Abstract: Accurate recognition of modern and traditional porcelain styles is a challenging issue in cantonese porcelain management due to the large variety and complex elements and patterns. We propose a hybrid system with porcelain style identification and image recreation modules. In the identification module, prediction of an unknown porcelain sample is obtained by logistic regression of ensembled neural networks of top-ranked design signatures, which are obtained by discriminative analysis and transformed features in principal components. The synthesis module is developed based on a conditional generative adversarial network, which enables users to provide a designed mask with porcelain elements to generate synthesized images of cantonese porcelain. Experimental results of 603 cantonese porcelain images demonstrate that the proposed model outperforms other methods relative to precision, recall, area under curve of receiver operating characteristic, and confusion matrix. Case studies on image creation indicate that the proposed system has the potential to engage the community in understanding cantonese porcelain and promote this intangible cultural heritage.

基于机器学习的广彩瓷图案生成系统

[1]Bao H, Liang Y, Liu HZ, et al., 2010. A novel algorithm for extraction of the scripts part in traditional Chinese painting images. Proc 2^nd Int Conf on Software Technology and Engineering, p.V2-26-V2-30.

[2]Buades A, Coll B, Morel JM, 2005. A non-local algorithm for image denoising. IEEE Computer Society Conf on Computer Vision and Pattern Recognition, p.60-65.

[3]Chen KH, 2019. Image Operations with cGAN. http://www.k4ai.com/imageops/index.html

[4]China Intangible Cultural Heritage Network, 2008. Cantonese Porcelain Inheritance Project. http://www.ihchina.cn/project_details/14453/ [Accessed on July 16, 2019] (in Chinese).

[5]Cochran WG, 1954. Some methods for strengthening the common χ² tests. Int Biom Soc, 10(4):417-451.

[6]Dirvanauskas D, Maskeliūnas R, Raudonis V, et al., 2019. HEMIGEN: human embryo image generator based on generative adversarial networks. Sensors, 19(16):3578.

[7]Efros AA, Freeman WT, 2001. Image quilting for texture synthesis and transfer. Proc 28^th Annual Conf on Computer Graphics and Interactive Techniques, p.341-346.

[8]El Hattami A, Pierre-Doray É, Barsalou Y, 2019. Background removal using U-net, GAN and image matting. https://github.com/eti-p-doray/unet-gan-matting [Accessed on July 14, 2019].

[12]Emami H, Dong M, Nejad-Davarani SP, et al., 2018. Generating synthetic CTs from magnetic resonance images using generative adversarial networks. Med Phys, 45(8): 3627-3636.

[9]Goodfellow IJ, Pouget-Abadie J, Mirza M, et al., 2014. Generative adversarial nets. Proc 27^th Int Conf on Neural Information Processing Systems, p.2672-2680.

[10]Hertzmann A, Jacobs CE, Oliver N, et al., 2001. Image analogies. Proc 28^th Annual Conf on Computer Graphics and Interactive Techniques, p.327-340.

[11]Iizuka S, Simo-Serra E, Ishikawa H, 2016. Let there be color!: joint end-to-end learning of global and local image priors for automatic image colorization with simultaneous classification. ACM Trans Graph, 35(4), Article 110.

[12]Isola P, Zhu JY, Zhou TH, et al., 2017. Image-to-image translation with conditional adversarial networks. IEEE Conf on Computer Vision and Pattern Recognition, p.1125-1134.

[13]Ji Y, Tan P, Chen SC, et al., 2019. Kansei engineering for E-commerce Cantonese porcelain selection in China. 21^st Int Conf on Human-Computer Interaction, p.463-474.

[14]Jiang SQ, Huang QM, Ye QX, et al., 2006. An effective method to detect and categorize digitized traditional Chinese paintings. Patt Recogn Lett, 27(7):734-746.

[15]Kira K, Rendell LA, 1992. A practical approach to feature selection. Machine Learning Proc, p.249-256.

[16]Kurin R, 2004. Safeguarding intangible cultural heritage in the 2003 UNESCO convention: a critical appraisal. Museum Int, 56(1-2):66-77.

[17]Larsson G, Maire M, Shakhnarovich G, 2016. Learning representations for automatic colorization. Proc 14^th European Conf on Computer Vision, p.577-593.

[18]Lecoutre A, Négrevergne B, Yger F, 2017. Recognizing art style automatically in painting with deep learning. Proc 9^th Asian Conf on Machine Learning, p.327-342.

[19]Li WB, Zhang PC, Zhang L, et al., 2019. Object-driven text-to-image synthesis via adversarial training. https://arxiv.org/abs/1902.10740

[20]Lin TY, Maire M, Belongie S, et al., 2014. Microsoft COCO: common objects in context. Proc 13^th European Conf on Computer Vision, p.740-755.

[21]Liu YF, Qin ZC, Wan T, et al., 2018. Auto-painter: cartoon image generation from sketch by using conditional Wasserstein generative adversarial networks. Neurocomputing, 311:78-87.

[22]Lowenthal D, 2005. Natural and cultural heritage. Int J Herit Stud, 11(1):81-92.

[28]Mao XF, Wang SH, Zheng LY, et al., 2018. Semantic invariant cross-domain image generation with generative adversarial networks. Neurocomputing, 293:55-63.

[23]Meng QY, Zhang HH, Zhou MQ, et al., 2018. The classification of traditional Chinese painting based on CNN. Proc 4^th Int Conf on Cloud Computing and Security, p.232-241.

[24]Połap D, Woźniak M, Wei W, et al., 2018. Multi-threaded learning control mechanism for neural networks. Fut Gener Comput Syst, 87:16-34.

[25]Quinlan JR, 1986. Induction of decision trees. Mach Learn, 1(1):81-106.

[26]Smith L, Akagawa N, 2008. Intangible Heritage. Routledge, London, UK.

[27]Yu L, Liu H, 2003. Feature selection for high-dimensional data: a fast correlation-based filter solution. Proc 20^th Int Conf on Machine Learning, p.856-863.

[28]Zhang R, Isola P, Efros AA, 2016. Colorful image colorization. Proc 14^th European Conf on Computer Vision, p.649-666.

[29]Zhu JY, Park T, Isola P, et al., 2017. Unpaired image-to-image translation using cycle-consistent adversarial networks. IEEE Int Conf on Computer Vision, p.2223-2232.

[30]Zujovic J, Gandy L, Friedman S, et al., 2009. Classifying paintings by artistic genre: an analysis of features & classifiers. IEEE Int Workshop on Multimedia Signal Processing, p.1-5.