Learning RGB-D Feature Embeddings for Unseen Object Instance Segmentation

TL;DR

This paper introduces a novel RGB-D feature embedding approach for unseen object instance segmentation, leveraging synthetic data and metric learning to improve robot perception in cluttered environments.

Contribution

It proposes a new method using synthetic RGB-D data and metric learning to produce transferable feature embeddings for unseen object segmentation.

Findings

Effective segmentation of unseen objects in cluttered scenes.

Synthetic data can be used to train models that transfer well to real-world images.

Improved accuracy with a two-stage clustering algorithm.

Abstract

Segmenting unseen objects in cluttered scenes is an important skill that robots need to acquire in order to perform tasks in new environments. In this work, we propose a new method for unseen object instance segmentation by learning RGB-D feature embeddings from synthetic data. A metric learning loss function is utilized to learn to produce pixel-wise feature embeddings such that pixels from the same object are close to each other and pixels from different objects are separated in the embedding space. With the learned feature embeddings, a mean shift clustering algorithm can be applied to discover and segment unseen objects. We further improve the segmentation accuracy with a new two-stage clustering algorithm. Our method demonstrates that non-photorealistic synthetic RGB and depth images can be used to learn feature embeddings that transfer well to real-world images for unseen object instance segmentation.