Research Analyzer
← Back ICRA 2026

Three-Dimensional Needle Tip Estimation from Multi-View X-Ray Images for Interventional Pain Procedures

Seunghui Han, Ayoung Hong

PDF

AI summary

Key figure (auto-extracted from paper)
A classical image processing pipeline accurately reconstructs the 3D position of a needle tip from multi-view 2D X-ray images, achieving sub-millimeter triangulation error.
needle tip estimation 3D reconstruction X-ray imaging image processing triangulation interventional robotics

Problem

Estimating the 3D position of a needle tip from 2D X-ray images is challenging due to lost directional information, low contrast, high noise, and the scarcity of discriminative features for learning-based models.

Approach

The method detects a circular marker on the robotic end-effector to locate the needle head, applies noise reduction and morphological filtering to segment the shaft, extracts the trajectory with A*, and reconstructs the 3D tip position via linear triangulation from multiple X-ray views.

Key results

  • 4-pixel average 2D tip localization error on synthetic DRR images
  • 0.89 mm mean 3D triangulation error across 20° to 60° view angles
  • Successful validation on real in vivo X-ray images from live pig experiments
  • GPU-accelerated pipeline enabling real-time computational efficiency

Why it matters

Provides a robust, model-free tracking solution for robotic interventional pain procedures, benefiting surgeons and medical device developers seeking reliable needle guidance.

Abstract

This study addresses the challenge of estimating the three-dimensional (3D) position of a needle tip from two- dimensional (2D) X-ray images. We propose a classical im- age processing–based framework for needle tip localization and 3D reconstruction. The method first detects a circular marker attached to the robotic end-effector that controls the needle insertion and identifies the needle head position within the marker. Preprocessing steps, including bilateral filtering, thresholding, and iterative morphological operations, are ap- plied to improve image quality and ensure the continuity of the needle shaft. A flood-fill algorithm is then used to segment the needle body, after which the needle trajectory is extracted using the A⋆algorithm. Finally, the 3D position of the needle tip is reconstructed by Triangulation from multiple X-ray images acquired at different viewing angles.

Index terms

Computer Vision for Medical Robotics

Related papers