Visual Geometric Skill Inference by Watching Human Demonstration

TL;DR

This paper introduces a graph-based kernel regression approach using InMaxEnt IRL to infer geometric manipulation skills from human demonstration videos, enabling human-readable task definitions and robust control error outputs.

Contribution

It presents a novel method that infers geometric associations directly from videos without extensive feature selection, improving generalization and simplicity over traditional approaches.

Findings

Accurately infers geometric associations from a single demonstration

Generalizes well under variance in demonstrations

Eliminates need for tedious feature tracking in visual servoing

Abstract

We study the problem of learning manipulation skills from human demonstration video by inferring the association relationships between geometric features. Motivation for this work stems from the observation that humans perform eye-hand coordination tasks by using geometric primitives to define a task while a geometric control error drives the task through execution. We propose a graph based kernel regression method to directly infer the underlying association constraints from human demonstration video using Incremental Maximum Entropy Inverse Reinforcement Learning (InMaxEnt IRL). The learned skill inference provides human readable task definition and outputs control errors that can be directly plugged into traditional controllers. Our method removes the need for tedious feature selection and robust feature trackers required in traditional approaches (e.g. feature-based visual servoing). Experiments show our method infers correct geometric associations even with only one human demonstration video and can generalize well under variance.