Human Pose Estimation in Trampoline Gymnastics: Improving Performance Using a New Synthetic Dataset

TL;DR

This paper improves human pose estimation in trampoline gymnastics by fine-tuning a model on a synthetic dataset of extreme poses, leading to state-of-the-art 2D accuracy and significant 3D error reduction.

Contribution

The creation of a synthetic trampoline pose dataset and its use to enhance pose estimation accuracy in extreme human poses.

Findings

Achieved state-of-the-art 2D pose estimation on trampoline data.

Reduced 3D MPJPE by 12.5 mm, a 19.6% improvement over baseline.

Demonstrated the effectiveness of synthetic data for extreme pose scenarios.

Abstract

Trampoline gymnastics involves extreme human poses and uncommon viewpoints, on which state-of-the art pose estimation models tend to under-perform. We demonstrate that this problem can be addressed by fine-tuning a pose estimation model on a dataset of synthetic trampoline poses (STP). STP is generated from motion capture recordings of trampoline routines. We develop a pipeline to fit noisy motion capture data to a parametric human model, then generate multiview realistic images. We use this data to fine-tune a ViTPose model, and test it on real multi-view trampoline images. The resulting model exhibits accuracy improvements in 2D which translates to improved 3D triangulation. In 2D, we obtain state-of-the-art results on such challenging data, bridging the performance gap between common and extreme poses. In 3D, we reduce the MPJPE by 12.5 mm with our best model, which represents an improvement of 19.6% compared to the pretrained ViTPose model.