Weakly-supervised 3D Pose Transfer with Keypoints

TL;DR

This paper introduces a weakly-supervised, keypoint-based framework for 3D pose transfer that handles different topologies and requires minimal supervision, achieving state-of-the-art results on benchmark datasets.

Contribution

The authors propose a novel topology-agnostic, weakly-supervised method for 3D pose transfer that disentangles pose and shape without needing paired data or identical topology meshes.

Findings

Achieves superior performance compared to unsupervised methods.

Comparable results to fully supervised approaches.

Demonstrates strong generalization across datasets.

Abstract

The main challenges of 3D pose transfer are: 1) Lack of paired training data with different characters performing the same pose; 2) Disentangling pose and shape information from the target mesh; 3) Difficulty in applying to meshes with different topologies. We thus propose a novel weakly-supervised keypoint-based framework to overcome these difficulties. Specifically, we use a topology-agnostic keypoint detector with inverse kinematics to compute transformations between the source and target meshes. Our method only requires supervision on the keypoints, can be applied to meshes with different topologies and is shape-invariant for the target which allows extraction of pose-only information from the target meshes without transferring shape information. We further design a cycle reconstruction to perform self-supervised pose transfer without the need for ground truth deformed mesh with the same pose and shape as the target and source, respectively. We evaluate our approach on benchmark human and animal datasets, where we achieve superior performance compared to the state-of-the-art unsupervised approaches and even comparable performance with the fully supervised approaches. We test on the more challenging Mixamo dataset to verify our approach's ability in handling meshes with different topologies and complex clothes. Cross-dataset evaluation further shows the strong generalization ability of our approach.