Data Augmentation Method for Sweet Cherries Based on Improved Generative Adversarial Network

HAN Xiang, LI Yuqiang, LI Yuqiang, GAO Ang, MA Jingyi, GONG Qingfu, SONG Yuepeng

Abstract

To address the class imbalance in sweet cherry data, a novel image enhancement method based on sweet cherry generative adversarial network, SCGAN was proposed. The generator incorporated multi-scale residual blocks (MSRB) and the convolutional block attention module (CBAM), enhancing the model’s feature representation and the quality of generated images. These blocks captured features at various scales, and CBAM focused on channel and spatial information, improving image quality. In the discriminator, spectral normalization and the Wasserstein distance with a gradient penalty loss function were applied. This combination controled the discriminator’s power, prevented overfitting, and boosted training stability and speed. Experimental results showed that SCGAN produced higher quality defective sweet cherry images compared with traditional GANs, with Fréchet inception distance (FID) scores of 64.36 and 59.97 for two types of defects. After data augmentation with SCGAN, classification accuracy for VGG19 and MobileNetV3 was increased by 16.44 percentage points and 13.94 percentage points, respectively. The data augmentation method presented held significant potential in addressing data imbalance issues within the agricultural and food sectors. It not only improved the generalization capability of models but also provided a more reliable data foundation for practical applications. Through this approach, it was possible to more effectively tackle long-tail class imbalance issues, which enhanced the accuracy and efficiency of agricultural and food detection systems.

 

Keywords: sweet cherry images, data generation;generative adversarial network, long-tailed class imbalance

 

Download Full Text:

PDF


References


PENG Mingxia. Research on the method of identifying weeds in cotton seedling stage based on image processing [D] . Wuhan: Huazhong Agricultural University, 2023. (in Chinese)

CENGGORO T W. Deep learning for imbalance data classification using class expert generative adversarial network [ J ]. Procedia Computer Science, 2018, 135: 60 -67.

CHEN H, LI C, YANG W, et al. Deep balanced cascade forest: an novel fault diagnosis method for data imbalance [ J ]. ISA Transactions, 2022, 126: 428 -439.

SUN C, GUI H, ZHOU W, et al. Epileptic seizure detection with EEG textural features and imbalanced classification based on EasyEnsemble learning [j]. International Journal of Neural Systems, 2019, 29( 10) : 1950021.

LIU J, XU K, CAI B, et al. Fault prediction of on-board train control equipment using a CGAN - Enhanced XGBoost method with unbalanced samples[ J ]. Machines, 2023 , 11(1): 114.

WANG H, HU T, XU Z, et al. A radar waveform recognition method based on ambiguity function generative adversarial network data enhancement under the condition of small samples [ J] . IET Radar, Sonar & Navigation, 2023 , 17( 1 ) : 86 -98.

WANG Jun, GAO Fang, SHENG Haixian, et al. Image generation algorithm based on Swin Transformer to generate adversarial networks[ J ]. Journal of Chinese Computer Systems,2024, 45( 1 ) : 241 -248. (in Chinese)

SHORTEN C, KHOSHGOFTAAR T M. A survey on image data augmentation for deep learning [j]. Journal of Big Data, 2019, 6(1): 1 -48.

WEI Tingting, GE Xiaoyue, XIONG Juntao. Hierarchical multi-label classification of agricultural pest and disease interrogative questions [J] . Transactions of the Chinese Society for Agricultural Machinery,2024,55( 1 ) :263 -269,435. ( in Chinese)

GEETHARAMANI G, PANDIAN A. Identification of plant leaf diseases using a nine-layer deep convolutional neural network [J] . Computers & Electrical Engineering, 2019, 76: 323 -338.

OFUSA K, MIYAZAKI T, SUGAYA Y, et al. Glyph-based data augmentation for accurate kanji character recognition [С] // 2017 14th IAPR International Conference on Document Analysis and Recognition (ICDAR). IEEE, 2017, 1 : 597 -602.

ZHANG Y D, DONG Z, CHEN X, et al. Image based fruit category classification by 13-layer deep convolutional neural network and data augmentation[ J ]. Multimedia Tools Applications, 2019, 78; 3613 -3632.

ZHOU G, CHEN Y, CHIEN C. On the analysis of data augmentation methods for spectral imaged based heart sound classification using convolutional neural networks [J]. BMC Medical Informatics Decision Making, 2022, 22( 1 ) • 226.

CHEN Foji, ZHU Feng, WU Qingxiao, et al. A survey about image generation with generative adversarial nets [ J] . Chinese Journal of Computers, 2021 , 44(2) : 347 -369. ( in Chinese)

KHOSLA C, SAINI В S. Enhancing performance of deep learning models with different data augmentation techniques; a survey [С] //2020 International Conference on Intelligent Engineering and Management (ICIEM). IEEE, 2020; 79 -85.

PEREZ L, W ANG J. The effectiveness of data augmentation in image classification using deep learning [J] . arXiv preprint arXiv, 2017: 1712.04621.

LI Yunhong, ZHANG Leitao, LI Limin, et al. Image recognition of apple leaf disease based on CycleGAN — IA method and M — ConvNext network [ J ]. Transactions of the Chinese Society for Agricultural Machinery, 2024, 55 ( 4 ) : 204 - 212. (in Chinese)

GOODFELLOW I, POUGET-ABADIE J, MIRZA M, et al. Generative adversarial nets [C]//Advances in Neural Information Processing Systems, 2014; 2672 -2680.

ZHANG Linxuan, BA Yintana, ZENG Qingsong. Grape disease detection method based on StvleGAN2 — ADA and improved YOLO v7 [ J ]. Transactions of the Chinese Society for Agricultural Machinery, 2024,55 ( 1 ) : 241 -252. (in Chinese)

LIU Yixuan, GE Guangying, QI Zhenling, et al. Research on embroidery image restoration based on improved deep convolutional generative adversarial network[ J ]. Laser & Optoelectronics Progress, 2023, 60(20) ; 68 -78. (in Chinese)

RADFORD A, METZ L, CHINTALA S. Unsupervised representation learning with deep convolutional generative adversarial networks[ J ]. arXiv preprint arXiv, 2015: 1511.06434.

TAO Zhaosheng, GONG Baoguo, LI Qingping, et al. Wheat image inpainting based on residual networks and feature fusion [ J ]. Transactions of the Chinese Society for Agricultural Machinery ,2023 ,54( 3 ) : 318 -327. (in Chinese)

LI Jiesong, LIU Tao, WU Xing. Spectral normalization CycleGAN for bearing fault transfer diagnosis [J ]. Journal of Vibration and Shock, 2023, 42(24) : 282 -289. (in Chinese)

LI Qingxu, W ANG Qiaohua, MA Meihu. Poultry egg image data generating based on generative adversarial network [ J], Transactions of the Chinese Society for Agricultural Machinery,2021 ,52(2) :236 -245. (in Chinese)


Refbacks

  • There are currently no refbacks.