基于卷积神经网络的肺表面纹理合成
|
摘要
以肺表面纹理为研究对象,提出一种基于卷积神经网络的纹理合成算法。算法以训练好的VGG-19模型为基础,用Gram矩阵来表示纹理的局部结构特征,引入一种结构约束关系来捕捉纹理的全局结构特征,并在网络的高层加入马尔科夫随机场以提高算法效率。实验结果表明该算法可以较好地合成具有局部结构特征以及非局部结构特征的肺表面纹理,并且对比现有相似的算法效率有了明显提升。
Taking the surface texture of lung as the research object,this paper proposes a texture synthesis algorithm based on convolutional neural network. Based on the trained VGG-19 model,the algorithm uses the Gram matrix to represent the local structural features of the texture,introduces a structural constraint relationship to capture the global structural features of the texture,and joins the Markov random field at the upper level of the network to improve algorithm efficiency. The experimental results show that the algorithm can synthesize the lung surface texture with local structural features and non-local structural features,and the efficiency of the existing similar algorithms is significantly improved.
Taking the surface texture of lung as the research object,this paper proposes a texture synthesis algorithm based on convolutional neural network. Based on the trained VGG-19 model,the algorithm uses the Gram matrix to represent the local structural features of the texture,introduces a structural constraint relationship to capture the global structural features of the texture,and joins the Markov random field at the upper level of the network to improve algorithm efficiency. The experimental results show that the algorithm can synthesize the lung surface texture with local structural features and non-local structural features,and the efficiency of the existing similar algorithms is significantly improved.
引文
[1]王栋.纹理合成技术及其在虚拟手术中的应用[D].国防科学技术大学,2008.
[2]Gatys L,Ecker A S,Bethge M.Texture Synthesis Using Convolutional Neural Networks[C].Advances in Neural Information Processing Systems.2015:262-270.
[3]Sendik O,Cohen-Or D.Deep Correlations for Texture Synthesis[J].ACM Transactions on Graphics(TOG),2017,36(5):161.
[4]Simonyan K,Zisserman A.Very Deep Convolutional Networks for Large-scale Image Recognition[J].ar Xiv preprint ar Xiv:1409.1556,2014.
[5]Li C,Wand M.Combining Markov Random Fields and Convolutional Neural Networks for Image Synthesis[J].2016:2479-2486.
[6]Johnson J,Alahi A,Li F F.Perceptual Losses for Real-Time Style Transfer and Super-Resolution[C]European Conference on Computer Vision.Springer,Cham,2016:694-711.
[7]Byrd R H,Lu P,Nocedal J,et al.A limited Memory Algorithm for Bound Constrained Optimization[J].Siam Journal on Scientific Computing,1995,16(5):1190-1208.
[8]Vedaldi A,Lenc K.Mat ConvNet:Convolutional Neural Networks for MATLAB[J].2014:689-692.
[9]成诚.纹理合成算法研究[D].合肥工业大学,2011.DOI:10.7666/d.d142972.
[10]Yang C,Lu X,Lin Z,et al.High-Resolution Image Inpainting using Multi-Scale Neural Patch Synthesis[J].2016:4076-4084.
[2]Gatys L,Ecker A S,Bethge M.Texture Synthesis Using Convolutional Neural Networks[C].Advances in Neural Information Processing Systems.2015:262-270.
[3]Sendik O,Cohen-Or D.Deep Correlations for Texture Synthesis[J].ACM Transactions on Graphics(TOG),2017,36(5):161.
[4]Simonyan K,Zisserman A.Very Deep Convolutional Networks for Large-scale Image Recognition[J].ar Xiv preprint ar Xiv:1409.1556,2014.
[5]Li C,Wand M.Combining Markov Random Fields and Convolutional Neural Networks for Image Synthesis[J].2016:2479-2486.
[6]Johnson J,Alahi A,Li F F.Perceptual Losses for Real-Time Style Transfer and Super-Resolution[C]European Conference on Computer Vision.Springer,Cham,2016:694-711.
[7]Byrd R H,Lu P,Nocedal J,et al.A limited Memory Algorithm for Bound Constrained Optimization[J].Siam Journal on Scientific Computing,1995,16(5):1190-1208.
[8]Vedaldi A,Lenc K.Mat ConvNet:Convolutional Neural Networks for MATLAB[J].2014:689-692.
[9]成诚.纹理合成算法研究[D].合肥工业大学,2011.DOI:10.7666/d.d142972.
[10]Yang C,Lu X,Lin Z,et al.High-Resolution Image Inpainting using Multi-Scale Neural Patch Synthesis[J].2016:4076-4084.