Related Experiment Video
Updated: May 7, 2026
07:13
Author Spotlight: An Efficient and Robust Software for Automated Fusion of Multiple Preclinical Imaging Modalities
Published on: October 27, 2023
2.0K
Automatic 3D ultrasound calibration for image guided therapy using intramodality image registration
Jeffrey Schlosser1, Can Kirmizibayrak, Vijay Shamdasani
1Department of Mechanical Engineering, Stanford University, Stanford, CA 94305, USA. Department of Bioengineering, Stanford University, Stanford, CA 94305, USA.
Physics in Medicine and Biology
|October 9, 2013
Summary
This study introduces an automated method for 3D ultrasound calibration, improving accuracy for image-guided therapies. The new technique enhances spatial accuracy by registering ultrasound volumes to anatomy models, crucial for real-time interventions.
Area of Science:
- Medical Imaging
- Ultrasound Technology
- Image-Guided Therapy
Background:
- Real-time ultrasound (US) guided therapies require precise anatomical alignment with pre-acquired models.
- Motion-induced anatomical changes necessitate robust spatial calibration for accurate US image integration.
- Current hand-eye calibration methods for 3D ultrasound have limitations in accuracy and automation.
Purpose of the Study:
- To develop and validate a fully automated method for calibrating 3D ultrasound volumes.
- To improve the accuracy and reproducibility of spatial calibration for image-guided interventions.
- To establish a novel validation technique for ultrasound calibration using physical phantoms.
Main Methods:
- A novel intramodality image registration method based on the hand-eye calibration technique was developed.
- Automation was achieved through sensor displacement data rejection, overlapping region registration, and continuous self-consistency error evaluation.
- A new validation method using physical phantom displacements within ultrasound images was introduced.
Main Results:
- The automated calibration method achieved high accuracy, with volumetric image alignment yielding <1.5 mm root mean square error for validation.
- Normalized mutual information and localized cross-correlation were identified as optimal registration algorithms for 3D US calibration.
- Validation results showed significant differences based on phantom type (p = 0.003), while calibration did not (p = 0.795).
Conclusions:
- The presented automated 3D ultrasound calibration method significantly improves accuracy and reproducibility compared to previous techniques.
- The novel phantom-based validation method provides a reliable means to assess calibration performance.
- This advancement is critical for enhancing the precision and safety of real-time ultrasound-guided medical procedures.
