HeatFormer: A Neural Optimizer for Multiview Human Mesh Recovery

TL;DR

HeatFormer is a neural optimizer that leverages multiple static camera views to accurately and robustly recover human shape and pose, adaptable to various camera configurations for applications like elderly care and safety monitoring.

Contribution

The paper introduces HeatFormer, a novel transformer-based neural optimizer that refines SMPL parameters from multiview images, handling varying camera setups and occlusions.

Findings

High accuracy in human shape and pose estimation

Robustness to occlusion and view variation

Effective generalization across different environments

Abstract

We introduce a novel method for human shape and pose recovery that can fully leverage multiple static views. We target fixed-multiview people monitoring, including elderly care and safety monitoring, in which calibrated cameras can be installed at the corners of a room or an open space but whose configuration may vary depending on the environment. Our key idea is to formulate it as neural optimization. We achieve this with HeatFormer, a neural optimizer that iteratively refines the SMPL parameters given multiview images, which is fundamentally agonistic to the configuration of views. HeatFormer realizes this SMPL parameter estimation as heat map generation and alignment with a novel transformer encoder and decoder. We demonstrate the effectiveness of HeatFormer including its accuracy, robustness to occlusion, and generalizability through an extensive set of experiments. We believe HeatFormer can serve a key role in passive human behavior modeling.