UnLoc dataset (Synthetic + Real)

<p>This dataset contains the synthetic and real data used in the article &quot; Virtual Training for a Real Application: Accurate Object-Robot Relative Localization Without Calibration &quot; to train 3 convolutional neural networks (CNNs) in order to perform uncalibrated relative localization of a cuboid block with respect to a robot, and to evaluate them.&nbsp;</p> <p>It consists of a dataset composed of synthetic pictures for training the CNNs and a dataset of real pictures for evaluation.&nbsp;</p> <p>The &quot;synthetic&quot; dataset is composed 3 sub-datasets (each of them composed of thousands of synthetic pictures and corresponding groundtruth) for training :&nbsp;</p> <ol> <li>a dataset for coarse relative localization subtask : coarse_estimation_data (~14.6 GB when extracted)</li> <li>a dataset for the tool localization subtask :&nbsp;tool_detection_data&nbsp;(~4.3&nbsp;GB when extracted)</li> <li>a dataset for the fine relative localization subtask :&nbsp;fine_estimation_data&nbsp;(~13.3 GB when extracted)</li> </ol> <p>These sub-datasets are composed of raw data as well as post-treated data ready to be used for training CNNs,&nbsp;in CSV format and in Torch format (.t7).&nbsp;</p> <p>The &quot;real&quot; dataset is&nbsp;composed of real pictures&nbsp;with a precisely localized cuboid block for evaluation only : UnLoc_real (~2.5&nbsp;GB when extracted).&nbsp;</p> <p>More information available on the project page :&nbsp;<a href="http://imagine.enpc.fr/~loingvi/unloc/">http://imagine.enpc.fr/~loingvi/unloc/</a></p>