CloSE: A Compact Shape- and Orientation-Agnostic Cloth State Representation

2025-04-07

AI Summary

To address inconsistent feature representation under non-rigid cloth deformation, this paper proposes CloSE: a compact, continuous, shape- and orientation-invariant cloth state representation. Its core is a topology-driven discrete Gaussian Linking Integral (dGLI) disk heatmap, which partitions the mesh boundary and performs circular sampling to extract topology-invariant features—such as corners and wrinkles—and maps them into a 1D differentiable circular representation. CloSE introduces, for the first time, dGLI disk modeling based on boundary-segment topological indices, enabling unified, compact, and differentiable encoding of cloth objects with arbitrary shape, scale, and pose. Evaluated on cloth semantic segmentation and hierarchical manipulation planning tasks, CloSE significantly improves model generalization and robustness. The code, dataset, and demonstration videos are publicly available.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Representation Learning
• Search and Optimization: Mixed Discrete/Continuous Search

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Security and Privacy: Data transparency and provenance

Abstract

Cloth manipulation is a difficult problem mainly because of the non-rigid nature of cloth, which makes a good representation of deformation essential. We present a new representation for the deformation-state of clothes. First, we propose the dGLI disk representation, based on topological indices computed for segments on the edges of the cloth mesh border that are arranged on a circular grid. The heat-map of the dGLI disk uncovers patterns that correspond to features of the cloth state that are consistent for different shapes, sizes of positions of the cloth, like the corners and the fold locations. We then abstract these important features from the dGLI disk onto a circle, calling it the Cloth StatE representation (CloSE). This representation is compact, continuous, and general for different shapes. Finally, we show the strengths of this representation in two relevant applications: semantic labeling and high- and low-level planning. The code, the dataset and the video can be accessed from : https://jaykamat99.github.io/close-representation

Problem

• Represents cloth deformation state compactly and generally
• Identifies consistent features across different cloth shapes
• Enables semantic labeling and planning for cloth manipulation

Innovation

• dGLI disk representation for cloth deformation
• CloSE: compact continuous cloth state representation
• Topological indices for shape-agnostic features