PostoMETRO: Pose Token Enhanced Mesh Transformer for Robust 3D Human Mesh Recovery

TL;DR

PostoMETRO introduces a novel transformer-based approach that uses a pose tokenizer to enhance 3D human mesh recovery, especially under occlusion, by effectively integrating 2D pose information with image features.

Contribution

The paper proposes a pose tokenization method within transformers to improve 3D human mesh recovery under occlusion, leveraging rich 2D pose annotations more effectively.

Findings

Outperforms existing methods on standard and occlusion-specific benchmarks.

Produces more accurate 3D meshes under occlusion scenarios.

Demonstrates the effectiveness of pose tokens in 3D reconstruction.

Abstract

With the recent advancements in single-image-based human mesh recovery, there is a growing interest in enhancing its performance in certain extreme scenarios, such as occlusion, while maintaining overall model accuracy. Although obtaining accurately annotated 3D human poses under occlusion is challenging, there is still a wealth of rich and precise 2D pose annotations that can be leveraged. However, existing works mostly focus on directly leveraging 2D pose coordinates to estimate 3D pose and mesh. In this paper, we present PostoMETRO($\textbf{Pos}$e $\textbf{to}$ken enhanced $\textbf{ME}$sh $\textbf{TR}$ansf$\textbf{O}$rmer), which integrates occlusion-resilient 2D pose representation into transformers in a token-wise manner. Utilizing a specialized pose tokenizer, we efficiently condense 2D pose data to a compact sequence of pose tokens and feed them to the transformer together with the image tokens. This process not only ensures a rich depiction of texture from the image but also fosters a robust integration of pose and image information. Subsequently, these combined tokens are queried by vertex and joint tokens to decode 3D coordinates of mesh vertices and human joints. Facilitated by the robust pose token representation and the effective combination, we are able to produce more precise 3D coordinates, even under extreme scenarios like occlusion. Experiments on both standard and occlusion-specific benchmarks demonstrate the effectiveness of PostoMETRO. Qualitative results further illustrate the clarity of how 2D pose can help 3D reconstruction. Code will be made available.