融合书画艺术元素的家居软装饰图案优化设计及应用

Optimization Design and Application of Home Soft Decoration Patterns Integrating Calligraphy and Painting Art Elements

中文关键词: 书画艺术  家居  软装饰  图案设计  图像迁移

英文关键词: calligraphy and painting art  home furnishing  soft decoration  pattern design  image migration

中文摘要:

      目的 利用中国书画元素对家居软装饰图案进行优化设计。方法 在分析中国书画艺术元素特点和内容的基础上，结合家居软装饰的构成因素及风格特征，将构图、线条、赋色和笔墨应用于软装饰的优化设计中。随后将基于卷积神经网络的超分辨率模型(Efficient Sub-Pixel Convolutional Neural Network，ESPCN)与Laplace算子相结合，形成Laplace-ESPCN模型，由此进行整体装饰图案的迁移，提出了基于泊松融合的局部迁移方法并加以完善。结果 所提出的整体装饰图案迁移模型在噪声密度超过50%时，品质因素仍然在0.5左右，并且峰值信噪比最低可达40 dB。将其与泊松融合的局部迁移结合后，目标区域分割下人数占比均在50%附近。结论 所提出的Laplace-ESPCN整体图案迁移模型得到的图像质量较高，具有更好的降噪效果。与泊松融合的局部迁移方法结合后，所得图像边缘更加平滑，可以在家居软装饰图案优化设计中有效应用，具有一定的推广价值。

英文摘要:

      The work aims to optimize the design of home soft decoration patterns with Chinese calligraphy and painting elements. On the basis of analyzing the characteristics and content of Chinese calligraphy and painting art elements, combined with the constituent factors and style characteristics of home soft decoration, composition, lines, coloring, and ink were applied to the optimization design of soft decoration. Subsequently, the Efficient Sub Pixel Convolutional Neural Network (ESPCN) based on convolutional neural networks was combined with the Laplace operator to form a Laplace ESPCN model for overall decorative pattern migration, and a local migration method based on Poisson fusion was proposed to improve it. The proposed overall decorative pattern transfer model still had a quality factor of around 0.5 when the noise density exceeded 50%, and the minimum peak signal-to-noise ratio could reach 40 dB. After combining it with the local transfer of Poisson fusion, the proportion of people in the target area segmentation was around 50%. The Laplace-ESPCN overall pattern transfer model proposed in the conclusion results in high image quality and good noise reduction effect. After combining with the local transfer method of Poisson fusion, the obtained image edges are smoother, which can be effectively applied in the optimization design of home soft decoration patterns and has certain promotion value.