# On Representation of 3D Rotation in the Context of Deep Learning

arXiv: https://arxiv.org/abs/2410.10350 · HTML: https://mlsanity.com/paper/2410.10350/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 8 of its 17 references that are in the map.
- First posted on arXiv 14 Oct 2024.

## Builds on (most cited first, 8 of 8)

- [Identity Mappings in Deep Residual Networks](https://mlsanity.com/paper/1603.05027/) (2016)
- [On the Continuity of Rotation Representations in Neural Networks](https://mlsanity.com/paper/1812.07035/) (2018)
- [Kornia: an Open Source Differentiable Computer Vision Library for PyTorch](https://mlsanity.com/paper/1910.02190/) (2019)
- [Image Matching across Wide Baselines: From Paper to Practice](https://mlsanity.com/paper/2003.01587/) (2020)
- [FS-Net: Fast Shape-based Network for Category-Level 6D Object Pose Estimation with Decoupled Rotation Mechanism](https://mlsanity.com/paper/2103.07054/) (2021)
- [On Object Symmetries and 6D Pose Estimation from Images](https://mlsanity.com/paper/1908.07640/) (2019)
- [Multi-sensor large-scale dataset for multi-view 3D reconstruction](https://mlsanity.com/paper/2203.06111/) (2022)
- [Towards Deep Learning-based 6D Bin Pose Estimation in 3D Scans](https://mlsanity.com/paper/2112.09598/) (2021)

## Built on by (most cited first, 1 of 1)

- [Grasp-MPC: Closed-Loop Visual Grasping via Value-Guided Model Predictive Control](https://mlsanity.com/paper/2509.06201/) (2025)

## Abstract (arXiv)

This paper investigates various methods of representing 3D rotations and their impact on the learning process of deep neural networks. We evaluated the performance of ResNet18 networks for 3D rotation estimation using several rotation representations and loss functions on both synthetic and real data. The real datasets contained 3D scans of industrial bins, while the synthetic datasets included views of a simple asymmetric object rendered under different rotations. On synthetic data, we also assessed the effects of different rotation distributions within the training and test sets, as well as the impact of the object's texture. In line with previous research, we found that networks using the continuous 5D and 6D representations performed better than the discontinuous ones.