Text Processing in Video Frames with Complex Background

Information deficiency is a huge problem when researching on video indexing and retrieval. On the other hand, text in video frames implies lots of semantics inherently, and can provide supplemental but important information for video data processing. In this paper, we present a fast and robust approach for text detection, localization, extraction, and reorganization in video frames with complex background. Here, block change rate (BCR for short) is imported to realize text detection and localization, smoothness model is used to narrow the scope of the text stroke, element image division in Lab color space is implied in binary text extraction, and Langue model is imported to evaluate the text extraction results. Experiments based on a large amount of video frames from different sources show that this approach is robust, effective and compatible for variety videos with complex background.