Use this catalog to identify pose estimators, video encoders, and lightweight detectors suitable for figure skating pipelines. Each entry summarises strengths, release cadence, and recommended integration patterns.
Systems that identify jumps, spins, and choreographic sequences from broadcast or practice footage, often powering judging review workflows.
Temporal segment network fine-tuned on keyframes to classify jump and spin elements from broadcast footage.
Two-stage pipeline combining 2D/3D pose estimation with a temporal segmentation network for element boundaries.
Pipeline that lifts AlphaPose skeletons into temporal graphs or probabilistic VIPE embeddings for figure skating element recognition.
Multimodal network blending RGB features with pose trajectories for robust element spotting under occlusion.
Depth-aware pose estimator trained on FS-Jump3D motion capture to recover 3D joint coordinates from monocular broadcast.
Real-time 2D keypoint detector tuned for low-latency figure skating edge and take-off detection.
Hierarchical transformer that models long-range pose dependencies for more stable spin and step-sequence tracking.
Research pipelines that break down rotations, airtime, and landing forces, blending sensors with computer vision for athlete safety and training insights.
Sensor-fusion pipeline combining monocular broadcast video with body-mounted IMUs to recover jump rotations and airtime.
Synchronized rink-side camera array estimating jump trajectories, landing zones, and clearance for training facilities.
Brigham Young University system that flags potential under-rotations by tracking blade yaw and flight angle in near real-time.
Automated blade edge classification pipeline using MediaPipe pose geometry and curated blade angle heuristics.
Transformer-based 3D pose estimation model with strong temporal reasoning, ideal for jump and spin kinematics.
Algorithms focused on differentiating flats, inside/outside edges, and intricate turn sequences that influence scoring and calls.
Automated blade edge classification pipeline using MediaPipe pose geometry and curated blade angle heuristics.
Sensor-fusion pipeline combining monocular broadcast video with body-mounted IMUs to recover jump rotations and airtime.
Foundational pose backbones and encoders that pair with the specialised systems above.
Masked autoencoder backbone for video that delivers strong general-purpose features for figure skating element discovery.