基于感兴趣区域的幅型比变换新方法

A Novel Method of Aspect Ratio Conversion Based on Region of Interest

基于人类视觉系统对运动目标比较敏感的特性,提出了一种新的幅型比变换方法.综合时空信息提取视频序列中的运动目标,利用图像修复技术对被提取运动目标后的视频图像进行背景修补.采用插值算法以不同的方式分别缩放运动目标和修复后的背景,并将变换后的运动目标与背景相叠加合成相应幅型比的视频图像.实验结果表明:此方法避免了已有幅型比变换方法的缺陷,保证了变换后运动目标的无失真性,与人类的视觉感知有较好的一致性.

Based on Ratio Conversion（ the fact that the human visual system is sensitive to the moving objects,a novel Aspect ARC） method was proposed for the format conversion between different aspect ratios. In the new ARC algorithm, the spatio-temporal information was integrated to separate the moving objects from the video sequences, and the image inpainting technique was used to restore the background. Interpolation method was employed to scale the moving objects and restored background in different ratios respectively. And,the scaled moving objects and background were merged into a new image with proper aspect ratio. The experimental results demonstrate that the proposed ARC method avoids the distortion of moving objects and achieves better perception effect.