Accurately quantifying the insect-damaged leaf areas is a core component in applied reaserch such as pest control, plant resistance breeding, nad genetic screening. However, leaves in natural environments are susceptible to insect feeding, resulting in irregular damage. To address the problem of insect-damaged leaf areas reconstruction, this paper proposed an image completion method based on Pix2Pix (image-to-image translation with conditional adversarial networks). Firstly, a residual structure and a convolutional block attention module (CBAM) were introduced into the generator to enhance the network's feature extraction and representation capabilities. Subsequently, the original discriminator was replaced with a WGAN-GP discriminator incorporated with spectral normalization (SN) to improve the stability of model training. The results indicated that, compared with the original model, the improved algorithm achieved an increase of 4.4% in structural similarity (SSIM), an increase of 13.8% in peak signal-to-noise ratio (PSNR), and a reduction of 53.7% in learned perceptual image patch similarity (LPIPS). This method can achieve restoration results with clearer leaf vein textures and better preservation of semantic structural consistency at leaf margins, providing technical support for the extraction of plant leaf phenotypic parameters.
LIY Q, WANGZ H.Leaf morphological traits:Ecological function,geographic distribution and drivers[J].Chinese Journal of Plant Ecology,2021,45(10):1154-1172.
[3]
JINX, YANGW, DOONANJ H,et al.Crop phenotyping studies with application to crop monitoring[J].The Crop Journal,2022,10(5):1221-1223.
LIY L, GAOY, YANJ L,et al.Image inpainting methods based on deep neural networks:A review[J].Chinese Journal of Computers,2021,44(11):2295-2316.
[6]
王阿川.基于变分PDE的单板缺陷图像检测及修补关键技术研究[D].哈尔滨:东北林业大学,2011.
[7]
WANGA C.Research on the key technology of veneer defect image detection and patching based on variational PDE[D].Harbin:Northeast Forestry University,2011.
[8]
PATHAKD, KRAHENBUHLP, DONAHUEJ,et al.Context encoders:Feature learning by inpainting[C]//Proceedings of the IEEE conference on computer vision and pattern recognition,2016:2536-2544.
[9]
IIZUKAS, SIMO-SERRAE, ISHIKAWAH.Globally and locally consistent image completion[J].ACM Transactions on Graphics,2017,36(4):1-14.
[10]
GOODFELLOWI, POUGET-ABADIEJ, MIRZAM,et al.Generative adversarial networks[J].Communications of the ACM,2020,63(11):139-144.
LOUY Y, ZHANGD Y, GEY L,et al.Image inpainting of veneer with improved learnable bidirectional attention maps[J].Forest Engineering,2023,39(2):132-138.
[15]
ISOLAP, ZHUJ Y, ZHOUT,et al.Image-to-image translation with conditional adversarial networks[C]//Proceedings of the IEEE conference on computer vision and pattern recognition,2017:1125-1134.
[16]
RONNEBERGERO, FISCHERP, BROXT.U-net:Convolutional networks for biomedical image segmentation[C]//International Conference on Medical Image Computing and Computer-Assisted Intervention.Cham:Springer International Publishing,2015:234-241.
[17]
WOO S, PARKJ, LEEJ Y,et al.Cbam:Convolutional block attention module[C]//2018 European Conference on Computer Vision (ECCV).Springer,Cham,2018:3-19.
[18]
GULRAJANII, AHMEDF, ARJOVSKYM,et al.Improved training of Wasserstein GANs[J].Advances in Neural Information Processing Systems,2017,30:1-11.
[19]
HUJ, SHENL, SUNG.Squeeze-and-excitation networks[C]//Proceedings of the IEEE conference on computer vision and pattern recognition,2018:7132-7141.
JOHNSONJ, ALAHIA, LIF F.Perceptual losses for real-time style transfer and super-resolution[C]//European Conference on Computer Vision.Cham:Springer International Publishing,2016:694-711.
[22]
WENL, LIX, LIX,et al.A new transfer learning based on VGG-19 network for fault diagnosis[C]//2019 IEEE 23rd International conference on computer supported cooperative work in design(CSCWD).IEEE,2019:205-209.
[23]
HORÉA, ZIOUD.Image quality metrics:PSNR vs.SSIM[C]//2010 20th international conference on pattern recognition.IEEE,2010:2366-2369.
[24]
SNELLJ, RIDGEWAYK, LIAOR,et al.Learning to generate images with perceptual similarity metrics[C]//2017 IEEE international conference on image processing(ICIP).IEEE,2017:4277-4281.
[25]
PASZKEA, GROSSS, MASSAF,et al.Pytorch:An imperative style,high-performance deep learning library[J].Advances in neural information processing systems,2019:32.
[26]
RADFORDA, METZL, CHINTALAS.Unsupervised representation learning with deep convolutional generative adversarial networks[J].arXiv preprint arXiv:1511. 06434,2015.
[27]
MAOX, LIQ, XIEH,et al.Least squares generative adversarial networks[C]//Proceedings of the IEEE international conference on computer vision,2017:2794-2802.
LIP Y, ZHANGY L, ZHANGY B,et al.Face image super-resolution reconstruction based on adaptive convolution and joint loss function[J].Science Technology and Engineering,2025,25(6):2442-2452.
LUP, SUNT W, CHENM,et al.Automatic extraction of phenotypic parameters from Anthurium andraeanum Linden based on YOLOv8 and CycleGAN[J].Transactions of the Chinese Society for Agricultural Machinery,2024,55(11):154-159,319.