Image-to-Image Translation with GAN for Synthetic Data Augmentation in Plant Disease Datasets
1
조회
0
다운로드
3
인용
0
참고문헌
키워드
초록
In recent research, deep learning-based methods have achieved state-of-the-art performance in various computer vision tasks. However, these methods are commonly supervised, and require huge amounts of annotated data to train. Acquisition of data demands an additional costly effort, particularly for the tasks where it becomes challenging to obtain large amounts of data considering the time constraints and the requirement of professional human diligence. In this paper, we present a data level synthetic sampling solution to learn from small and imbalanced data sets using Generative Adversarial Networks (GANs). The reason for using GANs are the challenges posed in various fields to manage with the small datasets and fluctuating amounts of samples per class. As a result, we present an approach that can improve learning with respect to data distributions, reducing the partiality introduced by class imbalance and hence shifting the classification decision boundary towards more accurate results. Our novel method is demonstrated on a small dataset of 2789 tomato plant disease images, highly corrupted with class imbalance in 9 disease categories. Moreover, we evaluate our results in terms of different metrics and compare the quality of these results for distinct classes.
Abstract
In recent research, deep learning-based methods have achieved state-of-the-art performance in various computer vision tasks. However, these methods are commonly supervised, and require huge amounts of annotated data to train. Acquisition of data demands an additional costly effort, particularly for the tasks where it becomes challenging to obtain large amounts of data considering the time constraints and the requirement of professional human diligence. In this paper, we present a data level synthetic sampling solution to learn from small and imbalanced data sets using Generative Adversarial Networks (GANs). The reason for using GANs are the challenges posed in various fields to manage with the small datasets and fluctuating amounts of samples per class. As a result, we present an approach that can improve learning with respect to data distributions, reducing the partiality introduced by class imbalance and hence shifting the classification decision boundary towards more accurate results. Our novel method is demonstrated on a small dataset of 2789 tomato plant disease images, highly corrupted with class imbalance in 9 disease categories. Moreover, we evaluate our results in terms of different metrics and compare the quality of these results for distinct classes.
참고문헌 (36)
- Y. LeCun, "Deep learning", Nature, 521, pp.436~444 (2015)
- K. He, "Delving deep into rectifiers:Surpassing human-level performance on ImageNet classification", 2015IEEE International Conference on Computer Vision (ICCV). IEEE Computer Society, pp.1026~1034 (2015)
- "Deep Learning" (2016)[URL]
- O. Ronneberger, "U-Net: Convolutional networks for biomedical image segmentation", Medical Image Computing and Computer-Assisted Intervention – MICCAI 2015, pp.234~241 (2015), Springer
- Hughes, D.P, "An open access repository of images on plant health to enable the development of mobile disease diagnostics", ~ (2016)
- Cire¸san, D.C, "High performance neural networks for visual object classification", ~ (2011)
- Krizhevsky, A, "Imagenet classification with deep convolutional neural networks", NIPS, 1097~1105 (2012)
- A. Rozantsev, "On rendering synthetic images for training an object detector", Computer Vision and Image Understanding (CVIU), 137, pp.24~37 (2015)
- G. Riegler, "Anatomical landmark detection in medical applications driven by synthetic data", 2015 IEEE International Conference on Computer Vision Workshop (ICCVW), pp.85~89 (2015)
- I. Goodfellow, "Generative Adversarial Nets", Advances in Neural Information Processing Systems, 27, pp.2672~2680 (2014), Curran Associates, Inc
- T. Karras, "Progressive growing of GANs for improved quality, stability, and variation", Proceedings of the Sixth International Conference on Learning Representations (ICLR) (2018)
- A. Shrivastava, "Learning from simulated and unsupervised images through adversarial training", 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp.2242~2251 (2017)
- T. Neff, "Generative Adversarial ˇNetwork based synthesis for supervised medical image segmentation", Proceedings of the OAGM&ARW Joint Workshop, pp.140~145 (2017)
- Tim Salimans, "Improved Techniques for Training GANs", ~ (2016)
- Japkowicz, N, "The class imbalance problem: Significance and strategies", Proceedings of the 2000International Conference on Artificial Intelligence (ICAI ’ 2000) (2000)
- P. Costa, "Towards adversarial retinal image synthesis", ~ (2017)
- D. Nie, "Image Synthesis with ContextAware Generative Adversarial Networks", pp.417~425 (2017), Springer International Publishing
- Z. Yi, "DualGAN: Unsupervised Dual Learning for Image-to-Image Translation", ~ (2017)
- M. -Y. Liu, "Unsupervised Image-to-Image Translation Networks", ~ (2017)
- J. -Y. Zhu, "Unpaired Image-to-Image Translation using CycleConsistent Adversarial Networks", ~ (2017)
- Y. Choi, "StarGAN:Unified Generative Adversarial Networks for Multi-Domain Image-to-Image Translation", ~ (2017)
- A. Radford, "Unsupervised representation learning with deep convolutional generative adversarial networks", ~ (2015)
- V. Dumoulin, "A guide to convolution arithmetic for deep learning", ~ (2016)
- Tung, H.-Y. F, "Adversarial inverse graphics networks: Learning 2d-to3d lifting and image-to-image translation from unpaired supervision", The IEEE International Conference on Computer Vision (ICCV), 2, pp.4364~4372 (2017)
- Wolf, L, "Unsupervised creation of parameterized avatars", ~ (2017)
- Hoffman, J., "Cycleconsistent adversarial domain adaptation", ~ (2017)
- Li, C., "Towards understanding adversarial learning for joint distribution matching", Advances in Neural Information Processing Systems, pp.5501~5509 (2017)
- Cire¸san, D. C., "High performance neural networks for visual object classification", ~ (2011)
- Simard, P. Y., "Best practices for convolutional neural networks applied to visual document analysis", ICDAR, 3, pp.958~962 (2003)
- I. Gulrajani, "Improved training of wasserstein gans", ~ (2017)
- O. Ronneberger, "U-net: Convolutional networks for biomedical image segmentation", International Conference on Medical Image Computing and Computer-Assisted Intervention, pp.234~241 (2015)
- M. Heusel, "Gans trained by a two time-scale update rule converge to a local nash equilibrium", Advances in Neural Information Processing Systems, pp.6629~6640 (2017)
- Talebi, H., "NIMA:neural image assessment", IEEE Trans. Image Process, 27, pp.3998~4011 (2018)
- 김창성, "다중 프레임에서의 보행자 검출 및 삭제 알고리즘", 스마트미디어저널, 4(2), pp.26~33 (2015), (사)한국스마트미디어학회
- 김서정, "딥러닝을 이용한 양파 밭의 잡초검출 연구", 스마트미디어저널, 7(3), pp.16~21 (2018), (사)한국스마트미디어학회[DOI]
- Nazki, H, "Synthetic Data Augmentation for Plant Disease Image Generation using GAN", Proceedings of the Korea Contents Association Conference, pp.459~460