TuSimple Lane Challenge Class Labels

This repository contains the class labels for the lane boundaries of the TuSimple lane detection dataset. You can download the dataset from TuSimple/tusimple-benchmark#3.

Classes

Each lane boundary in the dataset is annotated using 7 different classes. Lanes that cannot be uniquely identified are annotated as Unknown. A hierarchical representation of the classes is shown below. Even if the Double-dashed-continuous class has been considered during the annotation process for completeness, there are no examples in the dataset of that class. The names of the class ids are in class_mapping.txt.

Installation

First of all, download the dataset and extract it. Then, move the files inside the data folder inside the resulting folder.

mv data/* path/to/dataset/train_set

You can obtain a .json file with an additional class field with the converter.py script. Launch it with:

python converter.py --root /path/to/dataset

It is also possible to visualize the annotations with the visualizer.py script. Launch with:

python visualizer.py --root /path/to/dataset --labels labels_json_file.json

Citation

If this dataset is useful for you, we would appreciate a citation to our paper:

@misc{pizzati2019lane,
    title={Lane Detection and Classification using Cascaded CNNs},
    author={Fabio Pizzati and Marco Allodi and Alejandro Barrera and Fernando García},
    year={2019},
    eprint={1907.01294},
    archivePrefix={arXiv},
    primaryClass={cs.CV}
}