# Attaining human-level performance with atlas location autocontext for anatomical landmark detection in 3D CT data

arXiv: https://arxiv.org/abs/1805.08687 · HTML: https://mlsanity.com/paper/1805.08687/

## Key facts

- Cited by 8 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 14 May 2018.

## Built on by (most cited first, 8 of 8)

- [CLIP in Medical Imaging: A Survey](https://mlsanity.com/paper/2312.07353/) (2023)
- [SAM++: Enhancing Anatomic Matching using Semantic Information and Structural Inference](https://mlsanity.com/paper/2306.13988/) (2023)
- [Feature Aggregation and Refinement Network for 2D AnatomicalLandmark Detection](https://mlsanity.com/paper/2111.00659/) (2021)
- [Deep Learning Fetal Ultrasound Video Model Match Human Observers in Biometric Measurements](https://mlsanity.com/paper/2205.13835/) (2022)
- [Matching in the Wild: Learning Anatomical Embeddings for Multi-Modality Images](https://mlsanity.com/paper/2307.03535/) (2023)
- [Anatomical Landmarks Localization for 3D Foot Point Clouds](https://mlsanity.com/paper/2110.00937/) (2021)
- [HYATT-Net is Grand: A Hybrid Attention Network for Performant Anatomical Landmark Detection](https://mlsanity.com/paper/2412.06499/) (2024)
- [Rule-based Key-Point Extraction for MR-Guided Biomechanical Digital Twins of the Spine](https://mlsanity.com/paper/2508.14708/) (2025)

## Abstract (arXiv)

We present an efficient neural network method for locating anatomical landmarks in 3D medical CT scans, using atlas location autocontext in order to learn long-range spatial context. Location predictions are made by regression to Gaussian heatmaps, one heatmap per landmark. This system allows patchwise application of a shallow network, thus enabling multiple volumetric heatmaps to be predicted concurrently without prohibitive GPU memory requirements. Further, the system allows inter-landmark spatial relationships to be exploited using a simple overdetermined affine mapping that is robust to detection failures and occlusion or partial views. Evaluation is performed for 22 landmarks defined on a range of structures in head CT scans. Models are trained and validated on 201 scans. Over the final test set of 20 scans which was independently annotated by 2 human annotators, the neural network reaches an accuracy which matches the annotator variability, with similar human and machine patterns of variability across landmark classes.