Classification of foetal heart rate sequences based on fractal features

Area of Science:

• Cardiology
• Biomedical Engineering
• Signal Processing

Background:

• Visual inspection of fetal heart rate (FHR) sequences is crucial for assessing fetal well-being.
• Traditional classification methods struggle to differentiate certain FHR sequence patterns.
• FHR signals exhibit scale-invariant properties characteristic of fractal patterns.

Purpose of the Study:

• To apply fractal features for classifying physiologically relevant FHR sequence patterns.
• To overcome limitations of traditional classification schemes in FHR analysis.
• To explore novel methods for objective FHR assessment.

Main Methods:

• Characterizing FHR sequences using fractal features through two distinct approaches.
• Modeling FHR sequences as temporal fractals.
• Employing chaos-theory techniques to model the attractor based on FHR sequences.
• Designing a Bayesian classification scheme utilizing the derived fractal features.

Main Results:

• Satisfactory classification results were achieved for three distinct FHR classes.
• Demonstrated the efficacy of fractal-based methodologies in FHR pattern discrimination.
• Highlighted the potential of fractal analysis in enhancing fetal well-being evaluation.

Conclusions:

• Fractal features provide a robust method for classifying FHR sequences.
• This approach offers significant advantages over traditional classification techniques.
• Fractal analysis represents an important advancement in fetal well-being assessment.