Research

Jointly Optimize Data Augmentation and Network Training: Adversarial Data Augmentation in Human Pose Estimation

Computer Vision

Authors

Published on

05/24/2018

Random data augmentation is a critical technique to avoid overfitting in training deep neural network models. However, data augmentation and network training are usually treated as two isolated processes, limiting the effectiveness of network training. Why not jointly optimize the two? We propose adversarial data augmentation to address this limitation. The main idea is to design an augmentation network (generator) that competes against a target network (discriminator) by generating `hard’ augmentation operations online. The augmentation network explores the weaknesses of the target network, while the latter learns from hard augmentations to achieve better performance. We also design a reward/penalty strategy for effective joint training. We demonstrate our approach on the problem of human pose estimation and carry out a comprehensive experimental analysis, showing that our method can significantly improve state-of-the-art models without additional data efforts.

Please cite our work using the BibTeX below.


@article{DBLP:journals/corr/abs-1805-09707,
  author    = {Xi Peng and
               Zhiqiang Tang and
               Fei Yang and
               Rog{\'{e}}rio Schmidt Feris and
               Dimitris N. Metaxas},
  title     = {Jointly Optimize Data Augmentation and Network Training: Adversarial
               Data Augmentation in Human Pose Estimation},
  journal   = {CoRR},
  volume    = {abs/1805.09707},
  year      = {2018},
  url       = {http://arxiv.org/abs/1805.09707},
  archivePrefix = {arXiv},
  eprint    = {1805.09707},
  timestamp = {Sat, 31 Aug 2019 16:23:04 +0200},
  biburl    = {https://dblp.org/rec/journals/corr/abs-1805-09707.bib},
  bibsource = {dblp computer science bibliography, https://dblp.org}
}

Close Modal