End-to-end Recovery of Human Shape and Pose

Abstract

We describe Human Mesh Recovery (HMR), an end-to-end framework for reconstructing a full 3D mesh of a human body from a single RGB image. In contrast to most current methods that compute 2D or 3D joint locations, we produce a richer and more useful mesh representation that is parameterized by shape and 3D joint angles. The main objective is to minimize the reprojection loss of keypoints, which allow our model to be trained using images in-the-wild that only have ground truth 2D annotations. However, the reprojection loss alone leaves the model highly under constrained. In this work we address this problem by introducing an adversary trained to tell whether a human body parameter is real or not using a large database of 3D human meshes. We show that HMR can be trained with and without using any paired 2D-to-3D supervision. We do not rely on intermediate 2D keypoint detections and infer 3D pose and shape parameters directly from image pixels. Our model runs in real-time given a bounding box containing the person. We demonstrate our approach on various images in-the-wild and out-perform previous optimization based methods that output 3D meshes and show competitive results on tasks such as 3D joint location estimation and part segmentation.

CVPR 2018, Project page with code: https://akanazawa.github.io/hmr/

Cited by in corpus (13)

• Self-supervised Learning of Interpretable Keypoints from Unlabelled Videos
• DeepCap: Monocular Human Performance Capture Using Weak Supervision
• DIST: Rendering Deep Implicit Signed Distance Function with Differentiable Sphere Tracing
• Single Image 3D Hand Reconstruction with Mesh Convolutions
• NeuralBlox: Real-Time Neural Representation Fusion for Robust Volumetric Mapping
• Towards Robust RGB-D Human Mesh Recovery
• Monocular 3D Reconstruction of Interacting Hands via Collision-Aware Factorized Refinements
• Post-Data Augmentation to Improve Deep Pose Estimation of Extreme and Wild Motions
• 3D Human Texture Estimation from a Single Image with Transformers
• View Invariant 3D Human Pose Estimation
• ChallenCap: Monocular 3D Capture of Challenging Human Performances using Multi-Modal References
• Coherent Reconstruction of Multiple Humans from a Single Image
• 3D Scene Parsing via Class-Wise Adaptation