Yulin Li
ABSTRACT
With the development of remote sensing technology, remote sensing data has been widely used in agriculture, medicine, military, and other fields. However, due to the disadvantages of the high cost of data collection and high redundancy, regression experiments using remote sensing data have serious overfitting problems. It limits its application in practical work. To alleviate this problem, we propose a generative adversarial network to generate remote sensing signals. In this paper, a feature mixing module was proposed to reduce the bias of the discriminator for different signals, thereby increasing the diversity of generated data. At the same time, spectral normalization is utilized to improve the stability during generation, which makes the generated data closer to the real signal. After a series of ablation experiments on small-sample remote sensing data, it is proved that the data generated by the generative adversarial network significantly improves the diversity of data and effectively alleviates the over-fitting problem based on ensuring the reliability of the data.
- Shaomin Chen, Tiantian Hu, Lihua Luo, Qiong He, Shaowu Zhang, Mengyue Li, Xiaolu Cui, and Hongxiang Li. 2020. Rapid estimation of leaf nitrogen content in apple-trees based on canopy hyperspectral reflectance using multivariate methods. Infrared Physics & Technology 111, (December 2020), 103542. DOI: https://doi.org/10.1016/j.infrared.2020.103542Google Scholar
Cross Ref
- Pengfei Wen, Zujiao Shi, Ao Li, Fang Ning, Yuanhong Zhang, Rui Wang, and Jun Li. 2021. Estimation of the vertically integrated leaf nitrogen content in maize using canopy hyperspectral red edge parameters. Precision Agric 22, 3 (June 2021), 984–1005. DOI: https://doi.org/10.1007/s11119-020-09769-Google ScholarCross Ref
- Christian Janiesch, Patrick Zschech, and Kai Heinrich. 2021. Machine learning and deep learning. Electron Markets 31, 3 (September 2021), 685–695. DOI: https://doi.org/10.1007/s12525-021-00475-2Google ScholarCross Ref
- Wartini Ng, Budiman Minasny, Wanderson De Sousa Mendes, and José A. M. Demattê. 2019. Estimation of effective calibration sample size using visible near infrared spectroscopy: deep learning vs machine learning. Soil and methods, (September 2019), 1-21. DOI: https://doi.org/10.5194/soil-2019-48Google ScholarCross Ref
- J. Padarian, B. Minasny, and A.B. McBratney. 2019. Using deep learning to predict soil properties from regional spectral data. Geoderma Regional 16, (March 2019), e00198. DOI: https://doi.org/10.1016/j.geodrs.2018.e00198Google ScholarCross Ref
- Esben Jannik Bjerrum, Mads Glahder, and Thomas Skov. 2017. Data Augmentation of Spectral Data for Convolutional Neural Network (CNN) Based Deep Chemometrics. Computer Science, (2017), 1-10. DOI: https://doi.org/10.48550/ARXIV.1710.01927Google ScholarCross Ref
- Uladzislau Blazhko, Volha Shapaval, Vassili Kovalev, and Achim Kohler. 2021. Comparison of augmentation and pre-processing for deep learning and chemometric classification of infrared spectra. Chemometrics and Intelligent Laboratory Systems 215, (August 2021), 104367. DOI: https://doi.org/10.1016/j.chemolab.2021.104367Google ScholarCross Ref
- Andrew Hahn, Murali Tummala, and James Scrofani. 2019. Extended Semi-Supervised Learning GAN for Hyperspectral Imagery Classification. In 2019 13th International Conference on Signal Processing and Communication Systems (ICSPCS), IEEE, Gold Coast, Australia, 1–6. DOI: https://doi.org/10.1109/ICSPCS47537.2019.9008719Google ScholarCross Ref
- Shuai Zhang, Xiuqing Mao, Lei Sun, and Yu Yang. 2022. EEG data augmentation for Personal Identification Using SF-GAN. In 2022 3rd International Conference on Computer Vision, Image and Deep Learning & International Conference on Computer Engineering and Applications (CVIDL & ICCEA), IEEE, Changchun, China, 1–6. DOI: https://doi.org/10.1109/CVIDLICCEA56201.2022.9824276Google ScholarCross Ref
- Chuanli Jiang, Jianyun Zhao, Yuanyuan Ding, and Guorong Li. 2023. Vis–NIR Spectroscopy Combined with GAN Data Augmentation for Predicting Soil Nutrients in Degraded Alpine Meadows on the Qinghai–Tibet Plateau. Sensors 23, 7 (April 2023), 3686. DOI: https://doi.org/10.3390/s23073686Google ScholarCross Ref
- Nicolas Audebert, Bertrand Le Saux, and Sebastien Lefevre. 2018. Generative Adversarial Networks for Realistic Synthesis of Hyperspectral Samples. In IGARSS 2018 - 2018 IEEE International Geoscience and Remote Sensing Symposium, IEEE, Valencia, 4359–4362. DOI: https://doi.org/10.1109/IGARSS.2018.8518321Google ScholarCross Ref
