Learning to Estimate 3D Human Pose and Shape from a Single Color Image

TL;DR

This paper introduces a ConvNet-based method for estimating detailed 3D human pose and shape from a single image by integrating a parametric body model, enabling accurate predictions with limited training data and iterative optimization.

Contribution

It presents an end-to-end framework that predicts 3D human shape using 2D keypoints and masks, incorporating a differentiable renderer for refinement, reducing reliance on 3D ground truth data.

Findings

Outperforms previous methods on 3D human shape estimation

Uses only 2D keypoints and masks for reliable parameter prediction

Employs a differentiable renderer for improved accuracy

Abstract

This work addresses the problem of estimating the full body 3D human pose and shape from a single color image. This is a task where iterative optimization-based solutions have typically prevailed, while Convolutional Networks (ConvNets) have suffered because of the lack of training data and their low resolution 3D predictions. Our work aims to bridge this gap and proposes an efficient and effective direct prediction method based on ConvNets. Central part to our approach is the incorporation of a parametric statistical body shape model (SMPL) within our end-to-end framework. This allows us to get very detailed 3D mesh results, while requiring estimation only of a small number of parameters, making it friendly for direct network prediction. Interestingly, we demonstrate that these parameters can be predicted reliably only from 2D keypoints and masks. These are typical outputs of generic 2D human analysis ConvNets, allowing us to relax the massive requirement that images with 3D shape ground truth are available for training. Simultaneously, by maintaining differentiability, at training time we generate the 3D mesh from the estimated parameters and optimize explicitly for the surface using a 3D per-vertex loss. Finally, a differentiable renderer is employed to project the 3D mesh to the image, which enables further refinement of the network, by optimizing for the consistency of the projection with 2D annotations (i.e., 2D keypoints or masks). The proposed approach outperforms previous baselines on this task and offers an attractive solution for direct prediction of 3D shape from a single color image.