本文对现有的电视图象分类算法和检索算法作了的分析，并在图象分类方面提出了一种自适应分类算法；在检索方面提出了一种基于颜色的样本与非样本学习相结合的方法。自适应分类算法的计算量与相邻帧间的变化量成正比；对纡绝大多数相邻帧，只要计算数目极小的区域就可完成分类工作。实验证明，本文提出的分类与检索方法在速度与精度上均有明显的提高。

ummary of methods for video classification and indexing is given in this paper. An adaptive algorithm is proposed and a method using color trade-off between object and background is also suggested to video indexing. The time-cost for adaptive algouth is dependent on the amount of change between cosecutine frames. For most of neighbor frames, the number of local regions used for classification is very small. Experimental results show that the proposed methods for video classification and indexing are fast and accurate.