Datasets:

Languages:
English
ArXiv:
Tags:
mesh
License:

Improve metadata: Set task category to robotics

#1
by nielsr HF Staff - opened
Files changed (1) hide show
  1. README.md +79 -77
README.md CHANGED
@@ -1,13 +1,15 @@
1
  ---
2
- license: cc-by-nc-nd-4.0
3
  language:
4
  - en
5
- tags:
6
- - mesh
7
- - robotics
8
  size_categories:
9
  - 10B<n<100B
 
 
 
 
10
  ---
 
11
  # Dataset Card for Dataset Name
12
 
13
  <!-- Provide a quick summary of the dataset. -->
@@ -41,98 +43,98 @@ The dataset categories is inherited from [ClothesNet](https://arxiv.org/pdf/2308
41
 
42
  ### Four Folding Types in MetaFold
43
 
44
- 1. **No-Sleeve Folding**
45
- - Categories:
46
- - DLG (Dress-Long-Gallus)
47
- - DLNS (Dress-Long-NoSleeve)
48
- - DLT (Dress-Long-Tube)
49
- - DSNS (Dress-Short-NoSleeve)
50
- - SL (Skirt-Long)
51
- - SS (Skirt-Short)
52
- - TCNC (Top-Collar-NoSleeve-FrontClose)
53
- - TCNO (Top-Collar-NoSleeve-FrontOpen)
54
- - THNC (Top-Hooded-NoSleeve-FrontClose)
55
- - TNNC (Top-NoCollar-NoSleeve-FrontClose)
56
- - Total garments: **666**
57
- - Folding procedure: Single step (fold from bottom to top).
58
-
59
- 2. **Short-Sleeve Folding**
60
- - Categories:
61
- - DLSS (Dress-Long-ShortSleeve)
62
- - DSSS (Dress-Short-ShortSleeve)
63
- - TNSC (Top-NoCollar-ShortSleeve-FrontClose)
64
- - TCSC (Top-Collar-ShortSleeve-FrontClose)
65
- - Total garments: **121**
66
- - Folding procedure (3 steps):
67
- 1. Fold the left sleeve.
68
- 2. Fold the right sleeve.
69
- 3. Fold from the bottom hem upward.
70
-
71
- 3. **Long-Sleeve Folding**
72
- - Categories:
73
- - DLLS (Dress-Long-LongSleeve)
74
- - DSLS (Dress-Short-LongSleeve)
75
- - TCLC (Top-Collar-LongSleeve-FrontClose)
76
- - TCLO (Top-Collar-LongSleeve-FrontOpen)
77
- - THLC (Top-Hooded-LongSleeve-FrontClose)
78
- - THLO (Top-Hooded-LongSleeve-FrontOpen)
79
- - TNLC (Top-NoCollar-LongSleeve-FrontClose)
80
- - Total garments: **146**
81
- - Folding procedure (3 steps):
82
- 1. Fold the left sleeve.
83
- 2. Fold the right sleeve.
84
- 3. Fold from the bottom hem upward.
85
-
86
- 4. **Pants Folding**
87
- - Categories:
88
- - PS (Pants-Short)
89
- - PL (Pants-Long)
90
- - Total garments: **277**
91
- - Folding procedure (2 steps):
92
- 1. Fold the left pant leg over the right pant leg.
93
- 2. Fold from the waistband down toward the pant legs.
94
 
95
  ---
96
 
97
  ### Total Counts
98
 
99
- - **No-Sleeve:** 666 garments
100
- - **Short-Sleeve:** 121 garments
101
- - **Long-Sleeve:** 146 garments
102
- - **Pants:** 277 garments
103
- - **Overall Total:** 1,210 garments
104
 
105
  ---
106
 
107
  ### Filename Convention
108
 
109
- Each data file follows the pattern: `<Category>_<GarmentName>_<Foldstage>`
110
 
