What is spine pose estimation?

Spine pose estimation refers to predicting spine-related landmarks, curves, segments, or vertebral structures from images or video. Depending on the sensing setup, the target may be 2D spinal keypoints in RGB images, vertebra segmentation in CT or MRI, or posture and curvature estimates from radiographs or surface topography.

What is the difference between spine pose estimation and spine motion analysis?

Spine pose estimation usually refers to recovering structure or posture from a single image or frame, while spine motion analysis focuses on how the spine moves over time. Motion analysis therefore needs temporal consistency, 3D geometry, and ideally biomechanics-aware outputs such as vertebral rotations, curvature changes, or anatomically plausible segment motion.

Which sensing modalities are used for AI in spine biomechanics?

The field spans standing radiographs, biplanar X-ray, CT, MRI, structured light and rasterstereography, RGB and multiview video, and RGB-D or point-cloud sensing for surgical or navigation settings. Each modality exposes different tradeoffs between anatomy, motion, radiation, repeatability, and deployment cost.

Where can I find open datasets and benchmarks for spine motion and spine imaging AI?

Open resources exist, but they are unevenly distributed across the field. VerSe is a strong open CT benchmark, SPIDER represents lumbar MRI segmentation, AASCE covers scoliosis landmarking on radiographs, and my own SpinePose and SIMSPINE pages cover open RGB and biomechanics-aware motion resources.

Spine Vision Benchmarks, Datasets, and Systems

This matrix is intentionally cross-modal. It mixes public benchmarks, commercial clinical systems, and open research releases so visitors can compare what is measurable, what is reproducible, and what is still missing.

Representative resources

AASCE / SpineWeb

Challenge benchmark · Single-view radiographs · Established but narrow

Scoliosis landmark detection and curvature estimation. Useful for coronal alignment and deformity; not dynamic.

Evidence and evaluation: Annotated vertebral landmarks and alignment targets on AP X-rays. Challenge-style ranking centered on landmark and curvature accuracy.
Access: Public challenge resources via SpineWeb and Grand Challenge. Benchmark-oriented ecosystem, but no dominant reusable inference stack.

EOS / sterEOS

Clinical system · Biplanar low-dose X-ray · Deployed clinically

Weight-bearing 2D/3D reconstruction and alignment analysis. Very high for structural alignment under load.

Evidence and evaluation: Expert-guided reconstruction from paired standing radiographs. Clinical workflow rather than open leaderboard.
Access: Commercial platform. No public inference stack.

VerSe

Public challenge + dataset · CT · Strongest open anatomy benchmark

Vertebra localization, segmentation, and identification. Indirect; excellent for anatomy, weak for motion.

Evidence and evaluation: Volumetric vertebra masks and labels across heterogeneous CT scans. Standardized Grand Challenge tasks with shared metrics and reports.
Access: Public benchmark. Open challenge reports and baseline implementations are available.

SPIDER

Public challenge + dataset · Lumbar MRI · Growing open MRI benchmark

Lumbar spine segmentation and clinically relevant structure delineation. Indirect; more anatomy and pathology than motion.

Evidence and evaluation: Pixelwise annotations for lumbar structures on MRI. Continuous challenge setup with hidden test evaluation.
Access: Public challenge infrastructure. Shared baselines and evaluation framework.

DIERS formetric 4D

Clinical system · Structured light / rasterstereography · Useful in practice, weakly benchmarked

Back-surface posture analysis and radiation-free monitoring. Moderate; useful for posture and follow-up, indirect for vertebral anatomy.

Evidence and evaluation: Surface-shape capture and anatomical proxies rather than direct vertebral labels. Clinical repeatability and correlation studies.
Access: Commercial system. Public code and shared ML benchmarks are sparse.

SpineDepth

Research dataset / method · RGB-D and point clouds · Emerging lane

Dense 3D reconstruction of exposed spine anatomy for surgical registration. High anatomical fidelity in procedure-specific settings.

Evidence and evaluation: Paired RGB-D depth completion and anatomy-aware geometry targets. Research-task metrics for depth and geometry quality.
Access: Research release. Useful reference point, but not yet a field-standard benchmark stack.

SpinePose

Paper + dataset · Monocular RGB · Available now

2D spine-centric landmark estimation in natural images. Moderate; stronger than standard pose skeletons, still 2D.

Evidence and evaluation: Manual 2D spine landmarks with biomechanics-aware evaluation. Public paper benchmarks on spine-centric 2D tasks.
Access: Paper page, dataset, and inference entry point. Inference library available.

SIMSPINE

Paper + dataset + benchmark · Multiview RGB with simulation-derived supervision · Available now

3D spinal keypoints, rotational kinematics, and benchmark baselines. High; anatomy-aware motion is the core point.

Evidence and evaluation: Simulated vertebral markers, multi-view detections, and inverse-kinematics labels. Reference baselines across 2D, multiview 3D, monocular 3D, and biomechanics-aware checks.
Access: Paper page plus dataset on Hugging Face. Benchmark assets released; inference support is the next tooling step.

Benchmarks and Systems

Representative resources

AASCE / SpineWeb

EOS / sterEOS

VerSe

SPIDER

DIERS formetric 4D

SpineDepth

SpinePose

SIMSPINE

Further reading

Field review

Radiographs

Biplanar radiography

CT

MRI

Surface topography

RGB-D / surgery

RGB biomechanics