User: Guest  Login
Document type:
Forschungsdaten 
Publication date:
28.06.2021 
Responsible:
Thiel, Kevin Kennard 
Authors:
Thiel, Kevin Kennard 
Author affiliation:
TUM 
Publisher:
TUM 
Title:
C.DOT Ground Truth Dataset 
Time of production:
04.12.2020 
Subject area:
DAT Datenverarbeitung, Informatik; VER Technik der Verkehrsmittel 
Other subject areas:
Object Tracking ; Industry 4.0 ; Ground Truth ; Augmented Reality, Computer Vision, Artificial Intelligence 
Resource type:
Abbildungen von Objekten / image of objects; Statistik und Referenzdaten / statistics and reference data 
Data type:
Bilder / images; mehrdimensionale Visualisierungen oder Modelle / models; Texte / texts; Datenbanken / data bases 
Other data type:
Camera Calibration File 
Description:
A dataset for validation of 6-dof model trackers. It includes image sets, camera calibration data, ground truth data and projection data for each image. This data is provided for four different objects. For each object 3D model data is provided too. Two of those objects are taken from automotive engineering use cases of the Volkswagen Group. While the other two models are 3D printed versions of popular computer graphics models of the Stanford University Computer Graphics Laboratory. The data...    »
 
Method of data assessment:
The data was created for research of object tracking for superimposed augmented reality in the context of automotive engineering. The images were taken using a calibrated webcam. The ground truth was captured using an accurate model-based tracker. The specific reason for capturing the data was to evaluate machine-learning based object trackers trained on synthetic data only. More information can be found in the provided readme.txt file. 
Key words:
Image; Images; Image Sets; Camera Calibration; Ground Truth; Model; Models; Objects; Tracking; Object Tracking; Model-based Tracking; 6-dof Tracking; 6-dof, Six Degrees of Freedom; Automotive; Augmented Reality; Projection; Reference; Reference Points 
Technical remarks:
View and download (1.41 GB, 2 files)
The data server also offers downloads with FTP
The data server also offers downloads with rsync (password m1614575):
rsync rsync://m1614575.001@dataserv.ub.tum.de/m1614575.001/ 
Language:
en 
Rights:
by, http://creativecommons.org/licenses/by/4.0 
Other rights:
Exception: For the reuse of the 3D printed versions of popular computer graphics models of the Stanford University Computer Graphics Laboratory, please consider the conditions of reuse of the repository (http://graphics.stanford.edu/data/3Dscanrep/). When using the data you agree to acknowledge the following related publication: “K. K. Thiel, F. Naumann, E. Jundt, S. Guennemann and G. J. Klinker, C.DOT - Convolutional Deep Object Tracker for Augmented Reality Based Purely on Synthetic Data, in I...    »