111
- - **Category**: Garment category code (e.g., `DLG` = Dress-Long-Gallus).
112
- - **GarmentName**: Unique identifier for the specific garment model (e.g., `Dress032_1`, `010`).
113
- - **FoldStep**: Zero-based index indicating which folding step (e.g., `action0`, `action1`, …).
114
 
115
  ### Examples
116
 
117
- 1. **`DLG_Dress032_1_action0`**
118
- - **Category:** `DLG` (Dress-Long-Gallus → No-Sleeve)
119
- - **GarmentName:** `Dress032_1`
120
- - **FoldStep:** `action0` (since No-Sleeve has only one step: fold from bottom to top)
121
- - **Link:** [DLG_Dress032_1_action0](https://huggingface.co/datasets/chenhn02/MetaFold/tree/main/DLG/DLG_Dress032_1_action0)
122
 
123
- 2. **`TNLC_010_action1`**
124
- - **Category:** `TNLC` (Tops-NoCollar-LongSleeve-Close → Long-Sleeve)
125
- - **GarmentName:** `010`
126
- - **FoldStep:** `action1` (second step: fold the right sleeve)
127
- - **Link:** [TNLC_010_action1](https://huggingface.co/datasets/chenhn02/MetaFold/tree/main/TNLC/TNLC_010_action1)
128
 
129
 
130
  ### Directory Content
131
 
132
  Each Folding Trajectory directory contains:
133
- - Mesh files from `frame_00.obj` to `frame_20.obj`.
134
- - `initial.obj` is identical to `frame_00.obj`.
135
- - `keypoints_idx.npy` (specifies the folding start and end points for each step: `[start_0, end_0, start_1, end_1]`).
136
 
137
  If you need the point cloud, simply extract the vertex coordinates from the mesh files.
138
 
@@ -180,4 +182,4 @@ We use data from [ClothesNet](https://arxiv.org/pdf/2308.09987) and employ DiffC
180
  ## Dataset Card Contact
181
 
182
  First Author: [Haonan Chen](mailto:chenhaonan@u.nus.edu)
183
- Corresponding Author: [Lin Shao](mailto:linshao@nus.edu.sg)
 
1
  ---
 
2
  language:
3
  - en
4
+ license: cc-by-nc-nd-4.0
 
 
5
  size_categories:
6
  - 10B<n<100B
7
+ task_categories:
8
+ - robotics
9
+ tags:
10
+ - mesh
11
  ---
12
+
13
  # Dataset Card for Dataset Name
14
 
