# Towards Rotation Invariance in Object Detection

arXiv: https://arxiv.org/abs/2109.13488 · HTML: https://mlsanity.com/paper/2109.13488/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 11 of its 41 references that are in the map.
- First posted on arXiv 28 Sep 2021.

## Builds on (most cited first, 11 of 11)

- [PyTorch: An Imperative Style, High-Performance Deep Learning Library](https://mlsanity.com/paper/1912.01703/) (2019)
- [Microsoft COCO: Common Objects in Context](https://mlsanity.com/paper/1405.0312/) (2014)
- [Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks](https://mlsanity.com/paper/1506.01497/) (2015)
- [EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks](https://mlsanity.com/paper/1905.11946/) (2019)
- [CutMix: Regularization Strategy to Train Strong Classifiers with Localizable Features](https://mlsanity.com/paper/1905.04899/) (2019)
- [YOLOv4: Optimal Speed and Accuracy of Object Detection](https://mlsanity.com/paper/2004.10934/) (2020)
- [Improved Regularization of Convolutional Neural Networks with Cutout](https://mlsanity.com/paper/1708.04552/) (2017)
- [MMDetection: Open MMLab Detection Toolbox and Benchmark](https://mlsanity.com/paper/1906.07155/) (2019)
- [Albumentations: fast and flexible image augmentations](https://mlsanity.com/paper/1809.06839/) (2018)
- [Learning Data Augmentation Strategies for Object Detection](https://mlsanity.com/paper/1906.11172/) (2019)
- [Learning a Rotation Invariant Detector with Rotatable Bounding Box](https://mlsanity.com/paper/1711.09405/) (2017)

## Built on by (most cited first, 2 of 2)

- [A Survey on Data Synthesis and Augmentation for Large Language Models](https://mlsanity.com/paper/2410.12896/) (2024)
- [Not Just Streaks: Towards Ground Truth for Single Image Deraining](https://mlsanity.com/paper/2206.10779/) (2022)

## Abstract (arXiv)

Rotation augmentations generally improve a model's invariance/equivariance to rotation - except in object detection. In object detection the shape is not known, therefore rotation creates a label ambiguity. We show that the de-facto method for bounding box label rotation, the Largest Box Method, creates very large labels, leading to poor performance and in many cases worse performance than using no rotation at all. We propose a new method of rotation augmentation that can be implemented in a few lines of code. First, we create a differentiable approximation of label accuracy and show that axis-aligning the bounding box around an ellipse is optimal. We then introduce Rotation Uncertainty (RU) Loss, allowing the model to adapt to the uncertainty of the labels. On five different datasets (including COCO, PascalVOC, and Transparent Object Bin Picking), this approach improves the rotational invariance of both one-stage and two-stage architectures when measured with AP, AP50, and AP75. The code is available at https://github.com/akasha-imaging/ICCV2021.