A Comparative Analysis of Optimization Algorithms for Gastrointestinal Abnormalities Recognition and Classification...

Area of Science:

• Medical Imaging
• Artificial Intelligence in Medicine
• Gastroenterology

Background:

• Gastrointestinal disorders are a major cause of mortality, with traditional diagnostic methods being time-consuming.
• Accurate and early diagnosis is crucial for effective treatment and reducing mortality rates.

Purpose of the Study:

• To propose a hybrid method for accurate diagnosis of gastrointestinal tract abnormalities.
• To promote early treatment and reduce mortality associated with these disorders.

Main Methods:

• The study employed dataset augmentation, preprocessing, and feature engineering (extraction, fusion, optimization).
• Image enhancement used hybrid contrast stretching; deep features were extracted via transfer learning (ResNet18, XcepNet23) and ensembled with texture features.
• Feature optimization utilized Binary Dragonfly Algorithm (BDA), Moth-Flame Optimization (MFO), and Particle Swarm Optimization (PSO).

Main Results:

• The proposed method achieved superior accuracy on both the Hybrid (5 classes) and Kvasir-V1 (8 classes) datasets.
• Quadratic Support Vector Machine (Q_SVM) achieved 100% accuracy on the Hybrid dataset and 99.24% on the Kvasir-V1 dataset.
• The results demonstrated the effectiveness of the hybrid approach compared to recent methods.

Conclusions:

• The developed hybrid method offers a promising, highly accurate solution for diagnosing gastrointestinal abnormalities.
• This approach can significantly aid in early detection and treatment, potentially reducing mortality.
• The integration of deep learning, texture features, and advanced optimization algorithms proved effective.