"In vivo" pose estimation of artificial knee implants using computer vision

Area of Science:

• Biomedical Engineering
• Computer Vision
• Orthopedic Surgery

Background:

• Premature wear of polyethylene components limits the lifespan of current artificial knee implants.
• Understanding in vivo motion and contact points is crucial for improving implant design.
• Existing methods may not fully capture the complex 6-degree-of-freedom motion of knee implants.

Purpose of the Study:

• To develop and validate a computer vision algorithm for estimating the 6-degree-of-freedom pose of artificial knee implants.
• To determine in vivo bone motion and joint contact positions during activities like deep knee bends.
• To provide data supporting the design of next-generation prosthetic knee implants with enhanced longevity.

Main Methods:

• Utilizing fluoroscopy videos of patients with knee implants performing deep knee bends.
• Developing a computer vision algorithm to estimate the 3-D pose (translation and rotation) of femoral and tibial components.
• Employing an image matching technique against a library of known orientations.
• Transforming coordinate systems to determine component contact positions.

Main Results:

• The algorithm accurately determines the full 6-degree-of-freedom pose of artificial knee components.
• Significant rotations, including internal/external rotations, are accurately captured.
• Contact positions are determined by transforming component poses into a common coordinate system.
• The process provides a representation of contact positions during normal knee motion.

Conclusions:

• The developed algorithm enables precise tracking of artificial knee implant motion using fluoroscopy.
• Accurate in vivo contact position determination is achievable, offering valuable insights for implant design.
• This approach can contribute to the development of more wear-resistant and longer-lasting prosthetic knee implants.