GENERATING SYNTHETIC TRAINING DATASETS USING CONDITIONAL GENERATIVE ADVERSARIAL NETWORK TO IMPROVE IMAGES SEGMENTATION
DOI:
https://doi.org/10.34123/jurnalasks.v15i1.532Abstract
Limited amount of training datasets in deep learning research could impact the accuracy of the resulting models. This situation can cause overfit, so the model cannot work correctly. A conditional Generative Adversarial Network (CGAN) was introduced to generate synthetic data by considering certain conditions. This study aims to generate additional synthetic training datasets to improve the accuracy of the object segmentation model of images. Firstly, we evaluated CGAN-based dataset generator accuracy against several open datasets. Then, we applied the generator to train two object segmentation models, i.e., FCN and CNN U-Net. Our evaluation shows that CGAN can generate synthetic datasets well. Complex datasets require more training iterations. It also improves the validation loss and validation accuracy of both segmentation models, although other metrics still need further improvement
Downloads
References
Y. Roh, G. Heo och S. E. Whang, ”A Survey on Data Collection for Machine Learning: a Big Data -- AI Integration Perspective,” IEEE Transactions on Knowledge and Data Engineering, vol. 33, nr 4, pp. 1328 - 1347, 2019.
S. Mahapatra, ”Towards Data Science,” 22 March 2018. [Online]. Available: https://towardsdatascience.com/why-deep-learning-is-needed-over-traditional-machine-learning-1b6a99177063. [Använd 26 November 2021].
N. Papernot, P. McDaniel, S. Jha, M. Fredrikson, Z. B. Celik och A. Swami, ”The Limitations of Deep Learning in Adversarial Settings,” IEEE European Symposium on Security & Privacy, IEEE 2016. Saarbrucken, Germany, pp. 372-387, 2016.
D. Nyberg, ”Exploring the Capabilities of Generative Adversarial Networks in Remote Sensing Applications,” Linkoping University Electronic Press, Linkoping, 2021.
A. Antoniou, A. Storkey och H. Edwards, ”Data Augmentation Generative Adversarial Networks,” pp. 1-14, 2018.
Ministry of Defence, UK, Machine Learning with Limited Data, London: Data Science and Technology Laboratory, 2020.
D. Zhu, S. Xia, J. Zhao, Y. Zhou, Y. Zhou, M. Jian, Q. Niu, Q. Niu, R. Yao och Y. Chen, ”Diverse sample generation with multi-branch conditional generative adversarial network for remote sensing objects detection,” Neurocomputing, vol. 381, pp. 40-51, 2020.
M. D. Hossain och D. Chen, ”Segmentation for Object-Based Image Analysis (OBIA): A review of algorithms and challenges from remote sensing perspective,” ISPRS Journal of Photogrammetry and Remote Sensing, vol. 150, pp. 15-134, 2019.
T. Kattenborn, J. Leitloff, F. Schiefer och F. Schiefer, ”Review on Convolutional Neural Networks (CNN) in vegetation remote sensing,” ISPRS Journal of Photogrammetry and Remote Sensing, vol. 173, pp. 24-49, 2021.
A. Kulkarni, T. Mohandoss, D. Northrup, B. Hill, S. Louis, E. Mwebaze och H. Alemohammad, ”Semantic Segmentation of Medium-Resolution Satellite Imagery using Conditional Generative Adversarial Networks,” AI for Earth Sciences Workshop at NeurIPS , 2020.
M. Rezaei, K. Harmuth, W. Gierke, K. T. Gierke, M. Fischer, H. Yang och C. Meinel, ”Conditional Adversarial Network for Semantic Segmentation of Brain Tumor,” 2017.
I. J. Goodfellow, J. Pouget-Abadie, M. Mirza, B. Xu, D. Warde-Farley, S. Ozair, A. Courville och Y. Bengio, ”Generative Adversarial Nets,” Advances in Neural Information Processing Systems, pp. 1-9, June 2014.
M. Mirza och S. Osindero, ”Conditional Generative Adversarial Nets,” ArXiv, vol. abs/1411.1784, pp. 1-7, November 2014.
P. Yu, K. Song och J. Lu, ”Generating Adversarial Examples With Conditional Generative Adversarial Net,” i 2018 24th International Conference on Pattern Recognition (ICPR), Beijing, 2018.
G. Heilemann, M. Matthewman, P. Kuess, G. Goldner, J. Widder, D. Georg och L. Zimmermann, ”Can Generative Adversarial Networks help to overcome the limited data problem in segmentation?,” Zeitschrift fur Medizinische Physik, pp. 1-8, 2021.
A. F. Frangi, J. A. Schnabel, C. Davatzikos, C. Alberola-López och G. Fichtinger, ”Conditional Generative Adversarial and Convolutional Networks for X-ray Breast Mass Segmentation and Shape Classification,” Medical Image Computing and Computer Assisted Intervention, vol. 11071, pp. 833-840, 2018.
K. Cheng, R. Tahir, L. K. Eric och M. Li, ”An Analysis of Generative Adversarial Networks and Variants for Image Synthesis on MNIST Dataset,” Multimedia Tools and Applications, vol. 79, nr 19-20, pp. 13725-13750, 2020.
J. Brownlee, Generative Adversarial Networks with Python: Deep Learning Generative Models for Image Synthesis and Image Translation, Machine Learning Mastery, 2019.
