home research publications code sandbox contact

sandbox

tango

This page contains evaluation datasets from the paper:

How to Localize Humanoids with a Single Camera? Pablo F. Alcantarilla, Olivier Stasse, Sebastien Druon, Luis M. Bergasa and Frank Dellaert. Autonomous Robots, September 2012

We provide several stereo-vision datasets with calibration parameters and ground truth information for localization. In this way, other researchers working with humanoid robots can use these datasets to evaluate their algorithms.

In addition, we will upload several videos showing the results of the experiments described in the paper.

Please check the README file before downloading any sequence.


to_straight_3m

Toulouse Dataset: Straight 3m Sequence

Stereo Rig Calibration File

Visual SLAM Translation Output File

Visual SLAM Orientation Output File

MOCAP Translation File


to_circle

Toulouse Dataset: Circle Sequence

Stereo Rig Calibration File

Visual SLAM Translation Output File

Visual SLAM Orientation Output File

MOCAP Translation File


ts_circle

Tsukuba Dataset: Circle Sequence

Stereo Rig Calibration File

Visual SLAM Translation Output File

Visual SLAM Orientation Output File


Visual SLAM Video


Vision-Based Localization

 


Robot Kidnapping