Computer Science

Computer Science ›› 2020, Vol. 47 ›› Issue (11A): 209-214.doi: 10.11896/jsjkx.200100090

• Computer Graphics & Multimedia • Previous Articles     Next Articles

Neural Style Transfer Method Based on Laplace Operator to Suppress Artifacts

ZHANG Mei-yu1, LIU Yue-hui1, QIN Xu-jia1, WU Liang-wu2   

  1. 1 College of Computer Science and Technology,Zhejiang University of Technology,Hangzhou 310023,China
    2 Dalian Institute of Test and Control Technology,Dalian,Liaoning 116013,China
  • Online:2020-11-15 Published:2020-11-17
  • About author:ZHANG Mei-yu,born in 1965,professor.Her research interests include ima-ge analysis and image processing.
    QIN Xu-jia,born in 1968,Ph.D,professor,Ph.D candidate supervisor,is a member of China Computer Federation.His main research interests include computer graphics,image processing and data visualization.
  • Supported by:
    This work was supported the National Natural Science Foundation of China (61672463) and Natural Science Foundation of Zhejiang Pvovince,China (LY20F020025,LY18F020035).

PDF (PC)

684

Abstract: In image neural style transfer technology,most algorithms have artifacts that affect visual effects:checkerboard effects and textures that affect the semantic content of the original image.In this paper,an image style transfer method based on Laplacian suppression artifacts is proposed.Firstly,a transformation network for real-time neural style transfer is redesigned using hole convolution and 1×1 convolution filter kernels.Then,the transformed result is input to VGG for feature map detection,and the multi-layer feature and the original VGG feature are extracted and filtered by Laplace operator to calculate the L1 error.Constrain image changes to suppress artifacts.In the final encoder stage,the image content is modified using an encoder with added dropout.While deepening the network,the model size was controlled by 1×1 convolution filter kernels,which reduced the model size about 6%.Finally,experiments show that the results of this method are better than traditional methods in suppressing artifacts,and can produce images with good visual effects.

Key words: Convolutional Neural Networks, Gram matrix, Laplacian operator, Residual, Style transfer

CLC Number: 

