基于多策略改进麻雀搜索算法优化SVM的古建筑彩画分割

Optimization of SVM for Segmentation of Ancient Architectural Paintings Based on Multi Strategy Improved Sparrow Search Algorithm

为精确地提取古建筑彩画的色彩、图案与轮廓信息,本文利用网络搜索优化SLIC超像素分割,通过多策略改进的麻雀搜索算法优化SVM的超参数选择过程,获取古建筑彩画分割的精确分类结果,最后利用ArcGIS软件进行分割结果的矢量化.实验表明:本文算法较原SSA-SVM具有更高的精度和效率,最终的彩画矢量化结果优异.

In order to accurately extract the color,pattern,and contour information of ancient architectural paintings,this article uses network search to optimize SLIC super-pixel segmentation.The Multi-strategy Improved Sparrow Search Algorithm has been used to optimize the hyperparameter selection process of SVM classification,and the accurate classification results of ancient architecture painting were obtained in this study.Finally,the ArcGIS software has been used for vectorization of segmentation results.Experimental results show that the proposed algorithm has higher accuracy and efficiency than the original SSA-SVM,and the final color painting vectorization results are excellent.