Colon flattening using heat diffusion Riemannian metric

Area of Science:

• Medical Imaging
• Computational Geometry
• Computer-Aided Surgery

Background:

• Virtual colonoscopy requires accurate colon surface representation for effective navigation and polyp detection.
• Existing colon flattening methods often struggle with topological noise, necessitating complex denoising steps.
• Shape preservation is crucial for maintaining the integrity of anatomical structures and pathologies.

Purpose of the Study:

• To introduce an efficient and robust colon flattening algorithm.
• To eliminate the need for topological denoising in colon flattening.
• To improve the usability of virtual colonoscopy and colon registration pipelines.

Main Methods:

• Replaced the Euclidean metric with a heat diffusion metric, inherently robust to topological noise.
• Solved a Laplacian equation using the new metric to compute the flattening.
• Integrated the solution to obtain the final flattened colon surface.

Main Results:

• The algorithm efficiently flattens the colon surface while preserving its shape.
• Polyp shapes are accurately maintained in the flattened representation.
• The method demonstrated robustness against topological noise without prior denoising.

Conclusions:

• The proposed heat diffusion-based colon flattening is an efficient, shape-preserving, and noise-robust alternative.
• This technique enhances navigation and inspection in virtual colonoscopy.
• It also improves the robustness of existing colon registration pipelines.