# Learning Data Augmentation Strategies for Object Detection

## TL;DR

This paper investigates the impact of data augmentation on object detection, demonstrating that learned, specialized augmentation policies significantly improve detection accuracy and transfer well across datasets and models.

## Contribution

The authors introduce a method to learn optimized data augmentation policies specifically for object detection, outperforming traditional augmentation and regularization techniques.

## Key findings

- Optimized augmentation policies improve detection accuracy by over 2.3 mAP.
- A single learned policy can be transferred across datasets and models.
- Learned policies outperform state-of-the-art regularization methods.

## Abstract

Data augmentation is a critical component of training deep learning models. Although data augmentation has been shown to significantly improve image classification, its potential has not been thoroughly investigated for object detection. Given the additional cost for annotating images for object detection, data augmentation may be of even greater importance for this computer vision task. In this work, we study the impact of data augmentation on object detection. We first demonstrate that data augmentation operations borrowed from image classification may be helpful for training detection models, but the improvement is limited. Thus, we investigate how learned, specialized data augmentation policies improve generalization performance for detection models. Importantly, these augmentation policies only affect training and leave a trained model unchanged during evaluation. Experiments on the COCO dataset indicate that an optimized data augmentation policy improves detection accuracy by more than +2.3 mAP, and allow a single inference model to achieve a state-of-the-art accuracy of 50.7 mAP. Importantly, the best policy found on COCO may be transferred unchanged to other detection datasets and models to improve predictive accuracy. For example, the best augmentation policy identified with COCO improves a strong baseline on PASCAL-VOC by +2.7 mAP. Our results also reveal that a learned augmentation policy is superior to state-of-the-art architecture regularization methods for object detection, even when considering strong baselines. Code for training with the learned policy is available online at https://github.com/tensorflow/tpu/tree/master/models/official/detection