作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程 ›› 2011, Vol. 37 ›› Issue (17): 286-287,290. doi: 10.3969/j.issn.1000-3428.2011.17.096

• 开发研究与设计技术 • 上一篇    下一篇

基于优化三维均值坐标的混合式视频融合

刘奇松a,马利庄b   

  1. (上海交通大学 a. 软件学院;b. 电子信息与电气工程学院,上海 200240)
  • 收稿日期:2011-03-15 出版日期:2011-09-05 发布日期:2011-09-05
  • 作者简介:刘奇松(1985-),男,硕士研究生,主研方向:视频融合,图像增强技术;马利庄,教授、博士生导师
  • 基金资助:
    国家“863”计划基金资助项目(2009AA01Z334);国家自然科学基金资助项目(61073089)

Hybrid Video Fusion Based on Optimized 3D Mean-value Coordinates

LIU Qi-song  a, MA Li-zhuang  b   

  1. (a. School of Software; b. School of Electronic, Information and Electrical Engineering, Shanghai Jiaotong University, Shanghai 200240, China)
  • Received:2011-03-15 Online:2011-09-05 Published:2011-09-05

摘要: 基于二维均值坐标插值的视频融合方法会产生局部色彩失衡和画面闪烁等失真现象。为此,提出一种基于优化三维均值坐标的视频融合方法。在连续的视频序列上建立三维均值坐标,根据不同的边界情况优化均值坐标的计算,并利用图像matting技术提高融合效果。实验结果表明,与现有方法相比,该方法可以更有效地提高融合质量。

关键词: 均值坐标, 混合式融合, 视频融合, 视频处理, 时空一致性

Abstract: Interpolation-based fusion method using 2D mean-value coordinates may product some unnatural effects. Aiming at this problem, this paper presents a new method of video fusion to avoid these flaws. Based on 3D mean-value coordinates, a boundary-aware algorithm is provided to optimize the determination of 3D mean-value coordinates. Matting-based methods are used to refine produced video. Experimental results show that this method can enhance the quality of composition more effectively compared with the existing approaches.

Key words: mean-value coordinates, hybrid fusion, video fusion, video processing, temporal and spatial consistency

中图分类号: