Vertebra Segmentation Using Active Shape Model
Author Information
Author(s): Benjelloun Mohammed, Mahmoudi Saïd, Lecron Fabian
Primary Institution: Faculty of Engineering, University of Mons
Hypothesis
Can an Active Shape Model approach effectively segment vertebrae in X-ray images?
Conclusion
The proposed method allows for fast and efficient extraction of vertebra contours from X-ray images.
Supporting Evidence
- The method was tested on a large dataset and showed promising results.
- Using 20 landmarks per vertebra improved segmentation quality.
- A sample size of 75 images was found to be a good compromise for effective model training.
- The Active Shape Model approach allows for effective vertebra edge detection.
Takeaway
This study shows a way to help computers find and outline vertebrae in X-ray pictures, making it easier for doctors to see and understand them.
Methodology
The study used an Active Shape Model for vertebra segmentation, involving a training phase to create a statistical model and a semiautomatic initialization process for better accuracy.
Limitations
The method's effectiveness is highly dependent on the quality of the initialization and the representativeness of the training sample.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website