基于图像表征特性的轮廓线提取的研究

CONTOUR GROUPING BASED ON IMAGE PATCHES AROUND EDGES

轮廓线提取是计算机视觉中的一个重要问题。但是,用来解决这个问题的传统方法总是不能得到像人的视觉系统一样令人满意的结果。人的视觉系统一个重要的特点,就是人眼在获得轮廓线的时候,不仅对轮廓线两侧图像的差异及轮廓线本身几何拓扑结构上有所考虑,同时也考虑到了轮廓线周围区域图像的特性(图像的亮度、颜色、纹理等特征)的连续性。然而计算机解决这个问题的传统方法通常忽略了后一个问题。这些方法总是被分成边缘检测和边缘连接两步骤而当边缘被检测出来之后,它们周围区域图像的特性就不会再被考虑到,导致了属于不同物体的边缘被错误地连在一个轮廓线中。通过提取表示边缘的线段并连接它们的方法找出一幅图像中最显著的轮廓线。引入了一个"表征特性"来描述以表示边缘的线段周围的图像的特性,然后这个特征被用来定义一个度量轮廓线显著程度的尺度"显著度估计"中。这样,图像中最显著的一个完整轮廓线就可以通过对这个"显著度估计"的优化来得到。最后,这个方法的有效性在实验中得到了验证。

Contour grouping is an important issue in computer vision, However,traditional ways tackling the problem usually fail to provide as satisfying results as human vision can do. One important feature of human vision mechanism is that human vision tends to group together ed- ges that are not only geometrically and topologically related, but also similar in their appearances-the appearances of image patches around them including their brightness, color, texture cues, etc. But in traditional grouping approaches, after edges or lines have been detected, the ap- pearances of image patches around them are seldom considered again, which leads to the results that edges belonging to boundaries of different objects are sometimes falsely grouped together. An appearance feature is introduced to describe the appearance of an image patch around a line segment ,and this appearance feature is incorporated into a saliency measure to evaluate contours of an image. The most salient contour is found through the optimization of this saliency measure by using a genetic algorithm. Experimental resulfs prove the effectiveness of the approach.