Depth Restoration of Hand-Held Transparent Objects for Human-to-Robot Handover

TL;DR

This paper introduces a novel neural network-based method for restoring accurate depth information of hand-held transparent objects in human-robot interaction, leveraging hand posture cues and a synthetic dataset for training.

Contribution

The paper proposes a new hand-aware depth restoration method using implicit neural representations and introduces a synthetic dataset for training and evaluation.

Findings

Outperforms existing methods in accuracy and generalization

Effective in real-world human-robot handover scenarios

Utilizes hand posture to improve depth estimation of transparent objects

Abstract

Transparent objects are common in daily life, while their optical properties pose challenges for RGB-D cameras to capture accurate depth information. This issue is further amplified when these objects are hand-held, as hand occlusions further complicate depth estimation. For assistant robots, however, accurately perceiving hand-held transparent objects is critical to effective human-robot interaction. This paper presents a Hand-Aware Depth Restoration (HADR) method based on creating an implicit neural representation function from a single RGB-D image. The proposed method utilizes hand posture as an important guidance to leverage semantic and geometric information of hand-object interaction. To train and evaluate the proposed method, we create a high-fidelity synthetic dataset named TransHand-14K with a real-to-sim data generation scheme. Experiments show that our method has better performance and generalization ability compared with existing methods. We further develop a real-world human-to-robot handover system based on HADR, demonstrating its potential in human-robot interaction applications.