Wavelet-based energy features for glaucomatous image classification.

Area of Science:

• Medical Imaging
• Computer Vision
• Signal Processing

Background:

• Glaucoma classification relies on accurate image texture analysis.
• Wavelet transforms are effective for extracting texture features.

Purpose of the Study:

• To investigate the potential of various wavelet filters (daubechies, symlets, biorthogonal) for glaucoma classification.
• To develop a novel technique for extracting energy signatures from wavelet subbands.
• To evaluate feature ranking and selection strategies for optimal classification.

Main Methods:

• Utilized 2-D discrete wavelet transform (DWT) for energy signature extraction.
• Applied daubechies (db3), symlets (sym3), and biorthogonal (bio3.3, bio3.5, bio3.7) wavelet filters.
• Employed feature ranking and selection strategies.
• Evaluated classification performance using Support Vector Machine (SVM), Sequential Minimal Optimization (SMO), Random Forest (RF), and Naïve Bayes (NB).

Main Results:

• Achieved approximately 93% accuracy in glaucoma classification using tenfold cross-validation.
• Demonstrated the effectiveness of the proposed energy signature extraction and feature selection methods.
• Identified specific wavelet features with high discriminatory power for glaucoma.

Conclusions:

• Wavelet-derived texture features are highly effective for glaucoma classification.
• The proposed novel technique for energy signature extraction and feature selection enhances classification accuracy.
• This approach shows significant potential for developing automated glaucoma diagnostic tools.