Tasks Overview

Tasks

We provide a set of 10 tasks involving deformable objects, most tasks contains 5 handmade deformable objects. There are also two procedurally generated tasks, ButtonProc and HangProcCloth, in which the deformable objects are procedurally generated. Furthermore, to improve generalzation, the v0 of each task will randomizes textures and meshes.

All tasks have -v1 and -v2 with a particular choice of meshes and textures that is not randomized. Most tasks have versions up to -v5 with additional mesh and texture variations.

Tasks with procedurally generated cloth (ButtonProc and HangProcCloth) generate random cloth objects for all versions (but randomized textures only in v0).

HangBag

python -m dedo.demo_preset --env=HangBag-v1 --viz

HangBag contains a set of 108 bag meshes. These bags have varying geometry but identical in topology. E.g. size of the handle bar, height of the body etc.

HangBag-v0: Randomly selects one of 108 bag meshes; randomized textures

HangBag-v[1-108]: Selects from one of the 108 bags

HangGarment

python -m dedo.demo_preset --env=HangGarment-v1 --viz

HangGarment-v0: hang garment with randomized textures (a few examples below):

HangGarment-v[1-5]: 5 apron meshes and texture combos shown below:

HangGarment-v[6-10]: 5 shirt meshes and texture combos shown below:

HangProcCloth

python -m dedo.demo_preset --env=HangProcCloth-v1 --viz

HangProcCloth-v0: random textures, procedurally generated cloth with 1 and 2 holes.

HangProcCloth-v[1-2]: same, but with either 1 or 2 holes

Buttoning

python -m dedo.demo_preset --env=Button-v1 --viz

ButtonProc-v0: randomized textures and procedurally generated cloth with 2 holes, randomized hole/button positions.

ButtonProc-v[1-2]: procedurally generated cloth, 1 or two holes.

Button-v0: randomized textures, but fixed cloth and button positions.

Button-v1: fixed cloth and button positions with one texture (see image below):

Hoop

python -m dedo.demo_preset --env=Hoop-v1 --viz

Hoop-v0: randomized textures Hoop-v1: pre-selected textures

Lasso

python -m dedo.demo_preset --env=Lasso-v1 --viz

Lasso-v0: randomized textures Lasso-v1: pre-selected textures

DressBag

python -m dedo.demo_preset --env=DressBag-v1 --viz

DressBag-v0, DressBag-v[1-5]: demo for -v1 shown below

Visualizations of the 5 backpack mesh and texture variants for DressBag-v[1-5]:

DressGarment

python -m dedo.demo_preset --env=DressGarment-v1 --viz

DressGarment-v0, DressGarment-v[1-5]: demo for -v1 shown below

Mask

python -m dedo.demo_preset --env=Mask-v1 --viz

Mask-v0, Mask-v[1-5]: a few texture variants shown below:

Robotics Tasks

HangGarmentRobot

python -m dedo.demo_preset --env=HangGarmentRobot-v1 --viz

HangGarmentRobot-v1: A environment for demonstrating integration with Franka Robot Arm

FoodPacking

python -m dedo.demo_preset --env=FoodPacking-v1 --viz

FoodPacking-v[0-3]: Demonstrating robotic manipulation of pushing YCB objects

Loading the Sewing Patterns Dataset

We tested DEDO on a dataset from Generating Datasets of 3D Garments with Sewing Patterns. Maria Korosteleva and Sung-Hee Lee. In Conference on Neural Information Processing Systems, Datasets and Benchmarks Track (Round 1), 2021

This dataset includes over 22000 garments. However, this dataset is using highly dense meshes, thus we process this dataset in Blender to reduce the vertices count.

To do so:

0. Make sure the latest version of Blender is installed
1. Download the dataset from author's Dropbox
2. Collect all of the *_sim.obj files from each subdirectory to a single directory (script not provided)
3. locate scripts/blender_sewing_dataset_decimation.py and point the hardcoded input dir specified in the previous step
4. run Blender through background mode blender --background --python scripts/blender_sewing_dataset_decimation.py