Data-driven main color map feature learning, design and simulation for smart ethnic cloth

How to protect and develop traditional ethnic cloth culture are the key problems of current societal interest. Through earlier research of smart techniques of pattern layout and texture appearance simulation of traditional Ethnic cloth, a data-driven intelligent design and simulation model based on main color feature learning is proposed in this paper. We employ a combination-based design technique, which uses pattern elements data and skeletons data to design a digital layout for ethnic cloth. We use Octree to quantize the color map for the designed ethnic cloth layout and extract main color map based on k-means clustering. Using a cubic convolution interpolation algorithm with yarn structure template, we render each region that is segmented through the main color map. Then, we can generate a good representation of the texture appearance of the designed layout which is shown as a realistic fabric material. Finally, the designed layout will be transferred to intelligent loom produced based on industrial Internet of things. We design several traditional Ethic cloths (Tujia brocade) layouts and simulate their textures based on our method to analyze its applicability and validity. We also compare the design and simulation results with previously proposed algorithms, which indicate that our model can design complex patterns and simulate exquisite material of Tujia brocade. (C) 2019 Elsevier B.V. All rights reserved.