计算机应用 ›› 2010, Vol. 30 ›› Issue (06): 1629-1631.

• 图形图像处理与模式识别 • 上一篇    下一篇

结合蚁群优化搜索的图像综合特征类比在风格生成中的应用

钱哨   

  1. 交通部管理干部学院计算机系
  • 收稿日期:2009-12-04 修回日期:2010-01-26 发布日期:2010-06-01 出版日期:2010-06-01
  • 通讯作者: 钱哨
  • 基金资助:
    北京市自然科学基金资助项目(70053073)

Artistic style generation by image analogy using ant algorithm based on comprehensive area feathers

Qian Shao   

  • Received:2009-12-04 Revised:2010-01-26 Online:2010-06-01 Published:2010-06-01
  • Contact: Qian Shao

摘要: 为了解决图像风格虚拟生成过程中的参数控制困难以及运算时间较长的问题,提出了一种鲁棒的方法。该方法首先采用基于图像分割的综合特征区域纹理特征生成算法,通过与样本图像相似区域匹配获得新风格的图像;并将蚁群优化算法运用到纹理的匹配过程中以加速整个流程。实验结果表明,该算法在缩短算法时间的同时可以通过参数控制所得图像风格;以蚁群优化为代表的随机性算法能较大程度提高图像风格虚拟生成系统的综合性能。

关键词: 图像类比, 纹理合成, 风格生成, 蚁群优化搜索

Abstract: This paper presented a robust scheme for texture synthesis in the process of artistic style learning based on the comprehensive area feathers of the segmented image blobs, which speeded up the process by exploiting the ant colony algorithm for the texture blobs matching firstly. The scheme also ensures the artistic style transition by the parameter setting of different sample images weights and shortens the time cost in the style generation process, which can transit the styles of several images to the targets. The preliminary experiments illustrate the good performance of the scheme. Above all, the random algorithms just like the ant colony searching algorithm can enhance the comprehensive performance of the artistic style generation system.

Key words: Image analogy, Texture synthesis, Artistic style generation, Ant conoly alogrithm for search