Articulated 3D Human-Object Interactions from RGB Videos: An Empirical Analysis of Approaches and Challenges

TL;DR

This paper systematically benchmarks five approaches for reconstructing articulated 3D human-object interactions from RGB videos, revealing significant challenges and guiding future research directions.

Contribution

It provides the first comprehensive empirical analysis of various methods for articulated 3D human-object interaction reconstruction from RGB videos.

Findings

All methods struggle to achieve high accuracy even with ground truth data.

Key factors influencing difficulty are identified.

Future research directions are suggested.

Abstract

Human-object interactions with articulated objects are common in everyday life. Despite much progress in single-view 3D reconstruction, it is still challenging to infer an articulated 3D object model from an RGB video showing a person manipulating the object. We canonicalize the task of articulated 3D human-object interaction reconstruction from RGB video, and carry out a systematic benchmark of five families of methods for this task: 3D plane estimation, 3D cuboid estimation, CAD model fitting, implicit field fitting, and free-form mesh fitting. Our experiments show that all methods struggle to obtain high accuracy results even when provided ground truth information about the observed objects. We identify key factors which make the task challenging and suggest directions for future work on this challenging 3D computer vision task. Short video summary at https://www.youtube.com/watch?v=5tAlKBojZwc