摘要
边避免小波对图像多层次、多尺度的细节增强,以突出前景对象的边缘信息并平滑复杂的背景结构。在滑动窗口采样过程中,利用边避免小波突出采样空间的边缘细节信息以增强采样像素点的质量。在YUV色彩空间利用小波变换增强前景对象的细节信息。实验证明,利用小波对边缘细节增强后,在处理边缘颜色相差较小或者前景和背景交错的图像时可以取得高质量的结果。
The edge avoidance wavelet is used to highlight the edge information of the foreground item and smooth the complex background structure in this work. By applying the edge wavelet to improve the quality of the sample pixels, the edge details of the sampling space are avoided throughout the sliding window sampling procedure. The wavelet transform is used in this paper to improve the details of foreground objects in YUV color space. Experiments reveal that when processing photos with modest edge color changes or interlaced foreground and background images, high-quality results can be achieved after employing wavelet to enhance edge features.
作者
李增禄
郭孝玉
林志兴
刘孙发
余建
许章华
LI Zeng-lu;GUO Xiaoyu;LIN Zhi-xing;LIU Sun-fa;YU Jian;XU Zhang-hua(Network Center(Information Construction Office),Sanming University,Sanming 365004,China;Fujian Provincial Key Laboratory of Resources and Environment Monitoring and Sustainable Management and Utilization,School of Resources and Chemical Engineering,Sanming University,Sanming 365004,China;Geography and Ecological Environment Research Center,College of Environment and Safety Engineering,Fuzhou University,Fuzhou 350001,China)
出处
《三明学院学报》
2022年第3期66-75,共10页
Journal of Sanming University
基金
国家教育部教师函[2021]13号第二批人工智能助推教师队伍建设试点项目(三明学院),三明学院科学研究发展基金暨福建省中青年教师教育科研项目(JAT200620/B202011,JAT200627/B202018)
三明市科技局引导性项目(2020-S-16)。
关键词
抠图
边避免小波
贝叶斯
边缘增强
matting
edge avoiding wavelet
bayesian
edge enhancement
