Cardiac 3D Motion Reconstruction Using Dual-Camera Defocused Speckle Imaging With Multi-Scale Amplification

Area of Science:

• Biomedical Engineering
• Cardiovascular Imaging
• Medical Physics

Background:

• Cardiovascular diseases are a leading global cause of mortality.
• Accurate assessment of cardiac motion is vital for diagnosis and rehabilitation.
• Existing single-camera methods struggle with motion coupling, limiting 3D cardiac behavior characterization.

Purpose of the Study:

• To develop a non-contact method for decoupling and reconstructing cardiac linear vibration (SCG) and rotational components (GCGx, GCGy).
• To overcome motion coupling issues inherent in single-camera imaging.
• To enhance the accuracy of characterizing complex 3D cardiac mechanical behavior.

Main Methods:

• Utilized dual-camera imaging with varying defocus levels to capture speckle motion signals.
• Integrated signals to decouple and reconstruct SCG, GCGx, and GCGy.
• Employed a sternum-mounted inertial sensor as a reference for validation in 42 subjects.

Main Results:

• Reconstructed 3D cardiac motion signals showed >87.471% waveform similarity to reference signals.
• Achieved high localization accuracy for key biomarkers (e.g., 99.998% for mitral valve closing) within an 8 ms error tolerance.
• Performance surpassed that of raw speckle motion signals from a single camera.

Conclusions:

• Dual-camera imaging with different defocus levels effectively reconstructs SCG, GCGx, and GCGy.
• This novel approach offers a promising method for accurate cardiac 3D motion capture.
• The technique can significantly improve cardiac function assessment for diagnosis and rehabilitation.