Automatic Detection and Classification of Knee Osteoarthritis Using Hu's Invariant Moments

Area of Science:

• Medical Imaging
• Image Processing
• Biomedical Engineering

Background:

• Geometric distortions in X-ray images complicate information extraction.
• Accurate analysis of cartilage region is crucial for diagnosing knee osteoarthritis.

Purpose of the Study:

• To utilize Hu's invariant moments for early detection and grading of knee osteoarthritis.
• To analyze geometric transformations in the knee cartilage region of X-ray images.

Main Methods:

• Computed seven invariant moments for rotated knee X-ray images.
• Applied Hu's invariant moments to understand geometric transformations of the cartilage.
• Validated findings with orthopedic surgeons and rheumatologists.

Main Results:

• Demonstrated competitive and promising results in knee osteoarthritis detection and grading.
• Hu's moments effectively captured geometric transformations in distorted images.
• Expert validation confirmed the clinical relevance of the findings.

Conclusions:

• Hu's invariant moments are effective for analyzing geometrically transformed medical images.
• This technique shows potential for early and accurate knee osteoarthritis diagnosis.
• The method provides a valuable tool for radiologists and clinicians.