Datasets

CVC-06 DPM Virtual-World Pedestrian Dataset

This dataset contains 2534 pedestrian images and 2000 background images. The pedestrian images have frontal view and left view, which are annotated as 'M' and 'L'. You may flip the pedestrians to get right view examples. Part annotations are also provided. 
Here are some examples of the dataset: 
References to this pedestrian datset should be made to at least one of the following articles:
Download: CVC07 DPM Virtual-World Dataset (1.3GB) 
Please, read the terms of use before downloading the dataset: disclaimer
More info here

CVC-06 Partially Occluded Virtual-World Pedestrian Dataset

This is a partially occluded pedestrian dataset extracted from our virtual world. The dataset consists of two sets of 2400 and 2414 croped pedestrians (with their corresponding horizontal mirrors).
Here are some examples of the dataset:
References to this pedestrian dataset should be made to the following article: Download: CVC06-Partially Occluded Pedestrian Dataset (137MB)
Please, read the terms of use before downloading the dataset: disclaimer

CVC-05 Partially Occluded Pedestrian Dataset

This is a partially occluded pedestrian dataset. The dataset consists of 593 positive frames with annotated pedestrians (with their corresponding horizontal mirrors).
Here are some examples of the dataset:
References to this pedestrian datasetshould be made to the following article: Download: CVC05-Partially Occluded Pedestrian Dataset (274MB)
Please, read the terms of use before downloading the dataset: disclaimer
 

CVC-04 Virtual-World Pedestrian Dataset 2

This is the virtual pedestrian dataset, generated using Half-Life 2 graphics engine. The dataset consists of 1208 virtual pedestrians (with their corresponding horizontal mirrors) and 6828 pedestrian-free background images to extract negatives for training.
Here are some examples of the dataset:
References to this pedestrian datasetshould be made to the following article: Download: CVC04-Virtual-Pedestrian Dataset 2 (2,86GB)
Please, read the terms of use before downloading the dataset: disclaimer
 

CVC-03 Virtual-World Pedestrian Dataset

This is the virtual pedestrian dataset, generated using Half-Life 2 graphics engine. The dataset consists of 1678 virtual pedestrians (with their corresponding horizontal mirrors) and 2048 pedestrian-free background images to extract negatives for training.
Here are some examples of the dataset:
References to this pedestrian datasetshould be made to the following article: Download: CVC03-Virtual-Pedestrian Dataset (915MB)
Please, read the terms of use before downloading the dataset: disclaimer

CVC-01 Semantic Segmentation Dataset

The color images contained in this dataset are part of the KITTI odometry dataset [Geiger]. In addition to the 70 labelled images of this dataset released with the publication of [Valentin], we have manually labelled a set of 146 images more, which we release here. We used the 70 labelled images of [Valentin] as part of our training set, as well as 100 more from our own labelled images. 46 of our labelled images were used for testing.
Here there is an example of the dataset: 
000000000000
Please, note that we provide here not only our labelled images but, for the convenience of the interested researchers, we provide also the associated images of the KITTI odometry dataset. Thus, you have to respect their condition of use too. Regarding the 146 labelled images that we provide, you can use them as far as you cite our associated WACV paper:
  • G. Ros*, S. Ramos*, M. Granados, A. Bakhtiary, D. Vazquez and A.M. Lopez. (*equal contribution). Vision-based Offline-Online Perception Paradigm for Autonomous Driving. In Winter Conference on Applications of Computer Vision (WACV), 2015.
@inproceedings{ros:2015,
author    = {G. Ros and S. Ramos and M. Granados and A. Bakhtiary and D. Vazquez and {A.M.} Lopez},
title     = {Vision-based Offline-Online Perception Paradigm for Autonomous Driving},
booktitle = {WACV},
year      = {2015}
}
  More info here
Please, read the terms of use before downloading the dataset: disclaimer

Road Image dataset


This database addresses the need for experimental data to quantitatively evaluate emerging road detection and photometric invariant algorithms. These images were captured using an onboard camera mounted on the windshield of a moving vehicle. The main challenge is dealing with lighting variations and shadows. The database consists (right now) of two sequences acquired on the same scenario at different daytime and under different weather conditions. The first sequence was acquired at noon of a sunny day. The second was acquired in the morning after raining. The camera is a Bumblebee (stereo pairs will be available soon) working with fixed parameters. Road groundtruth has been manually generated. Here are some examples of the database:

Day sequence Rain sequence

References to this database should be made to the following article:

Road Detection Based on Illuminant Invariance. José M. Álvarez and Antonio M. López.  In IEEE Transactions on Intelligent Transportation Systems, 2011.

Download: Day Sequence and Rain Sequence