A simple database for image classification using images from a robot car.
This repository was created by Paula Moraes and Felipe Salvatore. You can find here a very simple computer vision dataset for classification. It consists of images like
examples of: right, left, up and up images.
To download the data set just run
$ bash download.sh
The data is split into three parts: 56,172 data points of training data (train_data.npy, train_labels.npy), 7,022 points of test data (test_data.npy, test_labels.npy), and 7,022 points of validation data (valid_data.npy, valid_labels.npy). Each data point is a flattened 45x80x3 image (a 10800-dimensional vector). Each image has a corresponding label ('0', '1' and '2') representing a command for the robot car ('0' = up, '1' = left, '2' = right).
@misc{self_driving_data2018,
author = {Paula Moraes and Felipe Salvatore},
title = {Self Driving Data},
year = {2018},
howpublished = {\url{https://github.com/felipessalvatore/self_driving_data/}},
note = {commit xxxxxxx}
}