A single algorithm can be used to automatically label spinal vertebrae and intervertebral discs on any type of 3D image, according to research published in the International Journal of Computer Assisted Radiology and Surgery.
Here are four things to know:
1. The researchers created a cross-modality labeling solution for clinical practice. They used entropy-optimized texture models to fix seed points for labeling and used the sacral region as the starting point for labeling.
2. The study authors evaluated the algorithm on 161 publicly available spinal scans acquired on various scanner types.
3. The algorithm accurately detected the seed points in the sacrum 93.6 percent of the time.
4. The algorithm does not need to be retrained for different types of imaging equipment.