物体基视频编码中的运动物体分割

针对物体基视频编码应用，提出了一种基于运动的图像序列分割方案，它同时结合了亮度和运动信息，使用重构式开－闭形态滤波器等形态变换工具。此外，还提出了一种区域生长型的分水岭算法，它直接对要分割的图像而不是其形态梯度信号进行分割，克服了传统形态分割方法使用形态梯度信号而导致轮廓信息丢失的缺点。在具体实现算法时，文中采用了一种基于分级队列结构进行图像扫描的巧妙方法，大大提高了算法的效率。模拟结果表明，该方

Segmentation of Moving Objects for Object-Based Video Coding

A motion based segmentation scheme for the object based video coding is proposed, which incorporates luminance and motion information simultaneously and uses morphological tools such as open close by reconstruction. It also proposes a region growing version of watershed algorithm. This algorithm works directly on the image to be segmented instead of the morphological gradient signal and solves the problem existing in the classical morphological segmentation approach in which the use of the morphological gradient results in a loss of information on the contour position. Furthermore, in order to implement this algorithm efficiently, this paper proposes a clever scanning of the images defined by hierarchical queues. Simulation results demonstrate that the resulting objects are characterized by a coherent motion and the moving object boundaries are precisely located.