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

arXiv: https://arxiv.org/abs/2004.13681 · HTML: https://mlsanity.com/paper/2004.13681/

## Key facts

- Cited by 7 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 28 Apr 2020.

## Built on by (most cited first, 7 of 7)

- [2D GANs Meet Unsupervised Single-view 3D Reconstruction](https://mlsanity.com/paper/2207.10183/) (2022)
- [Sim2Real Instance-Level Style Transfer for 6D Pose Estimation](https://mlsanity.com/paper/2203.02069/) (2022)
- [Synthetic Data Generation for Bridging Sim2Real Gap in a Production Environment](https://mlsanity.com/paper/2311.11039/) (2023)
- [6IMPOSE: Bridging the Reality Gap in 6D Pose Estimation for Robotic Grasping](https://mlsanity.com/paper/2208.14288/) (2022)
- [A Concept for Reconstructing Stucco Statues from historic Sketches using synthetic Data only](https://mlsanity.com/paper/2402.05593/) (2024)
- [One-to-many Reconstruction of 3D Geometry of cultural Artifacts using a synthetically trained Generative Model](https://mlsanity.com/paper/2402.08310/) (2024)
- [Investigation of the Impact of Synthetic Training Data in the Industrial Application of Terminal Strip Object Detection](https://mlsanity.com/paper/2403.04809/) (2024)

## Abstract (arXiv)

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 non-trivial 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 pixel-level 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 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.