摘要
视频流里的文本字幕包含了丰富的语义信息,对于视频内容的理解、索引和检索具有重要的作用.文中首先对视频图像进行分割,提取它们在小波变换域中的统计特征,然后用支持向量机(SVM)分类器对这些特征进行分类,实现了字幕的自动检测与提取.仿真结果表明该方法具有较好的适用性.
The captions in a video stream contain rich semantic information and play an important role in perceiving, indexing and retrieving video contents. In this paper, video images are segmented and the statistical attributes in wavelet translation domain are extracted. Then, by using the Support Vector Machine, the attributes are classified to implement the automatic detection and extraction of video captions. Simulation results indicate that the proposed method is of good applicability.
出处
《华南理工大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2004年第z1期63-66,共4页
Journal of South China University of Technology(Natural Science Edition)
关键词
视频检索
文本检测
HAAR变换
支持向量机
video retrieval
text detection
Haar translation
Support Vector Machine
