Monocular 3D Human Pose Estimation in the Wild Using Improved CNN Supervision

TLDR

The authors propose a CNN‑based method for estimating 3D human pose from single RGB images that overcomes the limited generalizability of models trained only on scarce 3D pose datasets. They train the CNN using transfer learning from 2D pose data, augment it with a new monocular dataset captured by a multi‑camera marker‑less motion‑capture system, and evaluate it on a newly introduced indoor‑outdoor benchmark that offers greater pose, appearance, and viewpoint diversity. The approach achieves state‑of‑the‑art results on standard benchmarks and outperforms existing datasets in in‑the‑wild scenarios, confirming that transfer learning combined with richer data is essential for general 3D pose estimation.

Abstract

We propose a CNN-based approach for 3D human body pose estimation from single RGB images that addresses the issue of limited generalizability of models trained solely on the starkly limited publicly available 3D pose data. Using only the existing 3D pose data and 2D pose data, we show state-of-the-art performance on established benchmarks through transfer of learned features, while also generalizing to in-the-wild scenes. We further introduce a new training set for human body pose estimation from monocular images of real humans that has the ground truth captured with a multi-camera marker-less motion capture system. It complements existing corpora with greater diversity in pose, human appearance, clothing, occlusion, and viewpoints, and enables an increased scope of augmentation. We also contribute a new benchmark that covers outdoor and indoor scenes, and demonstrate that our 3D pose dataset shows better in-the-wild performance than existing annotated data, which is further improved in conjunction with transfer learning from 2D pose data. All in all, we argue that the use of transfer learning of representations in tandem with algorithmic and data contributions is crucial for general 3D body pose estimation.