Efficient automatic text location method and content-based indexing and structuring of the video database

semantic audiovisual objects for indexing and retrieval of image and video documents has used relatively low-levelAn efficient automatic text detection and location method for video documents is proposed and its application for the perceptual content analysis such as texture (Picard and content-based retrieval of video is presented and discussed. Minka, 1995), shape (Sclaroff and Pentland, 1995; Tegolo, Target frames are selected at fixed time intervals from shots 1994), and color (Sakamoto et al., 1994) in image data, and detected by a scene-change detection method. For each selected various sound parameters for audio (Blum et al., 1995). frame, segmentation by color clustering is performed around Automatic or semi-automatic methods are favored, for the color peaks using a color histogram. For each color plane, text-sake of excluding subjectivism and saving the labor costlines are detected using heuristics, and the temporal and spatial and tedious repetitive work associated with human in-position and the text-image of each text-line are stored in a dexing. It turns out that some important semantic cluesdatabase. Experimental results for text detection in video im-can be recovered directly from the perceptual contents.ages and the performance of the method are reported for various Recently, researchers have tried to propose methods forvideo documents. A user interface for text-image based brows-ing is designed for direct content-based access to video docu- semantic speech or text-based retrieval of video. Some ments, and other applications are discussed. 1996 Academic Press researchers have proposed the textual indexing of nontex