15
  <!-- Provide a quick summary of the dataset. -->
 
43
 
44
  ### Four Folding Types in MetaFold
45
 
46
+ 1. **No-Sleeve Folding**
47
+ - Categories:
48
+ - DLG (Dress-Long-Gallus)
49
+ - DLNS (Dress-Long-NoSleeve)
50
+ - DLT (Dress-Long-Tube)
51
+ - DSNS (Dress-Short-NoSleeve)
52
+ - SL (Skirt-Long)
53
+ - SS (Skirt-Short)
54
+ - TCNC (Top-Collar-NoSleeve-FrontClose)
55
+ - TCNO (Top-Collar-NoSleeve-FrontOpen)
56
+ - THNC (Top-Hooded-NoSleeve-FrontClose)
57
+ - TNNC (Top-NoCollar-NoSleeve-FrontClose)
58
+ - Total garments: **666**
59
+ - Folding procedure: Single step (fold from bottom to top).
60
+
61
+ 2. **Short-Sleeve Folding**
62
+ - Categories:
63
+ - DLSS (Dress-Long-ShortSleeve)
64
+ - DSSS (Dress-Short-ShortSleeve)
65
+ - TNSC (Top-NoCollar-ShortSleeve-FrontClose)
66
+ - TCSC (Top-Collar-ShortSleeve-FrontClose)
67
+ - Total garments: **121**
68
+ - Folding procedure (3 steps):
69
+ 1. Fold the left sleeve.
70
+ 2. Fold the right sleeve.
71
+ 3. Fold from the bottom hem upward.
72
+
73
+ 3. **Long-Sleeve Folding**
74
+ - Categories:
75
+ - DLLS (Dress-Long-LongSleeve)
76
+ - DSLS (Dress-Short-LongSleeve)
77
+ - TCLC (Top-Collar-LongSleeve-FrontClose)
78
+ - TCLO (Top-Collar-LongSleeve-FrontOpen)
79
+ - THLC (Top-Hooded-LongSleeve-FrontClose)
80
+ - THLO (Top-Hooded-LongSleeve-FrontOpen)
81
+ - TNLC (Top-NoCollar-LongSleeve-FrontClose)
82
+ - Total garments: **146**
83
+ - Folding procedure (3 steps):
84
+ 1. Fold the left sleeve.
85
+ 2. Fold the right sleeve.
86
+ 3. Fold from the bottom hem upward.
87
+
88
+ 4. **Pants Folding**
89
+ - Categories:
90
+ - PS (Pants-Short)
91
+ - PL (Pants-Long)
92
+ - Total garments: **277**
93
+ - Folding procedure (2 steps):
94
+ 1. Fold the left pant leg over the right pant leg.
95
+ 2. Fold from the waistband down toward the pant legs.
96
 
97
  ---
98
 
99
  ### Total Counts
100
 
101
+ - **No-Sleeve:** 666 garments
102
+ - **Short-Sleeve:** 121 garments
103
+ - **Long-Sleeve:** 146 garments
104
+ - **Pants:** 277 garments
105
+ - **Overall Total:** 1,210 garments
106
 
107
  ---
108
 
109
  ### Filename Convention
110
 
111
+ Each data file follows the pattern: `<Category>_<GarmentName>_<Foldstage>`
112
 
113
+ - **Category**: Garment category code (e.g., `DLG` = Dress-Long-Gallus).
114
+ - **GarmentName**: Unique identifier for the specific garment model (e.g., `Dress032_1`, `010`).
115
+ - **FoldStep**: Zero-based index indicating which folding step (e.g., `action0`, `action1`, …).
116
 
117
  ### Examples
118
 
119
+ 1. **`DLG_Dress032_1_action0`**
120
+ - **Category:** `DLG` (Dress-Long-Gallus → No-Sleeve)
121
+ - **GarmentName:** `Dress032_1`
122
+ - **FoldStep:** `action0` (since No-Sleeve has only one step: fold from bottom to top)
123
+ - **Link:** [DLG_Dress032_1_action0](https://huggingface.co/datasets/chenhn02/MetaFold/tree/main/DLG/DLG_Dress032_1_action0)
124
 
125
+ 2. **`TNLC_010_action1`**
126
+ - **Category:** `TNLC` (Tops-NoCollar-LongSleeve-Close → Long-Sleeve)
127
+ - **GarmentName:** `010`
128
+ - **FoldStep:** `action1` (second step: fold the right sleeve)
129
+ - **Link:** [TNLC_010_action1](https://huggingface.co/datasets/chenhn02/MetaFold/tree/main/TNLC/TNLC_010_action1)
130
 
131
 
132
  ### Directory Content
133
 
134
  Each Folding Trajectory directory contains:
135
+ - Mesh files from `frame_00.obj` to `frame_20.obj`.
136
+ - `initial.obj` is identical to `frame_00.obj`.
137
+ - `keypoints_idx.npy` (specifies the folding start and end points for each step: `[start_0, end_0, start_1, end_1]`).
138
 
139
  If you need the point cloud, simply extract the vertex coordinates from the mesh files.
140
 
 
182
  ## Dataset Card Contact
183
 
184
  First Author: [Haonan Chen](mailto:chenhaonan@u.nus.edu)
185
+ Corresponding Author: [Lin Shao](mailto:linshao@nus.edu.sg)