Related Experiment Video
Updated: Mar 2, 2026
12:54
Simultaneous Brightfield, Fluorescence, and Optical Coherence Tomographic Imaging of Contracting Cardiac Trabeculae Ex Vivo
Published on: October 2, 2021
3.7K
Scale-space for empty catheter segmentation in PCI fluoroscopic images
Ketan Bacchuwar1,2, Jean Cousty3, Régis Vaillant4
1GE-Healthcare, 78530, Buc, France. ketan.bacchuwar@esiee.fr.
Summary
We developed a novel method using structural scale-space to automatically segment empty guiding catheters in X-ray fluoroscopy images. This technique improves Percutaneous Coronary Intervention (PCI) procedure modeling by accurately identifying catheter positions.
Area of Science:
- Medical Imaging
- Computer Vision
- Interventional Cardiology
Background:
- Guiding catheter segmentation is crucial for Percutaneous Coronary Intervention (PCI) procedure modeling.
- Empty guiding catheters present challenges due to low contrast and disconnected edges in fluoroscopic images.
Purpose of the Study:
- To present a novel method for segmenting empty guiding catheters in X-ray fluoroscopic images.
- To facilitate accurate PCI procedure modeling through reliable catheter segmentation.
Main Methods:
- Utilized level-set scale-space and min-tree to extract curve blobs.
- Developed a novel structural scale-space hierarchy based on curve blobs.
- Identified the deep connected component maximizing empty catheter likelihood for segmentation.
Main Results:
- Evaluated on 1250 fluoroscopic images from 6 patients.
- Achieved mean precision of 80.48% and recall of 63.04%.
- Demonstrated very good qualitative and quantitative segmentation performance.
Conclusions:
- A novel structural scale-space effectively segments sparse empty catheters in X-ray images.
- Fully-automatic segmentation is a vital preliminary step for PCI modeling.
- Aids in tracking interventional tools during PCI procedures.
