# Revisiting Feature Alignment for One-stage Object Detection

arXiv: https://arxiv.org/abs/1908.01570 · HTML: https://mlsanity.com/paper/1908.01570/

## Key facts

- Cited by 14 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 5 Aug 2019.

## Built on by (most cited first, 14 of 14)

- [Probabilistic two-stage detection](https://mlsanity.com/paper/2103.07461/) (2021)
- [Instance Localization for Self-supervised Detection Pretraining](https://mlsanity.com/paper/2102.08318/) (2021)
- [M3DSSD: Monocular 3D Single Stage Object Detector](https://mlsanity.com/paper/2103.13164/) (2021)
- [Disentangle Your Dense Object Detector](https://mlsanity.com/paper/2107.02963/) (2021)
- [Hierarchical Feature Alignment Network for Unsupervised Video Object Segmentation](https://mlsanity.com/paper/2207.08485/) (2022)
- [Temporal Feature Alignment and Mutual Information Maximization for Video-Based Human Pose Estimation](https://mlsanity.com/paper/2203.15227/) (2022)
- [Boosting R-CNN: Reweighting R-CNN Samples by RPN's Error for Underwater Object Detection](https://mlsanity.com/paper/2206.13728/) (2022)
- [CenterNet++ for Object Detection](https://mlsanity.com/paper/2204.08394/) (2022)
- [Attentional Feature Refinement and Alignment Network for Aircraft Detection in SAR Imagery](https://mlsanity.com/paper/2201.07124/) (2022)
- [PDNet: Toward Better One-Stage Object Detection With Prediction Decoupling](https://mlsanity.com/paper/2104.13876/) (2021)
- [Efficient Person Search: An Anchor-Free Approach](https://mlsanity.com/paper/2109.00211/) (2021)
- [Gaussian Guided IoU: A Better Metric for Balanced Learning on Object Detection](https://mlsanity.com/paper/2103.13613/) (2021)
- [Being Aware of Localization Accuracy By Generating Predicted-IoU-Guided Quality Scores](https://mlsanity.com/paper/2309.13269/) (2023)
- [RotaTR: Detection Transformer for Dense and Rotated Object](https://mlsanity.com/paper/2312.02821/) (2023)

## Abstract (arXiv)

Recently, one-stage object detectors gain much attention due to their simplicity in practice. Its fully convolutional nature greatly reduces the difficulty of training and deployment compared with two-stage detectors which require NMS and sorting for the proposal stage. However, a fundamental issue lies in all one-stage detectors is the misalignment between anchor boxes and convolutional features, which significantly hinders the performance of one-stage detectors. In this work, we first reveal the deep connection between the widely used im2col operator and the RoIAlign operator. Guided by this illuminating observation, we propose a RoIConv operator which aligns the features and its corresponding anchors in one-stage detection in a principled way. We then design a fully convolutional AlignDet architecture which combines the flexibility of learned anchors and the preciseness of aligned features. Specifically, our AlignDet achieves a state-of-the-art mAP of 44.1 on the COCO test-dev with ResNeXt-101 backbone.