摘要
首先介绍了立体视觉的基本原理,然后对立体图像的压缩方法分四类进行了综述。对其中用于立体图像序列的两种主要方法:基于“块”匹配的立体图像压缩方法和基于物体的立体图像压缩方法进行了深入探讨。通过对已有成果进行总结和分类,剖析了两种方法的优、缺点,并提出了一些还需要深入研究的问题,如:残差图像编码、遮挡检测、更精确的场景分割等。
This paper firstly introduces the basic theory of stereo vision. And then, compression methods for stereo image are summarized. The compression method based on block matching and object are discussed in detail. The advantages and disadvantages of these two methods are analyzed based on summing up and classifying the present achievements. Lastly, some problems which will be researched in the future are put forward, i.e., error image coding, occlusion detection, natural scene segmentation and so on.
出处
《通信学报》
EI
CSCD
北大核心
2003年第6期113-123,共11页
Journal on Communications
基金
教育部科研重点基金资助项目(02072)
关键词
立体视觉
图像压缩
块匹配
物体基
stereo vision
image compression
block matching
object-based
