ISA\(^2\) Dataset

Introduction

Here we release the ISA\(^2\) dataset. It is a benchmark for the problem of Intelligent Speed Adaptation from Appearance. The database consists of 5 video sequences taken from both urban and interurban scenarios in the Community of Madrid, Spain. In total, we provide a set of 149.055 frames, with a size of 640 x 384 pixels, with the annotation of the proper speed of the car (km/h). During the driving for the acquisition of the dataset, in addition to respecting the speed limits, our driver has carefully tried to adjust the speed of the vehicle to what he considers to be an appropriate speed, according to the traffic situation.

Download

Best practice: Recommendations on using the dataset

ISA\(^2\) dataset is divided into two dataset: training and test. Any approach reporting results for the ISA\(^2\) dataset must be trained using any data except the provided test images. Furthermore, the test data must be used strictly for reporting of results alone - it must not be used in any way to train or tune systems, for example by running multiple parameter choices and reporting the best results obtained.

We encourage two types of participation: (i) methods which are trained using only the training set of images provided by ISA\(^2\); (ii) methods trained using any data except the test data of ISA\(^2\).

Citing

If you make use of this data and software, please cite the following reference in any publications:

@InProceedings{isa2,
  Title                    = {ISA^2 : Intelligent Speed Adaptation from Appearance},
  Author                   = {Herranz-Perdiguero, C. and López-Sastre, R.~J.},
  Booktitle                = {IROS 2018 Workshop, 10th Planning, Perception and Navigation for Intelligent Vehicles},
  Year                     = {2018}
}