Monocular Human Shape and Pose with Dense Mesh-borne Local Image Features

TL;DR

This paper introduces a novel approach for human shape and pose estimation from monocular images by utilizing dense, vertex-specific local image features instead of global features, leading to improved accuracy.

Contribution

It is the first to incorporate pixel-aligned local image features per mesh vertex using DensePose, enhancing graph convolutional network performance for human shape and pose estimation.

Findings

Local features outperform global features in accuracy.

The method achieves competitive results on standard benchmarks.

Using dense local features improves pose and shape estimation quality.

Abstract

We propose to improve on graph convolution based approaches for human shape and pose estimation from monocular input, using pixel-aligned local image features. Given a single input color image, existing graph convolutional network (GCN) based techniques for human shape and pose estimation use a single convolutional neural network (CNN) generated global image feature appended to all mesh vertices equally to initialize the GCN stage, which transforms a template T-posed mesh into the target pose. In contrast, we propose for the first time the idea of using local image features per vertex. These features are sampled from the CNN image feature maps by utilizing pixel-to-mesh correspondences generated with DensePose. Our quantitative and qualitative results on standard benchmarks show that using local features improves on global ones and leads to competitive performances with respect to the state-of-the-art.