An Orthopedic Auxiliary Diagnosis System Based on Image Recognition Technology

Area of Science:

• Orthopedics
• Medical Image Processing
• Computer-Aided Diagnosis

Background:

• Orthopedic diseases present complex diagnostic challenges, leading to potential misdiagnosis.
• Computer-aided diagnosis (CAD) systems utilizing medical image processing can enhance diagnostic accuracy.
• Spatial inconsistency in medical images complicates accurate diagnosis.

Purpose of the Study:

• To develop an image registration method for improving spatial consistency in orthopedic medical images.
• To create a computer-aided diagnosis system for orthopedic diseases.
• To assist orthopedic doctors in diagnosing conditions like gout and cartilage defects.

Main Methods:

• Proposed an image registration method based on volume feature point selection and Powell algorithm optimization.
• Utilized golden section method for linear search and Powell algorithm for optimization to find optimal spatial transformation parameters.
• Developed a five-module computer-aided diagnosis system for orthopedic diseases.

Main Results:

• The proposed method ensures two-dimensional spatial consistency by maximizing normalized mutual information between images.
• The developed system demonstrated effective performance in cartilage segmentation, urate agglomeration segmentation, and artifact removal.
• The system achieved successful two-dimensional and three-dimensional image registration and visualization.

Conclusions:

• The developed computer-aided diagnosis system, incorporating an advanced image registration technique, shows significant potential for clinical application.
• The system provides a robust basis for assisting physicians in diagnosing and evaluating orthopedic conditions such as gout and cartilage defects.
• This approach enhances diagnostic accuracy and efficiency, potentially reducing patient suffering and treatment duration.