Getting and visualizing data from SVHN

The Street View House Numbers (SVHN) Dataset, as expressed in the link, "is a real-world image dataset for developing machine learning and object recognition algorithms with minimal requirement on data preprocessing and formatting". It is a dataset created using images of house numbers got from Google Street View and labeled.

Main Functions of the program:

• Downloading the dataset.

• Loading the dataset.

• Visualizing the images from the dataset and their labels.

  • The label '0' has been replaced by '10'.

The project has been developed using Jupyter Notebook and has been tested in Python 3.8.

Installation

First of all it is necessary to install Jupyter Notebook. Follow the instructions included in the link depending on the preferences of each one.

Use the requirements.txt file and pip to install all the required libraries automatically.

pip install -r .\requirements.txt

Data

All the necessary data is downloaded using the code, anyway if it is prefered to download the data from Google Drive (it can take a lot of time from the original sources) or there is any problem when downloading the data using the code, it is provided in the following Google Drive link. Include the files in the root path of the project.

Basics about the SVHN dataset:

• There are 10 classes: from '1' to '10' ('0' by default but changed to 10 in the code)
• In this case only the cropped digits have been included.
• Each sample contains the RGB data and the image index.

Usage

The code is self-explanatory, and additional comments are included to clarify what tasks are performed and/or why are perfomed.

The output of the application is the image of the house number in 32x32 pixels and the label, so, the digit that the image contains.

Contributing

Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change.

Please make sure to update tests as appropriate.

License

Details included in the LICENCE.txt file.