Reslicing axially sampled 3D shapes using elliptic Fourier descriptors

Area of Science:

• Medical image analysis
• Computational geometry
• Biomedical engineering

Background:

• Volumetric medical imagery provides rich 3D anatomical data.
• Reslicing and correspondence are crucial for analyzing and modeling organ shapes.
• Existing methods can be computationally intensive.

Purpose of the Study:

• To develop an efficient method for reslicing 3D organ shapes.
• To enable correspondence of organ shapes from medical imagery without explicit point mapping.
• To facilitate the construction of 3D shape models for improved segmentation.

Main Methods:

• Interpolation of elliptic Fourier descriptor coefficients between parallel contours.
• Direct generation of new slices at arbitrary axial locations.
• Avoidance of explicit point correspondence or 3D surface reconstruction.

Main Results:

• The proposed method successfully reslices synthetic and real prostate contours.
• Performance is comparable to variational implicit surface methods.
• Achieves significantly lower computational cost compared to existing techniques.

Conclusions:

• The novel interpolation method offers an efficient alternative for 3D shape reslicing.
• Enables accurate 3D organ shape modeling and segmentation from medical data.
• Reduces computational burden in medical image analysis workflows.