GarmentLab: A Unified Simulation and Benchmark for Garment Manipulation

Manipulating garments and fabrics has long been a critical endeavor in the development of home-assistant robots. However, due to complex dynamics and topological structures, garment manipulations pose significant challenges. Recent successes in reinforcement learning and vision-based methods offer promising avenues for learning garment manipulation. Nevertheless, these approaches are severely constrained by current benchmarks, which offer limited diversity of tasks and unrealistic simulation behavior. Therefore, we present GarmentLab, a content-rich benchmark and realistic simulation designed for deformable object and garment manipulation. Our benchmark encompasses a diverse range of garment types, robotic systems and manipulators. The abundant tasks in the benchmark further explores of the interactions between garments, deformable objects, rigid bodies, fluids, and human body. Moreover, by incorporating multiple simulation methods such as FEM and PBD, along with our proposed sim-to-real algorithms and real-world benchmark, we aim to significantly narrow the sim-to-real gap. We evaluate state-of-the-art vision methods, reinforcement learning, and imitation learning approaches on these tasks, highlighting the challenges faced by current algorithms, notably their limited generalization capabilities. Our proposed open-source environments and comprehensive analysis show promising boost to future research in garment manipulation by unlocking the full potential of these methods. We guarantee that we will open-source our code as soon as possible. You can watch the videos in supplementary files to learn more about the details of our work. Our project page is available at: https://garmentlab.github.io/

Paper

References (72)

Scroll for more · 38 remaining

Similar papers

Peer review

Reviewer 6kGg6/10 · confidence 3/52024-07-10

Summary

The authors present GarmentLab, a content-rich benchmark and realistic simulation designed for deformable object and garment manipulation. Multiple tasks such as interactions between garments, deformable objects, rigid bodies, fluids, and human body are explored. A first real-world deformable benchmark along with several sim2real methods is also presented in this paper. The integration of ROS and MoveIt is also supported.

Strengths

- The proposed framework is multi-functional supporting simulation of various interactions and coupling for garment manipulation scenarios. In addition to simulation, several sim-2-real algorithms such as Teleoperation, Visual Sim-Real Alignment are provided. The integration with ROS and MoveIt also help the framework to be a off-the-shelf tool for the community to use. - Extensive evaluations and experiments are presented. - The paper is well-written and organized.

Weaknesses

The main contributions of this paper lie in the engineering and benchmarking (or dataset) aspects. Although I acknowledge the substantial contribution to the community, the novelty of this work remains a concern.

Questions

See weaknesses.

Rating

6

Confidence

3

Soundness

3

Presentation

3

Contribution

3

Limitations

Yes.

Reviewer 1QRi6/10 · confidence 4/52024-07-11

Summary

This paper proposes a content-rich benchmark and realistic simulation for deformation object and garment manipulation. The GarmentLab consists of the GarmentLab Engine, Assets, and Benchmark, supporting various research topics.

Strengths

* The GarmentLab provides a realistic and rich environment for research topics related to garments. * The engine supports different interactions between garments and other objects. * The author further benchmarks various garment manipulation tasks based on this environment. * Different sim2real methods are integrated into the GarmentLab.

Weaknesses

Please refer to the questions.

Questions

1. Will all the 3 components, i.e. the Engine, Assets and Benchmark, be publicly available? 2. As for the physics engine, is it possible to simulate the interactions among different layers of garments? And will different layers of garments really affect each other instead of generating dynamics layer by layer? For example, a human wearing multiple layers of garments doing customized actions. 3. What about the interactions between garments and fluid? Will the garment dynamics also be affected through the interactions with fluid? Some simulation engines do not support the real interactions between garments and fluid. They may generate the garment dynamics first, fix the dynamic sequence, and then simulate the fluid. Can the proposed engine solve this issue? 4. Is it possible to generate customised dynamics using the proposed physics engine? For example, using customised garments and interacting with other deformable objects.

Rating

6

Confidence

4

Soundness

3

Presentation

2

Contribution

3

Limitations

Please refer to the questions.

Authorsrebuttal2024-08-12

Look forward to hearing from you

Thank you again for your constructive feedback! As the discussion period is coming to an end, we want to know whether our responses have addressed your concerns, and if you have further questions. Thank you!

Reviewer E1Nw7/10 · confidence 3/52024-07-12

Summary

This paper proposes GarmentLab, a content-rich benchmark and realistic simulation specifically designed for deformable objects and garment operations. It encompasses a diverse range of garment types, robotic systems, and manipulators.

