Bidimensional Multiscale Fuzzy Entropy and Its Application to Pseudoxanthoma Elasticum

Area of Science:

• Medical Imaging Analysis
• Biomedical Engineering
• Computational Dermatology

Background:

• Pseudoxanthoma elasticum (PXE) is a rare genetic disorder affecting the skin, eyes, and cardiovascular system.
• Accurate and early detection of PXE is crucial for effective management and preventing complications.
• Current diagnostic methods for PXE can be subjective and may benefit from objective, quantitative tools.

Purpose of the Study:

• To introduce novel bidimensional entropy measures, FuzEn2D and its multiscale extension MSF2D, for quantifying image irregularity and complexity.
• To evaluate the performance of these measures on synthetic and real-world texture datasets.
• To assess the utility of MSF2D in the detection of PXE in dermoscopic images.

Main Methods:

• Developed bidimensional fuzzy entropy (FuzEn2D) and its multiscale extension (MSF2D).
• Evaluated FuzEn2D and MSF2D on synthetic images and texture datasets.
• Applied MSF2D to dermoscopic images of PXE and compared results with bidimensional multiscale sample entropy (MSE2D).

Main Results:

• FuzEn2D demonstrated superior reliability and stability compared to bidimensional sample entropy (SampEn2D) for quantifying image irregularity, especially in small images.
• MSF2D proved to be an effective measure of image complexity.
• MSF2D successfully differentiated between normal skin and PXE papules in dermoscopic images with significant statistical difference and large effect size.

Conclusions:

• The proposed FuzEn2D and MSF2D measures offer reliable quantification of image irregularity and complexity.
• MSF2D shows significant potential as a tool to aid clinicians in the objective diagnosis of PXE from dermoscopic images.
• This work contributes novel quantitative methods for medical image analysis, potentially improving diagnostic accuracy for conditions like PXE.