X-rays is challenging.
• Grayscale medical images are crucial for analyzing lung pathologies.

Purpose of the Study:

• To estimate and analyze the complexity of coronavirus transmission using Generalized Fractal Dimensions (GFD).
• To evaluate the robustness of COVID-19 detection against noise in chest X-ray images.
• To compare the complexity of COVID-19 with other respiratory diseases via GFD curves.

Main Methods:

• Calculation of multifractal dimensions for original, noisy, and denoised grayscale chest X-ray images.
• Graphical comparison of COVID-19 X-ray images with healthy and other disease images using GFD curves.
• Evaluation of image denoising performance using Mean Absolute Error (MAE) and Peak Signal-to-Noise Ratio (PSNR).

Main Results:

• Generalized Fractal Dimensions (GFD) effectively quantify the complexity of lung diseases in chest X-rays.
• The multifractal analysis demonstrates robustness even with noisy or denoised image data.
• GFD curves provide a graphical method to distinguish COVID-19 complexity from other respiratory conditions.

Conclusions:

• GFD is a valuable tool for analyzing and differentiating respiratory diseases, including COVID-19, from chest X-rays.
• The proposed method offers a robust approach for medical image analysis in pandemic scenarios.
• This research contributes to improved diagnostic capabilities for lung infections through advanced image processing techniques.