# Robust vertebra identification using simultaneous node and edge predicting Graph Neural Networks

Source: [Robust vertebra identification using simultaneous node and edge predicting Graph Neural Networks](<https://www.sota2.com/research/paper/68fa70391e209131cea49b25>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-01-04T04:37:03.149Z

Authors: Vincent B\\"urgin, Raphael Prevost, Marijn F. Stollenga

Published: 2023-07-27T09:10:27.000Z

## Abstract

Automatic vertebra localization and identification in CT scans is important for numerous clinical applications. Much progress has been made on this topic, but it mostly targets positional localization of vertebrae, ignoring their orientation. Additionally, most methods employ heuristics in their pipeline that can be sensitive in real clinical images which tend to contain abnormalities. We introduce a simple pipeline that employs a standard prediction with a U-Net, followed by a single graph neural network to associate and classify vertebrae with full orientation. To test our method, we introduce a new vertebra dataset that also contains pedicle detections that are associated with vertebra bodies, creating a more challenging landmark prediction, association and classification task. Our method is able to accurately associate the correct body and pedicle landmarks, ignore false positives and classify vertebrae in a simple, fully trainable pipeline avoiding application-specific heuristics. We show our method outperforms traditional approaches such as Hungarian Matching and Hidden Markov Models. We also show competitive performance on the standard VerSe challenge body identification task.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Edge Classification on 2118 spine dataset (val)](<https://www.sota2.com/research/sota/edge-classification-on-2118-spine-dataset-val.md>): Edge classification; 2118 spine dataset (val); Edge F1 Score: 99.31.

- [Edge Classification on 2118 spine dataset hard subset](<https://www.sota2.com/research/sota/edge-classification-on-2118-spine-dataset-hard-subset.md>): Edge classification; 2118 spine dataset hard subset; Edge F1 Score: 97.77.

- [Vertebra Identification on 2118 spine dataset full (val)](<https://www.sota2.com/research/sota/vertebra-identification-on-2118-spine-dataset-full-val.md>): Vertebra Identification; 2118 spine dataset full (val); Identification Rate: 0.9719.

- [Vertebra Identification on 2118 spine dataset hard](<https://www.sota2.com/research/sota/vertebra-identification-on-2118-spine-dataset-hard.md>): Vertebra Identification; 2118 spine dataset hard; Identification Rate: 89.88.

- [Vertebra identification on VerSe 2019 (val)](<https://www.sota2.com/research/sota/vertebra-identification-on-verse-2019-val.md>): Vertebra Identification; VerSe 2019 (val); Identification Rate: 93.26.

- [Vertebra Identification on VerSe 2019 (test)](<https://www.sota2.com/research/sota/vertebra-identification-on-verse-2019-test.md>): Vertebra Identification; VerSe 2019 (test); Identification Rate: 93.02.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2308.02509>)

[PDF](<https://arxiv.org/pdf/2308.02509.pdf>)
