Dataset Card for Street View House Numbers#

The Street View House Numbers (SVHN) dataset is a large real-world image dataset used for developing machine learning and object recognition algorithms. It contains over 600,000 labeled images of house numbers taken from Google Street View.

The images are cropped to a fixed resolution of 32x32 pixels, centered around a single character but may contain some distractors at the sides.

SVHN is similar to the MNIST dataset but incorporates significantly more labeled data and comes from a harder, unsolved, real-world problem of recognizing digits and numbers in natural scene images.

The dataset here is provided as original images with character level bounding boxes

This is a FiftyOne dataset with 33402 samples.

The recipe notebook for creating this dataset can be found here

Installation#

If you haven’t already, install FiftyOne:

pip install -U fiftyone

Usage#

import fiftyone as fo
import fiftyone.utils.huggingface as fouh

# Load the dataset
# Note: other available arguments include 'max_samples', etc
dataset = fouh.load_from_hub("Voxel51/StreetViewHouseNumbers")

# Launch the App
session = fo.launch_app(dataset)

Dataset Details#

Curated by: Yuval Netzer, Tao Wang, Adam Coates, Alessandro Bissacco, Bo Wu, Andrew Y. Ng
Funded by: Google Inc., Stanford University
Shared by: [More Information Needed]
License: non-commercial use only

For questions regarding the dataset, please contact streetviewhousenumbers@gmail.com

Dataset Sources#

Repository: http://ufldl.stanford.edu/housenumbers
Paper: http://ufldl.stanford.edu/housenumbers/nips2011_housenumbers.pdf

Citation#

BibTeX:

@inproceedings{netzer2011reading,
  title={Reading digits in natural images with unsupervised feature learning},
  author={Netzer, Yuval and Wang, Tao and Coates, Adam and Bissacco, Alessandro and Wu, Bo and Ng, Andrew Y},
  booktitle={NIPS workshop on deep learning and unsupervised feature learning},
  volume={2011},
  number={2},
  pages={5},
  year={2011}
}