by Richter-Klug, Jesse and Frese, Udo
Abstract:
Image based object recognition and pose estimation is nowadays a heavily focused research field important for robotic object manipulation. Despite the impressive recent success of CNNs to our knowledge none includes a self-estimation of its predicted pose's uncertainty.
Reference:
Richter-Klug, Jesse and Frese, Udo, "Towards Meaningful Uncertainty Information for CNN Based 6D Pose Estimates", In Computer Vision Systems (Tzovaras, Dimitrios, Giakoumis, Dimitrios, Vincze, Markus, Argyros, Antonis, eds.), Springer International Publishing, Cham, pp. 408–422, 2019.
Bibtex Entry:
@InProceedings{10.1007/978-3-030-34995-0_37,
author="Richter-Klug, Jesse
and Frese, Udo",
editor="Tzovaras, Dimitrios
and Giakoumis, Dimitrios
and Vincze, Markus
and Argyros, Antonis",
title="Towards Meaningful Uncertainty Information for CNN Based 6D Pose Estimates",
booktitle="Computer Vision Systems",
year="2019",
publisher="Springer International Publishing",
address="Cham",
pages="408--422",
abstract="Image based object recognition and pose estimation is nowadays a heavily focused research field important for robotic object manipulation. Despite the impressive recent success of CNNs to our knowledge none includes a self-estimation of its predicted pose's uncertainty.",
isbn="978-3-030-34995-0",
url = {https://link.springer.com/chapter/10.1007/978-3-030-34995-0_37},
keywords={easecrc_perception}
}