HUMBI: A Large Multiview Dataset of Human Body Expressions

TL;DR

HUMBI is a comprehensive multiview dataset capturing diverse human body expressions with synchronized HD cameras, enabling detailed 3D reconstruction and modeling of appearance and geometry across various subjects.

Contribution

The paper introduces HUMBI, a large-scale multiview dataset with 772 subjects and 107 cameras, facilitating advanced modeling of human body expressions in natural clothing.

Findings

HUMBI enables high-fidelity 3D reconstruction of body expressions.

It is effective for learning view-specific appearance and geometry.

Complementary to existing datasets like MPII-Gaze and Human3.6M.

Abstract

This paper presents a new large multiview dataset called HUMBI for human body expressions with natural clothing. The goal of HUMBI is to facilitate modeling view-specific appearance and geometry of gaze, face, hand, body, and garment from assorted people. 107 synchronized HD cameras are used to capture 772 distinctive subjects across gender, ethnicity, age, and physical condition. With the multiview image streams, we reconstruct high fidelity body expressions using 3D mesh models, which allows representing view-specific appearance using their canonical atlas. We demonstrate that HUMBI is highly effective in learning and reconstructing a complete human model and is complementary to the existing datasets of human body expressions with limited views and subjects such as MPII-Gaze, Multi-PIE, Human3.6M, and Panoptic Studio datasets.