TV symbol recognition based on shape and color features

被引：0

作者：

Shi, Ying-Chun

Zhou, Xian-Zhong

Fang, Peng-Fei

机构：

[1] Department of Automation, Nanjing University of Science and Technology, Nanjing 210094, China

[2] College of Engineer Management, Nanjing University, Nanjing 210093, China

[3] Department of Computer Science and Engineering, Southeast University, Nanjing 210096, China

来源：

Moshi Shibie yu Rengong Zhineng/Pattern Recognition and Artificial Intelligence | 2005年 / 18卷 / 02期

关键词：

D O I：

暂无

中图分类号：

学科分类号：

摘要：

As a sign of a TV station, its symbol often implies semantics about its name and program tendency and it is one of the important semantic sources for video analysis, comprehension and retrieval. According to spatio-temporal visual feature, TV station symbol is firstly segmented by the detection of spatio-temporal invariant region, then described by combining space-distribution histogram and color histogram in HSV space, and recognized by support vector machine (SVM) finally. The algorithm is experimentally satisfied.

引用

页码：216 / 222