Strengths

1. The GarmentLab Environment offers a realistic and comprehensive setting for garment manipulation. Additionally, the GarmentLab Benchmark covers a wide range of garment manipulation tasks and serves as the first real-world benchmark for these activities. 2. Although this job is primarily focused on engineering, it makes a certain contribution to the development of the garment manipulation community.

Weaknesses

1. The paper mentions a characteristic of GarmentLab is efficiency (Line 59), however, there are no quantifiable comparison with other enviroment in the speed and GPU memory during training/testing. 2. The modal of the GarmentLab environment is limited to visual (RGB-D), and other modalities such as touch and hearing can also be helpful when operating garment objects. Supporting this will help further development.

Questions

See weaknesses.

Rating

7

Confidence

3

Soundness

3

Presentation

3

Contribution

3

Limitations

The paper mentions that the limitation is the sim2real gap, perhaps with the development of simulation engines, this problem will be alleviated.

Reviewer wEM96/10 · confidence 4/52024-07-13

Summary

This paper introduces GarmentLab, a new set of garment manipulation environments, assets, annotations and tasks based on IssacSim. The differences compared to existing benchmarks is the support of a broader range of tasks (such as tasks involving interaction with other objects such as hanger, fluid, and human avatar), more robot types, more assets, and other features such as motion planning, teleoperation, and sim2real transfer. A real-world garment manipulation benchmark is also proposed with the aim for evaluating real-world performances. Several existing algorithms are tested on the proposed tasks, and their performances are analyzed.

Strengths

- The overall system looks very comprehensive and cover most of the needs for garment manipulation, including a broad range of tasks, assets, and robot control methods. - The sim2real analysis and support for real-world data collection are nice features. - Overall the whole system looks impressive and can be a good contribution to the deformable object manipulation community.

Weaknesses

I mainly have the following questions: - Regarding the real-world benchmarks: I appreciate the authors provide the 3d scans of the real objects and provide semantic annotations. However, I think the proposed real-world benchmark still lacks details such that every lab can easily replicate the set up in the real world. For example, it would be ideal to include purchase links for the chose objects. The protocols presented in F.3 seems to be just describing the task in a very high-level way, without details such as how the robot / objects should exactly be initialized, e.g., their relative position, and how to exactly measure the success / performance. Without such details it would not be possible for other labs to perfectly replicate this real-world setups and to form a fair comparison for different algorithms. I would encourage the authors to provide more details on such protocols to ensure that this can be well replicated in any lab, otherwise this proposed real-world benchmark would seem infeasible. - For some readers not familiar with the context, it is hard to understand the right part of figure 6. How can one interpret this figure to understand that the performance becomes better after applying the sim2real method such as point cloud align? E.g., what does the color scheme mean, and why is the color scheme after better than before? Also what is the figure in the dashed bounding box supposed to show? I would suggest to revise this figure to make these more clear. - Since this is a benchmarking and dataset paper, it is a little bit hard to access its quality and potential contribution to the community without actually using/seeing the code. For example, numerous robotics benchmarking has been proposed in the past, but only a few of them will be widely adopted by the community, where factor such as easy-to-use APIs, intuitive UI design, informative debugging tools, are all critical for the survival of the benchmarking. As the code of this paper is not released, it's hard to access the above qualities. I would encourage the author to release the code as soon as possible, and take into consideration the above factors, especially given the proposed system seems complicated and involve many different components. - An error in table 1: PyBullet actually support garments. See Antonova et al., Dynamic Environments with Deformable Objects, NeurIPS Datasets and Benchmarks Track, 2021. This paper is also highly relevant and should probably be discussed in related work. - There are some typos in the paper, e.g., line 6 in the abstract: "which exhibit offer limited diversity" -> "which offer limited diversity". Please have a thorough check over the whole paper for typos.

Questions

Please see the weakness sections.

Rating

6

Confidence

4

Soundness

3

Presentation

3

Contribution

3

Limitations

Yes.

Reviewer E1Nw2024-08-09

Official Comment by Reviewer E1Nw

Thanks for the well-written rebuttal. I keep my score and support the acceptance of this paper. Looking forward to the release of the code and hoping that you can continue to maintain it.

Reviewer wEM92024-08-11

Official Comment

I want to thank the authors for the detailed and well-written rebuttal. I don't have further questions, and I remain positive about this paper.

Reviewer 6kGg2024-08-11

Reply to rebuttal

Thanks for the reply to my concerns. I don't have further questions and will maintain my score.

Program Chairsdecision2024-09-25

Decision

Accept (poster)

© 2026 NYSGPT2525 LLC