“…There are many endeavors carried out on pattern synthesis, or broadly speaking, image synthesis, e.g., using Markov random field model to approximate an image to different levels by tuning the neighborhood window (Efros and Leung, 2002), using convolutional neural network (CNN) to learn image features across the layers with texture representations that are increasingly explicit (Gatys et al , 2015), integrating depth data captured by a Microsoft Kinect device into the image input to a CNN to achieve 3D textile reconstruction (Hu et al , 2017). While the digitalized pattern creation is primarily inspired by the visualization of mathematical models, among which there are the famous fractal algorithms including iterated function systems (Guo and Song, 2010; Meng et al , 2013), the Julia set (Ma and Li, 2006; Jia and Zhang, 2015) and the L systems (Meng and Li, 2011; Zhang and Wang, 2011). Another important model for visualization is the weak chaos system (Zhang, 2004) upon which there are the derived mechanisms like quasi-regular system (Liu and Zhang, 2009), uniform stochastic web (Luo and Hong, 2014) and Hamiltonian function (Suyi and Leduo, 2009).…”