TU Darmstadt / ULB / TUbiblio

Style-transfer GANs for bridging the domain gap in synthetic pose estimator training

Rojtberg, Pavel and Pollabauer, Thomas and Kuijper, Arjan (2020):
Style-transfer GANs for bridging the domain gap in synthetic pose estimator training.
In: Proceedings : 2020 IEEE International Conference on Artificial Intelligence and Virtual Reality, pp. 188-195,
IEEE, International Conference on Artificial Intelligence and Virtual Reality (AIVR 2020), virtual Conference, 14.-18.12., ISBN 978-1-7281-7463-1,
DOI: 10.1109/AIVR50618.2020.00039,
[Conference or Workshop Item]

Abstract

Given the dependency of current CNN architectures on a large training set, the possibility of using synthetic data is alluring as it allows generating a virtually infinite amount of labeled training data. However, producing such data is a nontrivial task as current CNN architectures are sensitive to the domain gap between real and synthetic data.We propose to adopt general-purpose GAN models for pixellevel image translation, allowing to formulate the domain gap itself as a learning problem. The obtained models are then used either during training or inference to bridge the domain gap. Here, we focus on training the single-stage YOLO6D [20] object pose estimator on synthetic CAD geometry only, where not even approximate surface information is available. When employing paired GAN models, we use an edge-based intermediate domain and introduce different mappings to represent the unknown surface properties.Our evaluation shows a considerable improvement in model performance when compared to a model trained with the same degree of domain randomization, while requiring only very little additional effort.

Item Type: Conference or Workshop Item
Erschienen: 2020
Creators: Rojtberg, Pavel and Pollabauer, Thomas and Kuijper, Arjan
Title: Style-transfer GANs for bridging the domain gap in synthetic pose estimator training
Language: English
Abstract:

Given the dependency of current CNN architectures on a large training set, the possibility of using synthetic data is alluring as it allows generating a virtually infinite amount of labeled training data. However, producing such data is a nontrivial task as current CNN architectures are sensitive to the domain gap between real and synthetic data.We propose to adopt general-purpose GAN models for pixellevel image translation, allowing to formulate the domain gap itself as a learning problem. The obtained models are then used either during training or inference to bridge the domain gap. Here, we focus on training the single-stage YOLO6D [20] object pose estimator on synthetic CAD geometry only, where not even approximate surface information is available. When employing paired GAN models, we use an edge-based intermediate domain and introduce different mappings to represent the unknown surface properties.Our evaluation shows a considerable improvement in model performance when compared to a model trained with the same degree of domain randomization, while requiring only very little additional effort.

Title of Book: Proceedings : 2020 IEEE International Conference on Artificial Intelligence and Virtual Reality
Publisher: IEEE
ISBN: 978-1-7281-7463-1
Uncontrolled Keywords: Training, Solid modeling, Pose estimation, Data models
Divisions: 20 Department of Computer Science
20 Department of Computer Science > Interactive Graphics Systems
20 Department of Computer Science > Mathematical and Applied Visual Computing
Event Title: International Conference on Artificial Intelligence and Virtual Reality (AIVR 2020)
Event Location: virtual Conference
Event Dates: 14.-18.12.
Date Deposited: 25 Jan 2021 11:28
DOI: 10.1109/AIVR50618.2020.00039
Export:
Suche nach Titel in: TUfind oder in Google
Send an inquiry Send an inquiry

Options (only for editors)
Show editorial Details Show editorial Details