Dynamic elastic interpolation for 3D medical image reconstruction from serial cross sections

Area of Science:

• Computer Graphics
• Medical Imaging
• Geometric Modeling

Background:

• 3D object reconstruction from serial cross-sections is crucial in various fields.
• Existing methods often struggle with data loss and complex anatomical structures.
• Efficient interpolation techniques are needed to bridge the gap between discrete slices.

Purpose of the Study:

• To propose an interpolation method for generating intermediate contours between serial cross-sections.
• To develop a robust tool for 3D object reconstruction from medical imaging data.
• To address limitations of existing methods, particularly concerning data gaps and branching structures.

Main Methods:

• An interpolation algorithm is presented to estimate contours between a start and goal contour.
• The method assumes smooth transitions between successive slices, validated by Nyquist rate sampling.
• Integration with voxel-based display methods facilitates real-time visualization and reconstruction.

Main Results:

• The interpolation method successfully generates intermediate contours, enabling accurate 3D model creation.
• The approach effectively handles the challenge of branching structures in objects.
• The algorithm demonstrates low space complexity by not requiring persistent storage of intermediate contours.

Conclusions:

• The proposed interpolation method offers a powerful and efficient solution for 3D object reconstruction from serial cross-sections.
• Its ability to manage branching and its low memory footprint make it suitable for practical applications in medical imaging and computer graphics.
• The technique provides a realistic reconstruction when sampling adheres to the Nyquist rate.