  • TP391
[1] GATYS L A,ECKER A S,BETHGE M.A Neural Algorithm of Artistic Style[J].arXiv:1508.06576,2015.
[2] SIMONYAN K,ZISSERMAN A.Very Deep Convolutional Networks for Large-Scale Image Recognition[C]//International Conference on Learning Representations.2015.
[3] SRIVASTAVA N,HINTON G E,KRIZHEVSKY A,et al.Dropout:a simple way to prevent neural networks from overfitting[J].Journal of Machine Learning Research,2014,15(1):1929-1958.
[4] MISRA D.Mish:A Self Regularized Non-Monotonic Neural Ac-tivation Function[J].arXiv:1908.08681,2019
[5] JOHNSON J,ALAHI A,FEIFEI L.Perceptual Losses for Real-Time Style Transfer and Super-Resolution[C]//European Conference on Computer Vision.2016:694-711.
[6] LONG J,SHELHAMER E,DARRELL T.Fully convolutional networks for semantic segmentation[C]//Computer Vision and Pattern Recognition.2015:3431-3440.
[7] LUAN F,PARIS S,SHECHTMAN E,et al.Deep PainterlyHarmonization[J].Computer Graphics Forum,2018,37(4):95-106.
[8] JING Y,LIU Y,YANG Y,et al.Stroke Controllable Fast Style Transfer with Adaptive Receptive Fields[C]//European Confe-rence on Computer Vision.2018:244-260.
[9] SHEN F,YAN S,ZENG G.Neural Style Transfer via MetaNetworks[C]//Computer Vision and Pattern Recognition.2018:8061-8069.
[10] RUSSAKOVSKY O,DENG J,SU H,et al.ImageNet LargeScale Visual Recognition Challenge[J].International Journal of Computer Vision,2015,115(3):211-252.
[11] HE K,ZHANG X,REN S,et al.Deep Residual Learning forImage Recognition[C]//Computer Vision and Pattern Recognition.2016:770-778.
[12] WANG M,WEI Q.Research on the algorithm of LaplacianSharpening[C]// Proceedings of the 16th Annual Conference on Information Theory of the Chinese Institute of Electronics.2009:464-468.
[13] JIA H,ZHANG Y,LONG G,et al.REsearch on Laplace Image Enhancement Algorithm Optimization Based on OpenCL[J].Chinese Journal of Computers,2012,39(5):271.
[14] YU F,KOLTUN V.Multi-Scale Context Aggregation by DilatedConvolutions[C]//International Conference on Learning Representations.2016.
[15] LIN M,CHEN Q,YAN S.Network In Network[C]//International Conference on Learning Representations.2014.
[16] IOFFE S,SZEGEDY C.Batch Normalization:Accelerating Deep Network Training by Reducing Internal Covariate Shift[C]//International Conference on Machine Learning.2015:448-456.
[17] HINTON G E,SRIVASTAVA N,KRIZHEVSKY A,et al.Improving neural networks by preventing co-adaptation of feature detectors[J].arXiv:1207.0580,2012.
[18] STEINER B,DEVITO Z,CHINTALA S,et al.PyTorch:An Imperative Style,High-Performance Deep Learning Library[C] //Neural Information Processing Systems.2019:8024-8035.
[19] KINGMA D P,BA J.Adam:A Method for Stochastic Optimization[C]//International Conference on Learning Representations.2015.
[1] WANG Xin-tong, WANG Xuan, SUN Zhi-xin. Network Traffic Anomaly Detection Method Based on Multi-scale Memory Residual Network [J]. Computer Science, 2022, 49(8): 314-322.
[2] ZHU Cheng-zhang, HUANG Jia-er, XIAO Ya-long, WANG Han, ZOU Bei-ji. Deep Hash Retrieval Algorithm for Medical Images Based on Attention Mechanism [J]. Computer Science, 2022, 49(8): 113-119.
[3] ZHANG Ying-tao, ZHANG Jie, ZHANG Rui, ZHANG Wen-qiang. Photorealistic Style Transfer Guided by Global Information [J]. Computer Science, 2022, 49(7): 100-105.
[4] GAO Rong-hua, BAI Qiang, WANG Rong, WU Hua-rui, SUN Xiang. Multi-tree Network Multi-crop Early Disease Recognition Method Based on Improved Attention Mechanism [J]. Computer Science, 2022, 49(6A): 363-369.
[5] ZHANG Jia-hao, LIU Feng, QI Jia-yin. Lightweight Micro-expression Recognition Architecture Based on Bottleneck Transformer [J]. Computer Science, 2022, 49(6A): 370-377.
[6] WANG Jian-ming, CHEN Xiang-yu, YANG Zi-zhong, SHI Chen-yang, ZHANG Yu-hang, QIAN Zheng-kun. Influence of Different Data Augmentation Methods on Model Recognition Accuracy [J]. Computer Science, 2022, 49(6A): 418-423.
[7] SUN Jie-qi, LI Ya-feng, ZHANG Wen-bo, LIU Peng-hui. Dual-field Feature Fusion Deep Convolutional Neural Network Based on Discrete Wavelet Transformation [J]. Computer Science, 2022, 49(6A): 434-440.
[8] LIU Chang, WEI Wei-min, MENG Fan-xing, CAI Zhi. Research Progress on Speech Style Transfer [J]. Computer Science, 2022, 49(6A): 301-308.
[9] YANG Yue, FENG Tao, LIANG Hong, YANG Yang. Image Arbitrary Style Transfer via Criss-cross Attention [J]. Computer Science, 2022, 49(6A): 345-352.
[10] HAN Hong-qi, RAN Ya-xin, ZHANG Yun-liang, GUI Jie, GAO Xiong, YI Meng-lin. Study on Cross-media Information Retrieval Based on Common Subspace Classification Learning [J]. Computer Science, 2022, 49(5): 33-42.
[11] GAO Yuan-hao, LUO Xiao-qing, ZHANG Zhan-cheng. Infrared and Visible Image Fusion Based on Feature Separation [J]. Computer Science, 2022, 49(5): 58-63.
[12] ZHAO Ren-xing, XU Pin-jie, LIU Yao. ECG-based Atrial Fibrillation Detection Based on Deep Convolutional Residual Neural Network [J]. Computer Science, 2022, 49(5): 186-193.
[13] QU Zhong, CHEN Wen. Concrete Pavement Crack Detection Based on Dilated Convolution and Multi-features Fusion [J]. Computer Science, 2022, 49(3): 192-196.
[14] ZHANG Lyu, ZHOU Bo-wen, WU Liang-hong. SSD Network Based on Improved Convolutional Attention Module and Residual Structure [J]. Computer Science, 2022, 49(3): 211-217.
[15] CHEN Zhi-yi, SUI Jie. DeepFM and Convolutional Neural Networks Ensembles for Multimodal Rumor Detection [J]. Computer Science, 2022, 49(1): 101-107.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
No Suggested Reading articles found!
渝ICP备16013121号-4
Add:18# Honghu West Rd., Yubei ,Chongqing,China    Post Code: 401121
Tel: 023-63500828/023-67039612/023-67039623
E-mail: jsjkx12@163.com
Powered by Beijing Magtech Co., Ltd.