- Junho Kim, Yunjey Choi, and Youngjung Uh. 2021. Feature Statistics Mixing Regularization for Generative Adversarial Networks. (2021). DOI: https://doi.org/10.48550/ARXIV.2112.04120Google ScholarCross Ref
- Esteban Piacentino, Alvaro Guarner, and Cecilio Angulo. 2021. Generating Synthetic ECGs Using GANs for Anonymizing Healthcare Data. Electronics 10, 4 (February 2021), 389. DOI: https://doi.org/10.3390/electronics10040389Google ScholarCross Ref
- Takeru Miyato, Toshiki Kataoka, Masanori Koyama, and Yuichi Yoshida. 2018. Spectral Normalization for Generative Adversarial Networks. (2018). DOI: https://doi.org/10.48550/ARXIV.1802.05957Google ScholarCross Ref
- Yuichi Yoshida and Takeru Miyato. 2017. Spectral Norm Regularization for Improving the Generalizability of Deep Learning. (2017). DOI: https://doi.org/10.48550/ARXIV.1705.10941Google ScholarCross Ref
- Lilian Weng. 2019. From GAN to WGAN. (2019). DOI: https://doi.org/10.48550/ARXIV.1904.08994Google ScholarCross Ref
- Dingfan Chen, Tribhuvanesh Orekondy, and Mario Fritz. 2020. GS-WGAN: A Gradient-Sanitized Approach for Learning Differentially Private Generators. (2020). DOI: https://doi.org/10.48550/ARXIV.2006.08265Google ScholarCross Ref
- Changsheng Zhou, Jiangshe Zhang, and Junmin Liu. 2018. Lp-WGAN: Using Lp-norm normalization to stabilize Wasserstein generative adversarial networks. Knowledge-Based Systems 161, (December 2018), 415–424. DOI: https://doi.org/10.1016/j.knosys.2018.08.004Google ScholarCross Ref
- Qimin Jin, Rongheng Lin, and Fangchun Yang. 2020. E-WACGAN: Enhanced Generative Model of Signaling Data Based on WGAN-GP and ACGAN. IEEE Systems Journal 14, 3 (September 2020), 3289–3300. DOI: https://doi.org/10.1109/JSYST.2019.2935457Google ScholarCross Ref
- Chunling Cao, Tianli Wang, Maofang Gao, Yang Li, Dandan Li, and Huijie Zhang. 2021. Hyperspectral inversion of nitrogen content in maize leaves based on different dimensionality reduction algorithms. Computers and Electronics in Agriculture 190, (November 2021), 106461. DOI: https://doi.org/10.1016/j.compag.2021.106461Google ScholarDigital Library
- Yuki Enokiya, Yutaro Iwamoto, Yen-Wei Chen, and Xian-Hua Han, "Automatic Liver Segmentation Using U-Net with Wasserstein GANs," Journal of Image and Graphics, Vol. 7, No. 3, pp. 94-101, September 2019. doi: 10.18178/joig.7.3.94-101Google ScholarCross Ref
- Yusuke Ikeda, Keisuke Doman, Yoshito Mekada, and Shigeru Nawano, "Lesion Image Generation Using Conditional GAN for Metastatic Liver Cancer Detection," Journal of Image and Graphics, Vol. 9, No. 1, pp. 27-30, March 2021. doi: 10.18178/joig.9.1.27-30Google ScholarCross Ref
- Takato Sakai, Masataka Seo, Naoki Matsushiro, and Yen-Wei Chen, "Simulation of Facial Palsy Using Cycle GAN with Skip-Layer Excitation Module and Self-Supervised Discriminator," Journal of Image and Graphics, Vol. 11, No. 2, pp. 132-139, June 2023.Google Scholar
- Farnoush Zohourian and Josef Pauli, "Coarse-to-Fine Semantic Road Segmentation Using Super-Pixel Data Model and Semi-Supervised Modified CycleGAN," Journal of Image and Graphics, Vol. 10, No. 4, pp. 132-144, December 2022.Google Scholar
Index Terms
- Feature Mixture Generative Adversarial Network for Data Augmentation on Small Sample Hyperspectral Data
Recommendations
Population-scale Genomic Data Augmentation Based on Conditional Generative Adversarial Networks
BCB '20: Proceedings of the 11th ACM International Conference on Bioinformatics, Computational Biology and Health InformaticsAlthough next generation sequencing technologies have made it possible to quickly generate a large collection of sequences, current genomic data still suffer from small data sizes, imbalances, and biases due to various factors including disease rareness,...
Diverse sample generation with multi-branch conditional generative adversarial network for remote sensing objects detection
AbstractThe remote sensing data is difficult to collect and lack of diversity, which extremely limits the performance of object detection on remote sensing images. In this paper, a multi-branch conditional generative adversarial network (MCGAN)...
Features kept generative adversarial network data augmentation strategy for hyperspectral image classification
Highlights- A novel data augmentation strategy called feature-preserving generative adversarial network data augmentation (FPGANDA) is designed to alleviate the small-...
AbstractIn recent years, significant breakthroughs have been achieved in hyperspectral image (HSI) processing using deep learning techniques, including classification, object detection, and anomaly detection. However, the practical application ...
Comments