Acta Scientiarum Naturalium Universitatis Pekinensis

Previous Articles     Next Articles

A Novel Algorithm for the Video Caption Extraction

CAO Xixin1, LIU Jing1, YANG Xudong1, WU Shuai1, ZHANG Qihui2   

  1. 1.School of Software and Microelectronics, Peking University, Beijing 102600; 2. School of Physics and Electronics, Henan University, Kaifeng 475001;
  • Received:2012-02-23 Online:2013-03-20 Published:2013-03-20



  1. 1. 北京大学软件与微电子学院, 北京102600; 2. 河南大学物理与电子学院, 开封475001;

Abstract: The authors propose a novel algorithm for the video caption extraction. By using the image sharpening method, a better extraction effect can be achieved for the special video sequences. The caption area and the non-caption area can be efficiently separated by neighborhood average operation. A projection algorithm is proposed to implement on a better decomposition and extraction based on the statistical characteristics, and this can efficiently separate the caption area and make the extraction coordinates more precise. The experiments show that the proposed algorithm has a better effect compared with the original methods. Because of the low computational complexity, it can be widely used in the field of video image processing.

Key words: caption extraction, edge intensity, image enhancement pre-processing, projection statistical analysis

摘要: 提出一种基于边缘强度的视频图像字幕提取算法。此算法通过对图像进行锐化预处理, 使得对特定的视频序列可以达到较好的提取效果; 通过领域平均处理, 更好地分离了字幕区域和背景区域; 提出一个基于投影统计特征分析的算法, 能够更好、更高效、更精确地分离和提取字幕。实验证明, 该算法在效果上优于其他的算法, 并且该算法复杂度低, 实用价值高, 可以广泛用于视频编解码等过程中。

关键词: 字幕提取, 边缘强度, 图像增强预处理, 投影统计分析

CLC Number: