Cloth Region Segmentation for Robust Grasp Selection

TL;DR

This paper presents a novel method for segmenting key regions like edges and corners of cloth from depth images to improve robotic grasping, demonstrating superior success rates on real robot experiments.

Contribution

The work introduces a new segmentation network and grasp estimation algorithm specifically for cloth edges and corners, enhancing robotic manipulation accuracy.

Findings

Outperforms baseline methods in grasping success

Effective segmentation of cloth edges and corners from depth images

Robust grasp estimation with directional uncertainty modeling

Abstract

Cloth detection and manipulation is a common task in domestic and industrial settings, yet such tasks remain a challenge for robots due to cloth deformability. Furthermore, in many cloth-related tasks like laundry folding and bed making, it is crucial to manipulate specific regions like edges and corners, as opposed to folds. In this work, we focus on the problem of segmenting and grasping these key regions. Our approach trains a network to segment the edges and corners of a cloth from a depth image, distinguishing such regions from wrinkles or folds. We also provide a novel algorithm for estimating the grasp location, direction, and directional uncertainty from the segmentation. We demonstrate our method on a real robot system and show that it outperforms baseline methods on grasping success. Video and other supplementary materials are available at: https://sites.google.com/view/cloth-segmentation.