【摘要】 随着多媒体技术的发展,图像编辑技术越来越受到人们的关注和重视,是计算机视觉和机器学习领域的热点和难点问题。图像合成是图像编辑中的一种重要方法,现有的图像合成方法或者未考虑图像的高度或法相信息,导致图像合成效果不真实;或者需要复杂的交互,费时费力。为此,本文提出了一种新的图像合成算法,基于三维变形来表示图像的高度或法相信息,不需要人工交互,计算速度快。多组实验结果证明,本文方法的效果优于目前其他方法。更多还原

【Abstract】 With the development of multimedia technologies,image editing are more and more popular and become a hot research topic in the computer vision and machine learning communities. Image composition is an important method in image editing. Existing image composition methods either do not consider the height or normal phase information of images so that resulting in non-realistic effect,or require complex interaction so that are time-consuming and laborious. Therefore,this paper proposes a new image composition algorithm based on three-dimensional deformation to represent the height or normal phase information of an image. It needs no manual interaction and is fast to calculate. Experimental results show that the performance of this method is better than other methods.更多还原