Lumbar spine disorder detection using segmentation and IBN-MGNet classification with enhanced feature extraction.

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Spine Disorders

Background:

• Back pain is a prevalent global health issue, significantly contributed to by lumbar spine disorders (LSDs).
• Accurate detection of LSDs is crucial for effective patient management and treatment.

Purpose of the Study:

• To develop an advanced deep learning framework for enhanced detection of lumbar spine disorders (LSDs).
• To improve the precision and reliability of LSD detection systems.

Main Methods:

• A novel pipeline involving bilateral filtering for preprocessing.
• Image segmentation using Proposed Residual Block with Patch-based RESU-NET (PRB-PRESU-NET).
• Feature extraction combining Gaussian Filter with Scharr Operator-based Gradient Local Ternary Patterns (GSO-GLTP), MT, and deep features.
• Classification utilizing the Improved Bottle Neck-based Modified GhostNet (IBN-MGNet).

Main Results:

• The proposed framework achieved a high classification accuracy of 0.9312 for LSD detection.
• The Scharr operator demonstrated superior edge detection capabilities compared to Sobel filters.
• The developed system outperformed existing methods in precision and reliability.

Conclusions:

• The integrated approach of refined segmentation, feature extraction, and classification offers a reliable system for detecting lumbar spine disorders.
• This deep learning framework represents a significant advancement in the diagnostic tools for LSDs.