Hugging Face's logo

Datasets:
Ramos-Ramos
/
smallnorb

Modalities:
Image
Formats:
parquet
Size:
10K - 100K
Libraries:
Datasets
pandas
Dataset card Viewer Files Community
Dataset Viewer
Auto-converted to Parquet
Full Screen
Search is not available for this dataset
image_lt
imagewidth (px)
96
96
image_rt
imagewidth (px)
96
96
category
int32
0
4
instance
int32
4
9
elevation
int32
0
8
azimuth
int32
0
34
lighting
int32
0
5
0
8
6
4
4
1
8
2
6
4
2
4
8
24
3
3
7
8
28
0
4
8
7
20
0
0
8
7
8
1
1
9
6
12
0
2
7
7
4
2
3
7
3
20
0
4
6
7
24
1
0
6
6
18
1
1
6
2
20
2
2
6
2
16
3
3
9
5
16
4
4
8
3
2
4
0
6
1
22
3
1
6
7
10
3
2
4
1
16
2
3
6
7
6
1
4
6
1
32
2
0
4
7
6
2
1
7
0
0
1
2
4
2
0
0
3
6
4
20
2
4
4
5
16
4
0
6
5
2
5
1
8
5
28
3
2
7
2
10
4
3
4
8
22
3
4
7
4
20
2
0
4
6
22
3
1
6
4
8
4
2
6
4
32
2
3
6
5
26
4
4
8
2
30
4
0
9
6
26
0
1
4
7
28
4
2
9
8
20
2
3
4
1
22
5
4
8
4
32
0
0
9
5
16
0
1
4
3
32
3
2
6
6
2
1
3
6
1
0
4
4
4
7
2
0
0
6
2
34
0
1
7
7
8
3
2
6
6
32
1
3
9
6
32
0
4
6
0
4
2
0
6
3
26
4
1
7
7
28
2
2
6
4
6
1
3
4
6
34
1
4
6
2
4
2
0
8
1
20
1
1
8
7
20
1
2
8
2
34
4
3
6
0
24
4
4
8
2
10
5
0
7
2
8
3
1
7
6
16
4
2
8
8
32
4
3
4
2
16
3
4
7
8
8
4
0
9
6
30
2
1
4
1
34
0
2
6
2
32
0
3
7
3
18
3
4
6
1
4
2
0
8
5
16
4
1
8
0
2
1
2
7
2
28
5
3
9
0
8
5
4
8
3
16
3
0
7
4
16
3
1
4
1
8
1
2
4
2
14
3
3
8
2
14
2
4
7
2
2
3
0
4
5
28
0
1
7
6
14
1
2
9
0
14
5
3
4
1
20
0
4
7
3
24
5
0
9
2
34
5
1
7
1
24
5
2
9
2
28
3
3
7
6
16
1
4
7
0
6
0
0
4
4
4
2
1
4
5
8
3
2
4
8
20
4
3
9
0
6
5
4
9
6
6
5
0
9
6
34
2
1
8
8
20
1
2
7
4
14
5
3
8
0
18
2
4
4
4
22
2
End of preview. Expand in Dataset Viewer.

Dataset Card for "smallnorb"

Dataset Description

NOTE: This dataset is an unofficial port of small NORB based on a repo from Andrea Palazzi using this script. For complete and accurate information, we highly recommend visiting the dataset's original homepage.

Dataset Summary

From the dataset's homepage:

This database is intended for experiments in 3D object reocgnition from shape. It contains images of 50 toys belonging to 5 generic categories: four-legged animals, human figures, airplanes, trucks, and cars. The objects were imaged by two cameras under 6 lighting conditions, 9 elevations (30 to 70 degrees every 5 degrees), and 18 azimuths (0 to 340 every 20 degrees).

The training set is composed of 5 instances of each category (instances 4, 6, 7, 8 and 9), and the test set of the remaining 5 instances (instances 0, 1, 2, 3, and 5).

Dataset Structure

Data Instances

An example of an instance in this dataset:

{
  'image_lt': <PIL.PngImagePlugin.PngImageFile image mode=L size=96x96 at 0x...>,
  'image_rt': <PIL.PngImagePlugin.PngImageFile image mode=L size=96x96 at 0x...>,
  'category': 0,
  'instance': 8,
  'elevation': 6,
  'azimuth': 4,
  'lighting': 4
}

Data Fields

Explanation of this dataset's fields:

  • image_lt: a PIL image of an object from the dataset taken with one of two cameras
  • image_rt: a PIL image of an object from the dataset taken with one of two cameras
  • category: the category of the object shown in the images
  • instance: the instance of the category of the object shown in the images
  • elevation: the label of the elevation of the cameras used in capturing a picture of the object
  • azimuth: the label of the azimuth of the cameras used in capturing a picture of the object
  • lighting: the label of the lighting condition used in capturing a picture of the object

For more information on what these categories and labels pertain to, please see Dataset Summary or the repo used in processing the dataset.

Data Splits

Information on this dataset's splits:

train test
size 24300 24300

Additional Information

Dataset Curators

Credits from the dataset's homepage:

Fu Jie Huang, Yann LeCun

Courant Institute, New York University

October, 2005

Licensing Information

From the dataset's homepage:

This database is provided for research purposes. It cannot be sold. Publications that include results obtained with this database should reference the following paper:

Y. LeCun, F.J. Huang, L. Bottou, Learning Methods for Generic Object Recognition with Invariance to Pose and Lighting. IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR) 2004

Citation Information

From the dataset's homepage:

Publications that include results obtained with this database should reference the following paper:

Y. LeCun, F.J. Huang, L. Bottou, Learning Methods for Generic Object Recognition with Invariance to Pose and Lighting. IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR) 2004

@inproceedings{lecun2004learning,
  title={Learning methods for generic object recognition with invariance to pose and lighting},
  author={LeCun, Yann and Huang, Fu Jie and Bottou, Leon},
  booktitle={Proceedings of the 2004 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2004. CVPR 2004.},
  volume={2},
  pages={II--104},
  year={2004},
  organization={IEEE}
}

DOI: 10.1109/CVPR.2004.1315150

Contributions

Code to process small NORB adapted from Andrea Palazzi's repo with this script.

Downloads last month
45
Size of downloaded dataset files:
237 MB
Size of the auto-converted Parquet files:
237 MB
Number of rows:
48,600