Related Experiment Video
Updated: Jul 16, 2026
04:48
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
Published on: November 30, 2022
Real-time image mosaicing for medical applications
Kevin E Loewke1, David B Camarillo, Christopher A Jobst
1Department of Mechanical Engineering, Stanford University, CA, USA.
Studies in Health Technology and Informatics
|March 23, 2007
Summary
A new robotically-assisted system creates medical image mosaics in real-time. Using 5-degree-of-freedom sensing and hand-eye calibration, it accurately captures camera motion for faster, error-free imaging.
Area of Science:
- Medical imaging
- Robotics
- Computer vision
Background:
- Medical image mosaicing is crucial for comprehensive visualization.
- Real-time processing and accurate motion tracking remain challenges.
- Near-field imaging involves complex camera movements (translations, pan, tilt).
Purpose of the Study:
- To develop a robotically-assisted image mosaicing system for medical applications.
- To achieve real-time processing through fast initial image alignment.
- To accurately measure and compensate for complex camera motions.
Main Methods:
- Implementation of a robotically-assisted system for image mosaicing.
- Utilizing robotic position sensing for rapid image alignment.
- Employing 5-degree-of-freedom (5-d.o.f.) sensing to measure translations, pan, and tilt.
- Incorporating hand-eye calibration to correct for sensor offset.
Main Results:
- The system achieves real-time processing speeds.
- Elimination of cumulative errors in mosaicing.
- Successful handling of arbitrary camera motions, including translations.
- Visually satisfactory medical image mosaics were produced on a dental model.
Conclusions:
- Robotically-assisted mosaicing with 5-d.o.f. sensing offers a robust solution for medical imaging.
- The developed system enhances speed and accuracy while minimizing errors.
- The approach is adaptable for various medical imaging modalities beyond dental models.
