Reconstruction accuracy enhancement of fiber shape sensing by Kabsch-algorithm-based pose correction

Area of Science:

• Medical instrumentation
• Robotics and intelligent systems
• Computational geometry

Background:

• Fiber-optic shape sensing is crucial for minimally invasive procedure navigation.
• Existing shape reconstruction algorithms (FS, RMF, HEM, TM) suffer from error accumulation, limiting distal tip accuracy.
• Noise sources like strain and angular deviations further degrade reconstruction fidelity.

Purpose of the Study:

• To theoretically compare the reconstruction accuracy and noise robustness of four common fiber-optic shape sensing algorithms.
• To propose and evaluate a novel Kabsch-algorithm-based pose correction method to enhance shape reconstruction accuracy.
• To assess the improved accuracy for complex paths relevant to clinical applications.

Main Methods:

• Theoretical comparison of Frenet-Serret frame (FS), rotation minimizing frame (RMF), helical extension method (HEM), and transformation matrix (TM) algorithms.
• Analysis of algorithm performance under noise-free and noisy conditions (strain and angular noise).
• Implementation and evaluation of Kabsch-algorithm-based pose correction to refine principal axis direction.

Main Results:

• In noise-free conditions, the TM algorithm demonstrated the lowest reconstruction error.
• TM algorithm accuracy degraded under strain and angular noise, particularly at high spatial resolution.
• Kabsch-algorithm-based pose correction significantly improved TM algorithm accuracy, exceeding 35.7% under coupled noise and up to 42.7% under angular noise.
• For a brachial-artery-to-heart path, pose correction reduced mean shape reconstruction error by 28% (from 1.25mm to 0.90mm).

Conclusions:

• The transformation matrix (TM) algorithm shows promise for clinical tip localization but requires noise mitigation.
• Kabsch-algorithm-based pose correction offers a simple yet effective method to enhance the accuracy of fiber-optic shape sensing, especially for complex paths.
• This study provides guidance for selecting noise-robust reconstruction methods and highlights the potential of pose correction for improving navigation in minimally invasive procedures.