Automatic chronic degenerative diseases identification using enteric nervous system images.

Area of Science:

• Neuroscience
• Computational Biology
• Medical Imaging

Background:

• Chronic degenerative diseases impact Enteric Glial Cells (EGC).
• Accurate identification of affected EGC is crucial for disease diagnosis.
• Developing automated recognition methods for EGC health is needed.

Purpose of the Study:

• To develop and evaluate machine learning models for classifying healthy versus diseased EGC images.
• To compare the efficacy of handcrafted and deep learning features for EGC image analysis.
• To assess the diagnostic potential of pattern recognition techniques for EGC affected by chronic diseases.

Main Methods:

• Utilized pattern recognition and machine learning for EGC image classification.
• Extracted handcrafted features using texture descriptors like Local Binary Pattern (LBP).
• Employed deep learning techniques including Convolutional Neural Networks (CNNs) (AlexNet, VGG16) with and without transfer learning.
• Applied late fusion techniques to combine handcrafted and non-handcrafted features.

Main Results:

• Achieved high recognition rates: 89.30% for Rheumatoid Arthritis, 98.45% for Cancer, and 95.13% for Diabetes Mellitus.
• Demonstrated the effectiveness of combining both feature types (handcrafted and deep learning).
• Successfully distinguished between healthy and diseased EGC images across multiple disease categories.

Conclusions:

• The proposed machine learning approach accurately identifies EGC affected by chronic degenerative diseases.
• Combining diverse feature extraction methods enhances classification performance.
• This technique shows promise for aiding in the diagnosis of EGC-related chronic diseases.