Detection and Interpretation of Text Information in Noisy Video Sequences

摘要: Text superimposed on the video frames provides supplemental but important information for indexing and retrieval. The detection recognition of text from is thus an issue in automated content-based visual archives. interest not limited to static text. They could be scrolling a linear motion where only part available during different video. problem further complicated if corrupted with noise. An algorithm proposed detect, classify segment both simple moving complex noisy background. extracted texts are processed using averaging attain quality suitable by commercial Optical Character Recognition (OCR) software.